diff --git a/Cargo.lock b/Cargo.lock
index a7a68203b2a5..d69c13da4644 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -2435,6 +2435,7 @@ dependencies = [
  "criterion",
  "databend-common-base",
  "databend-common-exception",
+ "databend-common-functions",
  "databend-common-io",
  "databend-common-meta-app",
  "enum-as-inner 0.5.1",
@@ -2751,7 +2752,6 @@ dependencies = [
  "borsh 1.2.1",
  "bstr 1.6.2",
  "bumpalo",
- "bytes",
  "chrono",
  "chrono-tz",
  "comfy-table 6.1.4",
@@ -2793,6 +2793,7 @@ dependencies = [
  "siphasher",
  "streaming_algorithms",
  "strength_reduce",
+ "stringslice",
  "twox-hash",
 ]
 
@@ -3195,6 +3196,7 @@ dependencies = [
  "databend-common-expression",
  "futures",
  "minitrace",
+ "once_cell",
  "petgraph",
  "serde",
  "serde_json",
@@ -3261,7 +3263,6 @@ dependencies = [
  "databend-common-exception",
  "databend-common-expression",
  "databend-common-pipeline-core",
- "databend-common-profile",
  "itertools 0.10.5",
  "jsonb 0.3.0 (git+https://github.com/datafuselabs/jsonb?rev=a0669bf)",
  "match-template",
@@ -3270,14 +3271,6 @@ dependencies = [
  "typetag",
 ]
 
-[[package]]
-name = "databend-common-profile"
-version = "0.1.0"
-dependencies = [
- "dashmap",
- "databend-common-base",
-]
-
 [[package]]
 name = "databend-common-proto-conv"
 version = "0.1.0"
@@ -3388,7 +3381,6 @@ dependencies = [
  "databend-common-pipeline-core",
  "databend-common-pipeline-sources",
  "databend-common-pipeline-transforms",
- "databend-common-profile",
  "databend-common-settings",
  "databend-common-storage",
  "databend-common-storages-delta",
@@ -3856,7 +3848,6 @@ dependencies = [
  "databend-common-metrics",
  "databend-common-pipeline-core",
  "databend-common-pipeline-sources",
- "databend-common-profile",
  "databend-common-sql",
  "databend-common-storage",
  "databend-common-storages-fuse",
@@ -4233,7 +4224,6 @@ dependencies = [
  "databend-common-pipeline-sinks",
  "databend-common-pipeline-sources",
  "databend-common-pipeline-transforms",
- "databend-common-profile",
  "databend-common-settings",
  "databend-common-sharing",
  "databend-common-sql",
@@ -12164,6 +12154,12 @@ version = "0.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "fe895eb47f22e2ddd4dabc02bce419d2e643c8e3b585c78158b349195bc24d82"
 
+[[package]]
+name = "stringslice"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8dad614e9c8a0773603982e0c1d3192f2bce1c0ee193caf0526b8ae3c6c3e38c"
+
 [[package]]
 name = "strsim"
 version = "0.10.0"
diff --git a/Cargo.toml b/Cargo.toml
index b7a7404a2d93..8771d38de23a 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -46,7 +46,6 @@ members = [
     "src/query/pipeline/sinks",
     "src/query/pipeline/sources",
     "src/query/pipeline/transforms",
-    "src/query/profile",
     "src/query/settings",
     "src/query/sql",
     "src/query/storages/common/blocks",
diff --git a/src/common/cloud_control/proto/task.proto b/src/common/cloud_control/proto/task.proto
index eb19965a06b6..dda4f3776405 100644
--- a/src/common/cloud_control/proto/task.proto
+++ b/src/common/cloud_control/proto/task.proto
@@ -33,6 +33,7 @@ message CreateTaskRequest {
   // DAG
   repeated string after = 11; //
   optional string when_condition = 12;
+  map<string, string> session_parameters = 13;
 }
 
 message TaskError {
@@ -90,6 +91,7 @@ message Task {
   optional string last_suspended_at = 16;
   repeated string after = 17;
   optional string when_condition = 18;
+  map<string, string> session_parameters = 19;
 }
 
 
@@ -134,6 +136,8 @@ message AlterTaskRequest {
   optional string when_condition = 11;
   repeated string add_after = 12;
   repeated string remove_after = 13;
+  bool set_session_parameters = 14;
+  map<string, string> session_parameters = 15;
 }
 
 message AlterTaskResponse {
@@ -179,6 +183,7 @@ message TaskRun {
   string query_id = 17;
   string condition_text = 18;
   string root_task_id = 19;
+  map<string, string> session_parameters = 20;
 }
 
 message ShowTaskRunsResponse {
diff --git a/src/common/cloud_control/src/task_utils.rs b/src/common/cloud_control/src/task_utils.rs
index 1615ec0753e9..be8b17cfcad5 100644
--- a/src/common/cloud_control/src/task_utils.rs
+++ b/src/common/cloud_control/src/task_utils.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::BTreeMap;
 use std::fmt::Display;
 use std::fmt::Formatter;
 
@@ -77,6 +78,7 @@ pub struct Task {
     pub created_at: DateTime<Utc>,
     pub updated_at: DateTime<Utc>,
     pub last_suspended_at: Option<DateTime<Utc>>,
+    pub session_params: BTreeMap<String, String>,
 }
 
 pub fn format_schedule_options(s: &ScheduleOptions) -> Result<String> {
@@ -203,6 +205,7 @@ impl TryFrom<crate::pb::Task> for Task {
             status,
             created_at,
             updated_at,
+            session_params: value.session_parameters,
         };
         Ok(t)
     }
@@ -227,6 +230,7 @@ pub struct TaskRun {
     pub error_code: i64,
     pub error_message: Option<String>,
     pub root_task_id: String,
+    pub session_params: BTreeMap<String, String>,
 }
 
 // convert from crate::pb::taskRun to struct taskRun
@@ -305,6 +309,7 @@ impl TryFrom<crate::pb::TaskRun> for TaskRun {
             scheduled_at,
             completed_at,
             root_task_id: value.root_task_id,
+            session_params: value.session_parameters,
         };
         Ok(tr)
     }
diff --git a/src/common/cloud_control/tests/it/task_client.rs b/src/common/cloud_control/tests/it/task_client.rs
index e8b09118fd20..68fa51f248a1 100644
--- a/src/common/cloud_control/tests/it/task_client.rs
+++ b/src/common/cloud_control/tests/it/task_client.rs
@@ -82,6 +82,7 @@ impl TaskService for MockTaskService {
                 last_suspended_at: None,
                 after: vec![],
                 when_condition: None,
+                session_parameters: Default::default(),
             }),
             error: None,
         }))
@@ -196,6 +197,7 @@ async fn test_task_client_success_cases() -> Result<()> {
         if_not_exist: false,
         after: vec![],
         when_condition: None,
+        session_parameters: Default::default(),
     });
 
     let response = client.create_task(request).await?;
diff --git a/src/common/hashtable/src/hashjoin_hashtable.rs b/src/common/hashtable/src/hashjoin_hashtable.rs
index 7c126159fa14..235200e83f7f 100644
--- a/src/common/hashtable/src/hashjoin_hashtable.rs
+++ b/src/common/hashtable/src/hashjoin_hashtable.rs
@@ -209,56 +209,80 @@ where
         count
     }
 
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    fn early_filtering_probe(&self, hashes: &mut [u64], bitmap: Option<Bitmap>) -> usize {
+    // Perform early filtering probe, store matched indexes in `matched_selection` and store unmatched indexes
+    // in `unmatched_selection`, return the number of matched and unmatched indexes.
+    fn early_filtering_probe(
+        &self,
+        hashes: &mut [u64],
+        bitmap: Option<Bitmap>,
+        matched_selection: &mut [u32],
+        unmatched_selection: &mut [u32],
+    ) -> (usize, usize) {
         let mut valids = None;
         if let Some(bitmap) = bitmap {
             if bitmap.unset_bits() == bitmap.len() {
-                hashes.iter_mut().for_each(|hash| {
-                    *hash = 0;
-                });
-                return 0;
+                unmatched_selection
+                    .iter_mut()
+                    .enumerate()
+                    .for_each(|(idx, val)| {
+                        *val = idx as u32;
+                    });
+                return (0, hashes.len());
             } else if bitmap.unset_bits() > 0 {
                 valids = Some(bitmap);
             }
         }
-        let mut count = 0;
+        let mut matched_idx = 0;
+        let mut unmatched_idx = 0;
         match valids {
             Some(valids) => {
-                valids
-                    .iter()
-                    .zip(hashes.iter_mut())
-                    .for_each(|(valid, hash)| {
+                valids.iter().zip(hashes.iter_mut().enumerate()).for_each(
+                    |(valid, (idx, hash))| {
                         if valid {
                             let header = self.pointers[(*hash >> self.hash_shift) as usize];
                             if header != 0 && early_filtering(header, *hash) {
                                 *hash = remove_header_tag(header);
-                                count += 1;
+                                unsafe {
+                                    *matched_selection.get_unchecked_mut(matched_idx) = idx as u32
+                                };
+                                matched_idx += 1;
                             } else {
-                                *hash = 0;
+                                unsafe {
+                                    *unmatched_selection.get_unchecked_mut(unmatched_idx) =
+                                        idx as u32
+                                };
+                                unmatched_idx += 1;
                             }
                         } else {
-                            *hash = 0;
+                            unsafe {
+                                *unmatched_selection.get_unchecked_mut(unmatched_idx) = idx as u32
+                            };
+                            unmatched_idx += 1;
                         }
-                    });
+                    },
+                );
             }
             None => {
-                hashes.iter_mut().for_each(|hash| {
+                hashes.iter_mut().enumerate().for_each(|(idx, hash)| {
                     let header = self.pointers[(*hash >> self.hash_shift) as usize];
                     if header != 0 && early_filtering(header, *hash) {
                         *hash = remove_header_tag(header);
-                        count += 1;
+                        unsafe { *matched_selection.get_unchecked_mut(matched_idx) = idx as u32 };
+                        matched_idx += 1;
                     } else {
-                        *hash = 0;
+                        unsafe {
+                            *unmatched_selection.get_unchecked_mut(unmatched_idx) = idx as u32
+                        };
+                        unmatched_idx += 1;
                     }
                 });
             }
         }
-        count
+        (matched_idx, unmatched_idx)
     }
 
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    fn early_filtering_probe_with_selection(
+    // Perform early filtering probe and store matched indexes in `selection`, return the number of matched indexes.
+    fn early_filtering_matched_probe(
         &self,
         hashes: &mut [u64],
         bitmap: Option<Bitmap>,
diff --git a/src/common/hashtable/src/hashjoin_string_hashtable.rs b/src/common/hashtable/src/hashjoin_string_hashtable.rs
index c7ca141e7f03..7112d2ee68fb 100644
--- a/src/common/hashtable/src/hashjoin_string_hashtable.rs
+++ b/src/common/hashtable/src/hashjoin_string_hashtable.rs
@@ -138,20 +138,31 @@ where A: Allocator + Clone + 'static
         count
     }
 
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    fn early_filtering_probe(&self, hashes: &mut [u64], bitmap: Option<Bitmap>) -> usize {
+    // Perform early filtering probe, store matched indexes in `matched_selection` and store unmatched indexes
+    // in `unmatched_selection`, return the number of matched and unmatched indexes.
+    fn early_filtering_probe(
+        &self,
+        hashes: &mut [u64],
+        bitmap: Option<Bitmap>,
+        matched_selection: &mut [u32],
+        unmatched_selection: &mut [u32],
+    ) -> (usize, usize) {
         let mut valids = None;
         if let Some(bitmap) = bitmap {
             if bitmap.unset_bits() == bitmap.len() {
-                hashes.iter_mut().for_each(|hash| {
-                    *hash = 0;
-                });
-                return 0;
+                unmatched_selection
+                    .iter_mut()
+                    .enumerate()
+                    .for_each(|(idx, val)| {
+                        *val = idx as u32;
+                    });
+                return (0, hashes.len());
             } else if bitmap.unset_bits() > 0 {
                 valids = Some(bitmap);
             }
         }
-        let mut count = 0;
+        let mut matched_idx = 0;
+        let mut unmatched_idx = 0;
         match valids {
             Some(valids) => {
                 hashes.iter_mut().enumerate().for_each(|(idx, hash)| {
@@ -159,32 +170,45 @@ where A: Allocator + Clone + 'static
                         let header = self.pointers[(*hash >> self.hash_shift) as usize];
                         if header != 0 && early_filtering(header, *hash) {
                             *hash = remove_header_tag(header);
-                            count += 1;
+                            unsafe {
+                                *matched_selection.get_unchecked_mut(matched_idx) = idx as u32
+                            };
+                            matched_idx += 1;
                         } else {
-                            *hash = 0;
+                            unsafe {
+                                *unmatched_selection.get_unchecked_mut(unmatched_idx) = idx as u32
+                            };
+                            unmatched_idx += 1;
                         }
                     } else {
-                        *hash = 0;
-                    };
+                        unsafe {
+                            *unmatched_selection.get_unchecked_mut(unmatched_idx) = idx as u32
+                        };
+                        unmatched_idx += 1;
+                    }
                 });
             }
             None => {
-                hashes.iter_mut().for_each(|hash| {
+                hashes.iter_mut().enumerate().for_each(|(idx, hash)| {
                     let header = self.pointers[(*hash >> self.hash_shift) as usize];
                     if header != 0 && early_filtering(header, *hash) {
                         *hash = remove_header_tag(header);
-                        count += 1;
+                        unsafe { *matched_selection.get_unchecked_mut(matched_idx) = idx as u32 };
+                        matched_idx += 1;
                     } else {
-                        *hash = 0;
+                        unsafe {
+                            *unmatched_selection.get_unchecked_mut(unmatched_idx) = idx as u32
+                        };
+                        unmatched_idx += 1;
                     }
                 });
             }
         }
-        count
+        (matched_idx, unmatched_idx)
     }
 
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    fn early_filtering_probe_with_selection(
+    // Perform early filtering probe and store matched indexes in `selection`, return the number of matched indexes.
+    fn early_filtering_matched_probe(
         &self,
         hashes: &mut [u64],
         bitmap: Option<Bitmap>,
diff --git a/src/common/hashtable/src/traits.rs b/src/common/hashtable/src/traits.rs
index f49ad9472b8b..14ff77790d80 100644
--- a/src/common/hashtable/src/traits.rs
+++ b/src/common/hashtable/src/traits.rs
@@ -351,6 +351,13 @@ impl FastHash for [u8] {
     }
 }
 
+impl FastHash for str {
+    #[inline(always)]
+    fn fast_hash(&self) -> u64 {
+        self.as_bytes().fast_hash()
+    }
+}
+
 // trick for unsized_hashtable
 impl<const N: usize> FastHash for ([u64; N], NonZeroU64) {
     #[inline(always)]
@@ -506,24 +513,21 @@ pub trait HashtableLike {
 pub trait HashJoinHashtableLike {
     type Key: ?Sized;
 
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    // same with `early_filtering_probe`, but we don't use early_filter
+    // Probe hash table, use `hashes` to probe hash table and convert it in-place to pointers for memory reuse.
     fn probe(&self, hashes: &mut [u64], bitmap: Option<Bitmap>) -> usize;
 
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    // 1. same with `early_filtering_probe_with_selection`, but we don't use selection to preserve the
-    // unfiltered indexes, we just set the filtered hashes as zero.
-    // 2. return the unfiltered counts.
-    fn early_filtering_probe(&self, hashes: &mut [u64], bitmap: Option<Bitmap>) -> usize;
-
-    // Using hashes to probe hash table and converting them in-place to pointers for memory reuse.
-    // we use `early_filtering_probe_with_selection` to do the first round probe.
-    // 1. `hashes` is the hash value of probe block's rows. we will use this one to
-    // do early filtering. if we can't early filter one row(at idx), we will assign pointer in
-    // the bucket to hashes[idx] to reuse the memory.
-    // 2. `selection` is used to preserved the indexes which can't be early_filtered.
-    // 3. return the count of preserved the indexes in `selection`
-    fn early_filtering_probe_with_selection(
+    // Perform early filtering probe, store matched indexes in `matched_selection` and store unmatched indexes
+    // in `unmatched_selection`, return the number of matched and unmatched indexes.
+    fn early_filtering_probe(
+        &self,
+        hashes: &mut [u64],
+        valids: Option<Bitmap>,
+        matched_selection: &mut [u32],
+        unmatched_selection: &mut [u32],
+    ) -> (usize, usize);
+
+    // Perform early filtering probe and store matched indexes in `selection`, return the number of matched indexes.
+    fn early_filtering_matched_probe(
         &self,
         hashes: &mut [u64],
         valids: Option<Bitmap>,
@@ -534,7 +538,7 @@ pub trait HashJoinHashtableLike {
     // the ptr is the link header.
     fn next_contains(&self, key: &Self::Key, ptr: u64) -> bool;
 
-    /// 1. `key` is the serialize build key from one row
+    /// 1. `key` is the serialize probe key from one row
     /// 2. `ptr` pointers to the *RawEntry for of the bucket correlated to key.So before this method,
     /// we will do a round probe firstly. If the ptr is zero, it means there is no correlated bucket
     /// for key
diff --git a/src/common/storage/src/statistics.rs b/src/common/storage/src/statistics.rs
index 0debe1ed8930..5907a5a58a04 100644
--- a/src/common/storage/src/statistics.rs
+++ b/src/common/storage/src/statistics.rs
@@ -50,7 +50,7 @@ impl Datum {
             }
             Scalar::Number(NumberScalar::Float64(v)) => Some(Datum::Float(v)),
             Scalar::Binary(v) => Some(Datum::Bytes(v)),
-            Scalar::String(v) => Some(Datum::Bytes(v)),
+            Scalar::String(v) => Some(Datum::Bytes(v.as_bytes().to_vec())),
             _ => None,
         }
     }
diff --git a/src/query/ast/Cargo.toml b/src/query/ast/Cargo.toml
index 4121907b9166..183004b44f70 100644
--- a/src/query/ast/Cargo.toml
+++ b/src/query/ast/Cargo.toml
@@ -13,6 +13,7 @@ doctest = false
 # Workspace dependencies
 databend-common-base = { path = "../../common/base" }
 databend-common-exception = { path = "../../common/exception" }
+databend-common-functions = { path = "../functions" }
 databend-common-io = { path = "../../common/io" }
 databend-common-meta-app = { path = "../../meta/app" }
 
diff --git a/src/query/ast/src/ast/expr.rs b/src/query/ast/src/ast/expr.rs
index a0b19450086a..3522ba10ab29 100644
--- a/src/query/ast/src/ast/expr.rs
+++ b/src/query/ast/src/ast/expr.rs
@@ -18,6 +18,7 @@ use std::fmt::Formatter;
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_exception::Span;
+use databend_common_functions::aggregates::AggregateFunctionFactory;
 use databend_common_io::display_decimal_256;
 use databend_common_io::escape_string_with_quote;
 use enum_as_inner::EnumAsInner;
@@ -1340,3 +1341,90 @@ pub fn split_equivalent_predicate_expr(expr: &Expr) -> Option<(Expr, Expr)> {
         _ => None,
     }
 }
+
+// If contain agg function in Expr
+pub fn contain_agg_func(expr: &Expr) -> bool {
+    match expr {
+        Expr::ColumnRef { .. } => false,
+        Expr::IsNull { expr, .. } => contain_agg_func(expr),
+        Expr::IsDistinctFrom { left, right, .. } => {
+            contain_agg_func(left) || contain_agg_func(right)
+        }
+        Expr::InList { expr, list, .. } => {
+            contain_agg_func(expr) || list.iter().any(contain_agg_func)
+        }
+        Expr::InSubquery { expr, .. } => contain_agg_func(expr),
+        Expr::Between {
+            expr, low, high, ..
+        } => contain_agg_func(expr) || contain_agg_func(low) || contain_agg_func(high),
+        Expr::BinaryOp { left, right, .. } => contain_agg_func(left) || contain_agg_func(right),
+        Expr::JsonOp { left, right, .. } => contain_agg_func(left) || contain_agg_func(right),
+        Expr::UnaryOp { expr, .. } => contain_agg_func(expr),
+        Expr::Cast { expr, .. } => contain_agg_func(expr),
+        Expr::TryCast { expr, .. } => contain_agg_func(expr),
+        Expr::Extract { expr, .. } => contain_agg_func(expr),
+        Expr::DatePart { expr, .. } => contain_agg_func(expr),
+        Expr::Position {
+            substr_expr,
+            str_expr,
+            ..
+        } => contain_agg_func(substr_expr) || contain_agg_func(str_expr),
+        Expr::Substring {
+            expr,
+            substring_for,
+            substring_from,
+            ..
+        } => {
+            if let Some(substring_for) = substring_for {
+                contain_agg_func(expr) || contain_agg_func(substring_for)
+            } else {
+                contain_agg_func(expr) || contain_agg_func(substring_from)
+            }
+        }
+        Expr::Trim { expr, .. } => contain_agg_func(expr),
+        Expr::Literal { .. } => false,
+        Expr::CountAll { .. } => false,
+        Expr::Tuple { exprs, .. } => exprs.iter().any(contain_agg_func),
+        Expr::FunctionCall { name, .. } => {
+            AggregateFunctionFactory::instance().contains(name.to_string())
+        }
+        Expr::Case {
+            operand,
+            conditions,
+            results,
+            else_result,
+            ..
+        } => {
+            if let Some(operand) = operand {
+                if contain_agg_func(operand) {
+                    return true;
+                }
+            }
+            if conditions.iter().any(contain_agg_func) {
+                return true;
+            }
+            if results.iter().any(contain_agg_func) {
+                return true;
+            }
+            if let Some(else_result) = else_result {
+                if contain_agg_func(else_result) {
+                    return true;
+                }
+            }
+            false
+        }
+        Expr::Exists { .. } => false,
+        Expr::Subquery { .. } => false,
+        Expr::MapAccess { expr, .. } => contain_agg_func(expr),
+        Expr::Array { exprs, .. } => exprs.iter().any(contain_agg_func),
+        Expr::Map { kvs, .. } => kvs.iter().any(|(_, v)| contain_agg_func(v)),
+        Expr::Interval { expr, .. } => contain_agg_func(expr),
+        Expr::DateAdd { interval, date, .. } => {
+            contain_agg_func(interval) || contain_agg_func(date)
+        }
+        Expr::DateSub { interval, date, .. } => {
+            contain_agg_func(interval) || contain_agg_func(date)
+        }
+        Expr::DateTrunc { date, .. } => contain_agg_func(date),
+    }
+}
diff --git a/src/query/ast/src/ast/statements/task.rs b/src/query/ast/src/ast/statements/task.rs
index b101c4b31925..9a164eec4919 100644
--- a/src/query/ast/src/ast/statements/task.rs
+++ b/src/query/ast/src/ast/statements/task.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::BTreeMap;
 use std::fmt::Display;
 use std::fmt::Formatter;
 
@@ -23,6 +24,7 @@ pub struct CreateTaskStmt {
     pub name: String,
     pub warehouse_opts: WarehouseOptions,
     pub schedule_opts: Option<ScheduleOptions>,
+    pub session_parameters: BTreeMap<String, String>,
     pub suspend_task_after_num_failures: Option<u64>,
     pub comments: String,
     pub after: Vec<String>,
@@ -43,8 +45,14 @@ impl Display for CreateTaskStmt {
             write!(f, "{}", schedule_opt)?;
         }
 
+        if !self.session_parameters.is_empty() {
+            for (key, value) in &self.session_parameters {
+                write!(f, " {} = '{}'", key, value)?;
+            }
+        }
+
         if let Some(num) = self.suspend_task_after_num_failures {
-            write!(f, " SUSPEND TASK AFTER {} FAILURES", num)?;
+            write!(f, " SUSPEND_TASK_AFTER {} FAILURES", num)?;
         }
 
         if !self.comments.is_empty() {
@@ -117,6 +125,7 @@ pub enum AlterTaskOptions {
         schedule: Option<ScheduleOptions>,
         suspend_task_after_num_failures: Option<u64>,
         comments: Option<String>,
+        session_parameters: Option<BTreeMap<String, String>>,
     },
     Unset {
         warehouse: bool,
@@ -137,6 +146,7 @@ impl Display for AlterTaskOptions {
                 warehouse,
                 schedule,
                 suspend_task_after_num_failures,
+                session_parameters,
                 comments,
             } => {
                 if let Some(wh) = warehouse {
@@ -151,6 +161,11 @@ impl Display for AlterTaskOptions {
                 if let Some(comments) = comments {
                     write!(f, " COMMENTS = '{}'", comments)?;
                 }
+                if let Some(session) = session_parameters {
+                    for (key, value) in session {
+                        write!(f, " {} = '{}'", key, value)?;
+                    }
+                }
                 Ok(())
             }
             AlterTaskOptions::Unset { warehouse } => {
diff --git a/src/query/ast/src/parser/expr.rs b/src/query/ast/src/parser/expr.rs
index 5cbd99193bd7..175b92c0785b 100644
--- a/src/query/ast/src/parser/expr.rs
+++ b/src/query/ast/src/parser/expr.rs
@@ -1574,7 +1574,7 @@ pub fn type_name(i: Input) -> IResult<TypeName> {
     );
     let ty_binary = value(
         TypeName::Binary,
-        rule! { ( BINARY | VARBINARY ) ~ ( "(" ~ ^#literal_u64 ~ ^")" )? },
+        rule! { ( BINARY | VARBINARY | LONGBLOB | MEDIUMBLOB |  TINYBLOB| BLOB ) ~ ( "(" ~ ^#literal_u64 ~ ^")" )? },
     );
     let ty_string = value(
         TypeName::String,
diff --git a/src/query/ast/src/parser/statement.rs b/src/query/ast/src/parser/statement.rs
index 4d3186e6dd0f..9b1ca0a9bdae 100644
--- a/src/query/ast/src/parser/statement.rs
+++ b/src/query/ast/src/parser/statement.rs
@@ -107,12 +107,14 @@ pub fn statement(i: Input) -> IResult<StatementWithFormat> {
     let create_task = map(
         rule! {
             CREATE ~ TASK ~ ( IF ~ ^NOT ~ ^EXISTS )?
-            ~ #ident ~ #task_warehouse_option
+            ~ #ident
+            ~ #task_warehouse_option
             ~ (SCHEDULE ~ "=" ~ #task_schedule_option)?
             ~ (AFTER ~ #comma_separated_list0(literal_string))?
             ~ (WHEN ~ #expr )?
             ~ (SUSPEND_TASK_AFTER_NUM_FAILURES ~ "=" ~ #literal_u64)?
             ~ ( (COMMENT | COMMENTS) ~ ^"=" ~ ^#literal_string )?
+            ~ (#set_table_option)?
             ~ AS ~ #statement
         },
         |(
@@ -126,11 +128,12 @@ pub fn statement(i: Input) -> IResult<StatementWithFormat> {
             when_conditions,
             suspend_opt,
             comment_opt,
+            session_opts,
             _,
             sql,
         )| {
             let sql = format!("{}", sql.stmt);
-
+            let session_opts = session_opts.unwrap_or_default();
             Statement::CreateTask(CreateTaskStmt {
                 if_not_exists: opt_if_not_exists.is_some(),
                 name: task.to_string(),
@@ -144,6 +147,7 @@ pub fn statement(i: Input) -> IResult<StatementWithFormat> {
                 },
                 when_condition: when_conditions.map(|(_, cond)| cond.to_string()),
                 sql,
+                session_parameters: session_opts,
             })
         },
     );
@@ -2956,12 +2960,16 @@ pub fn alter_task_option(i: Input) -> IResult<AlterTaskOptions> {
              ~ ( SCHEDULE ~ "=" ~ #task_schedule_option )?
              ~ ( SUSPEND_TASK_AFTER_NUM_FAILURES ~ "=" ~ #literal_u64 )?
              ~ ( COMMENT ~ "=" ~ #literal_string )?
-        },
-        |(_, warehouse_opts, schedule_opts, suspend_opts, comment)| AlterTaskOptions::Set {
-            warehouse: warehouse_opts.map(|(_, _, warehouse)| warehouse),
-            schedule: schedule_opts.map(|(_, _, schedule)| schedule),
-            suspend_task_after_num_failures: suspend_opts.map(|(_, _, num)| num),
-            comments: comment.map(|(_, _, comment)| comment),
+             ~ (#set_table_option)?
+        },
+        |(_, warehouse_opts, schedule_opts, suspend_opts, comment, session_opts)| {
+            AlterTaskOptions::Set {
+                warehouse: warehouse_opts.map(|(_, _, warehouse)| warehouse),
+                schedule: schedule_opts.map(|(_, _, schedule)| schedule),
+                suspend_task_after_num_failures: suspend_opts.map(|(_, _, num)| num),
+                comments: comment.map(|(_, _, comment)| comment),
+                session_parameters: session_opts,
+            }
         },
     );
     let unset = map(
diff --git a/src/query/ast/src/parser/token.rs b/src/query/ast/src/parser/token.rs
index 4ee536190e24..945f7b343ef8 100644
--- a/src/query/ast/src/parser/token.rs
+++ b/src/query/ast/src/parser/token.rs
@@ -364,6 +364,14 @@ pub enum TokenKind {
     BIGINT,
     #[token("BINARY", ignore(ascii_case))]
     BINARY,
+    #[token("LONGBLOB", ignore(ascii_case))]
+    LONGBLOB,
+    #[token("MEDIUMBLOB", ignore(ascii_case))]
+    MEDIUMBLOB,
+    #[token("TINYBLOB", ignore(ascii_case))]
+    TINYBLOB,
+    #[token("BLOB", ignore(ascii_case))]
+    BLOB,
     #[token("BINARY_FORMAT", ignore(ascii_case))]
     BINARY_FORMAT,
     #[token("BITMAP", ignore(ascii_case))]
@@ -971,6 +979,12 @@ pub enum TokenKind {
     TABLES,
     #[token("TEXT", ignore(ascii_case))]
     TEXT,
+    #[token("LONGTEXT", ignore(ascii_case))]
+    LONGTEXT,
+    #[token("MEDIUMTEXT", ignore(ascii_case))]
+    MEDIUMTEXT,
+    #[token("TINYTEXT", ignore(ascii_case))]
+    TINYTEXT,
     #[token("TENANTSETTING", ignore(ascii_case))]
     TENANTSETTING,
     #[token("TENANTS", ignore(ascii_case))]
diff --git a/src/query/ast/tests/it/parser.rs b/src/query/ast/tests/it/parser.rs
index 2b0b837796af..d9636a64dc54 100644
--- a/src/query/ast/tests/it/parser.rs
+++ b/src/query/ast/tests/it/parser.rs
@@ -507,13 +507,14 @@ fn test_statement() {
         r#"CREATE NETWORK POLICY mypolicy ALLOWED_IP_LIST=('192.168.10.0/24') BLOCKED_IP_LIST=('192.168.10.99') COMMENT='test'"#,
         r#"ALTER NETWORK POLICY mypolicy SET ALLOWED_IP_LIST=('192.168.10.0/24','192.168.255.1') BLOCKED_IP_LIST=('192.168.1.99') COMMENT='test'"#,
         // tasks
-        r#"CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 15 MINUTE SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' AS SELECT * FROM MyTable1"#,
+        r#"CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 15 MINUTE SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' DATABASE = 'target', TIMEZONE = 'America/Los Angeles' AS SELECT * FROM MyTable1"#,
         r#"CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 15 SECOND SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' AS SELECT * FROM MyTable1"#,
         r#"CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 1215 SECOND SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' AS SELECT * FROM MyTable1"#,
         r#"CREATE TASK IF NOT EXISTS MyTask1 SCHEDULE = USING CRON '0 6 * * *' 'America/Los_Angeles' COMMENT = 'serverless + cron' AS insert into t (c1, c2) values (1, 2), (3, 4)"#,
         r#"CREATE TASK IF NOT EXISTS MyTask1 SCHEDULE = USING CRON '0 12 * * *' AS copy into streams_test.paper_table from @stream_stage FILE_FORMAT = (TYPE = PARQUET) PURGE=true"#,
         r#"CREATE TASK IF NOT EXISTS MyTask1 SCHEDULE = USING CRON '0 13 * * *' AS COPY INTO @my_internal_stage FROM canadian_city_population FILE_FORMAT = (TYPE = PARQUET)"#,
         r#"CREATE TASK IF NOT EXISTS MyTask1 AFTER 'task2', 'task3' WHEN SYSTEM$GET_PREDECESSOR_RETURN_VALUE('task_name') != 'VALIDATION' AS VACUUM TABLE t"#,
+        r#"CREATE TASK IF NOT EXISTS MyTask1 DATABASE = 'target', TIMEZONE = 'America/Los Angeles'  AS VACUUM TABLE t"#,
         r#"ALTER TASK MyTask1 RESUME"#,
         r#"ALTER TASK MyTask1 SUSPEND"#,
         r#"ALTER TASK MyTask1 ADD AFTER 'task2', 'task3'"#,
@@ -521,6 +522,7 @@ fn test_statement() {
         r#"ALTER TASK MyTask1 SET WAREHOUSE= 'MyWarehouse' SCHEDULE = USING CRON '0 6 * * *' 'America/Los_Angeles' COMMENT = 'serverless + cron'"#,
         r#"ALTER TASK MyTask1 SET WAREHOUSE= 'MyWarehouse' SCHEDULE = 13 MINUTE SUSPEND_TASK_AFTER_NUM_FAILURES = 10 COMMENT = 'serverless + cron'"#,
         r#"ALTER TASK MyTask1 SET WAREHOUSE= 'MyWarehouse' SCHEDULE = 5 SECOND SUSPEND_TASK_AFTER_NUM_FAILURES = 10 COMMENT = 'serverless + cron'"#,
+        r#"ALTER TASK MyTask1 SET DATABASE='newDB', TIMEZONE='America/Los_Angeles'"#,
         r#"ALTER TASK MyTask2 MODIFY AS SELECT CURRENT_VERSION()"#,
         r#"ALTER TASK MyTask1 MODIFY WHEN SYSTEM$GET_PREDECESSOR_RETURN_VALUE('task_name') != 'VALIDATION'"#,
         r#"DROP TASK MyTask1"#,
diff --git a/src/query/ast/tests/it/testdata/expr-error.txt b/src/query/ast/tests/it/testdata/expr-error.txt
index 0cdfbf5f9d5b..dda8730e2edd 100644
--- a/src/query/ast/tests/it/testdata/expr-error.txt
+++ b/src/query/ast/tests/it/testdata/expr-error.txt
@@ -30,7 +30,7 @@ error:
   --> SQL:1:14
   |
 1 | CAST(col1 AS foo)
-  | ----         ^^^ unexpected `foo`, expecting `BOOL`, `FLOAT`, `BOOLEAN`, `FLOAT32`, `FLOAT64`, `JSON`, `DOUBLE`, `UINT8`, `TINYINT`, `UINT16`, `SMALLINT`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `DECIMAL`, `ARRAY`, `MAP`, `BITMAP`, `TUPLE`, `DATE`, `DATETIME`, `TIMESTAMP`, `BINARY`, `VARBINARY`, `STRING`, `VARCHAR`, `CHAR`, `CHARACTER`, `TEXT`, `VARIANT`, or `NULLABLE`
+  | ----         ^^^ unexpected `foo`, expecting `BOOL`, `FLOAT`, `BOOLEAN`, `FLOAT32`, `FLOAT64`, `BLOB`, `JSON`, `DOUBLE`, `LONGBLOB`, `UINT8`, `TINYINT`, `UINT16`, `SMALLINT`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `DECIMAL`, `ARRAY`, `MAP`, `BITMAP`, `TUPLE`, `DATE`, `DATETIME`, `TIMESTAMP`, `BINARY`, `VARBINARY`, `MEDIUMBLOB`, `TINYBLOB`, `STRING`, `VARCHAR`, `CHAR`, `CHARACTER`, `TEXT`, `VARIANT`, or `NULLABLE`
   | |             
   | while parsing `CAST(... AS ...)`
   | while parsing expression
diff --git a/src/query/ast/tests/it/testdata/statement-error.txt b/src/query/ast/tests/it/testdata/statement-error.txt
index e7292e72492c..cde27fda1b6b 100644
--- a/src/query/ast/tests/it/testdata/statement-error.txt
+++ b/src/query/ast/tests/it/testdata/statement-error.txt
@@ -29,7 +29,7 @@ error:
   --> SQL:1:19
   |
 1 | create table a (c varch)
-  | ------          - ^^^^^ unexpected `varch`, expecting `VARCHAR`, `CHAR`, `VARIANT`, `CHARACTER`, `VARBINARY`, `ARRAY`, `BINARY`, `MAP`, `DATE`, `STRING`, `FLOAT32`, `FLOAT64`, `DECIMAL`, `SMALLINT`, `DATETIME`, `NULLABLE`, `BOOLEAN`, `BOOL`, `UINT8`, `TINYINT`, `UINT16`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `FLOAT`, `DOUBLE`, `BITMAP`, `TUPLE`, `TIMESTAMP`, `TEXT`, or `JSON`
+  | ------          - ^^^^^ unexpected `varch`, expecting `VARCHAR`, `CHAR`, `VARIANT`, `CHARACTER`, `VARBINARY`, `ARRAY`, `BINARY`, `MAP`, `DATE`, `STRING`, `FLOAT32`, `FLOAT64`, `DECIMAL`, `SMALLINT`, `DATETIME`, `NULLABLE`, `BOOLEAN`, `BOOL`, `UINT8`, `TINYINT`, `UINT16`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `FLOAT`, `DOUBLE`, `BITMAP`, `TUPLE`, `TIMESTAMP`, `LONGBLOB`, `MEDIUMBLOB`, `TINYBLOB`, `BLOB`, `TEXT`, or `JSON`
   | |               |  
   | |               while parsing `<column name> <type> [DEFAULT <expr>] [AS (<expr>) VIRTUAL] [AS (<expr>) STORED] [COMMENT '<comment>']`
   | while parsing `CREATE TABLE [IF NOT EXISTS] [<database>.]<table> [<source>] [<table_options>]`
@@ -42,7 +42,7 @@ error:
   --> SQL:1:25
   |
 1 | create table a (c tuple())
-  | ------          - ----- ^ unexpected `)`, expecting `BOOLEAN`, `BOOL`, `UINT8`, `TINYINT`, `UINT16`, `SMALLINT`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `FLOAT32`, `FLOAT`, `FLOAT64`, `DOUBLE`, `DECIMAL`, `ARRAY`, `MAP`, `BITMAP`, `TUPLE`, `DATE`, `DATETIME`, `TIMESTAMP`, `BINARY`, `VARBINARY`, `STRING`, `VARCHAR`, `CHAR`, `CHARACTER`, `TEXT`, `VARIANT`, `JSON`, `NULLABLE`, <Ident>, or <QuotedString>
+  | ------          - ----- ^ unexpected `)`, expecting `BOOLEAN`, `BOOL`, `UINT8`, `TINYINT`, `UINT16`, `SMALLINT`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `FLOAT32`, `FLOAT`, `FLOAT64`, `DOUBLE`, `DECIMAL`, `ARRAY`, `MAP`, `BITMAP`, `TUPLE`, `DATE`, `DATETIME`, `TIMESTAMP`, `BINARY`, `VARBINARY`, `LONGBLOB`, `MEDIUMBLOB`, `TINYBLOB`, `BLOB`, `STRING`, `VARCHAR`, `CHAR`, `CHARACTER`, `TEXT`, `VARIANT`, `JSON`, `NULLABLE`, <Ident>, or <QuotedString>
   | |               | |      
   | |               | while parsing type name
   | |               while parsing `<column name> <type> [DEFAULT <expr>] [AS (<expr>) VIRTUAL] [AS (<expr>) STORED] [COMMENT '<comment>']`
@@ -70,7 +70,7 @@ error:
   --> SQL:1:38
   |
 1 | create table a (b tuple(c int, uint64));
-  | ------          - -----              ^ unexpected `)`, expecting `BOOLEAN`, `BOOL`, `UINT8`, `TINYINT`, `UINT16`, `SMALLINT`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `FLOAT32`, `FLOAT`, `FLOAT64`, `DOUBLE`, `DECIMAL`, `ARRAY`, `MAP`, `BITMAP`, `TUPLE`, `DATE`, `DATETIME`, `TIMESTAMP`, `BINARY`, `VARBINARY`, `STRING`, `VARCHAR`, `CHAR`, `CHARACTER`, `TEXT`, `VARIANT`, `JSON`, or `NULLABLE`
+  | ------          - -----              ^ unexpected `)`, expecting `BOOLEAN`, `BOOL`, `UINT8`, `TINYINT`, `UINT16`, `SMALLINT`, `UINT32`, `INT`, `INTEGER`, `UINT64`, `UNSIGNED`, `BIGINT`, `INT8`, `INT16`, `INT32`, `INT64`, `SIGNED`, `FLOAT32`, `FLOAT`, `FLOAT64`, `DOUBLE`, `DECIMAL`, `ARRAY`, `MAP`, `BITMAP`, `TUPLE`, `DATE`, `DATETIME`, `TIMESTAMP`, `BINARY`, `VARBINARY`, `LONGBLOB`, `MEDIUMBLOB`, `TINYBLOB`, `BLOB`, `STRING`, `VARCHAR`, `CHAR`, `CHARACTER`, `TEXT`, `VARIANT`, `JSON`, or `NULLABLE`
   | |               | |                   
   | |               | while parsing TUPLE(<name> <type>, ...)
   | |               | while parsing type name
diff --git a/src/query/ast/tests/it/testdata/statement.txt b/src/query/ast/tests/it/testdata/statement.txt
index 966fd97396b8..53f13cd4fbd9 100644
--- a/src/query/ast/tests/it/testdata/statement.txt
+++ b/src/query/ast/tests/it/testdata/statement.txt
@@ -13856,9 +13856,9 @@ AlterNetworkPolicy(
 
 
 ---------- Input ----------
-CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 15 MINUTE SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' AS SELECT * FROM MyTable1
+CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 15 MINUTE SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' DATABASE = 'target', TIMEZONE = 'America/Los Angeles' AS SELECT * FROM MyTable1
 ---------- Output ---------
-CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = MyWarehouse SCHEDULE 900 SECOND SUSPEND TASK AFTER 3 FAILURES COMMENTS = 'This is test task 1' AS SELECT * FROM MyTable1
+CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = MyWarehouse SCHEDULE 900 SECOND database = 'target' timezone = 'America/Los Angeles' SUSPEND_TASK_AFTER 3 FAILURES COMMENTS = 'This is test task 1' AS SELECT * FROM MyTable1
 ---------- AST ------------
 CreateTask(
     CreateTaskStmt {
@@ -13874,6 +13874,10 @@ CreateTask(
                 900,
             ),
         ),
+        session_parameters: {
+            "database": "target",
+            "timezone": "America/Los Angeles",
+        },
         suspend_task_after_num_failures: Some(
             3,
         ),
@@ -13888,7 +13892,7 @@ CreateTask(
 ---------- Input ----------
 CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 15 SECOND SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' AS SELECT * FROM MyTable1
 ---------- Output ---------
-CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = MyWarehouse SCHEDULE 15 SECOND SUSPEND TASK AFTER 3 FAILURES COMMENTS = 'This is test task 1' AS SELECT * FROM MyTable1
+CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = MyWarehouse SCHEDULE 15 SECOND SUSPEND_TASK_AFTER 3 FAILURES COMMENTS = 'This is test task 1' AS SELECT * FROM MyTable1
 ---------- AST ------------
 CreateTask(
     CreateTaskStmt {
@@ -13904,6 +13908,7 @@ CreateTask(
                 15,
             ),
         ),
+        session_parameters: {},
         suspend_task_after_num_failures: Some(
             3,
         ),
@@ -13918,7 +13923,7 @@ CreateTask(
 ---------- Input ----------
 CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = 'MyWarehouse' SCHEDULE = 1215 SECOND SUSPEND_TASK_AFTER_NUM_FAILURES = 3 COMMENT = 'This is test task 1' AS SELECT * FROM MyTable1
 ---------- Output ---------
-CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = MyWarehouse SCHEDULE 1215 SECOND SUSPEND TASK AFTER 3 FAILURES COMMENTS = 'This is test task 1' AS SELECT * FROM MyTable1
+CREATE TASK IF NOT EXISTS MyTask1 WAREHOUSE = MyWarehouse SCHEDULE 1215 SECOND SUSPEND_TASK_AFTER 3 FAILURES COMMENTS = 'This is test task 1' AS SELECT * FROM MyTable1
 ---------- AST ------------
 CreateTask(
     CreateTaskStmt {
@@ -13934,6 +13939,7 @@ CreateTask(
                 1215,
             ),
         ),
+        session_parameters: {},
         suspend_task_after_num_failures: Some(
             3,
         ),
@@ -13965,6 +13971,7 @@ CreateTask(
                 ),
             ),
         ),
+        session_parameters: {},
         suspend_task_after_num_failures: None,
         comments: "serverless + cron",
         after: [],
@@ -13992,6 +13999,7 @@ CreateTask(
                 None,
             ),
         ),
+        session_parameters: {},
         suspend_task_after_num_failures: None,
         comments: "",
         after: [],
@@ -14019,6 +14027,7 @@ CreateTask(
                 None,
             ),
         ),
+        session_parameters: {},
         suspend_task_after_num_failures: None,
         comments: "",
         after: [],
@@ -14041,6 +14050,7 @@ CreateTask(
             warehouse: None,
         },
         schedule_opts: None,
+        session_parameters: {},
         suspend_task_after_num_failures: None,
         comments: "",
         after: [
@@ -14055,6 +14065,32 @@ CreateTask(
 )
 
 
+---------- Input ----------
+CREATE TASK IF NOT EXISTS MyTask1 DATABASE = 'target', TIMEZONE = 'America/Los Angeles'  AS VACUUM TABLE t
+---------- Output ---------
+CREATE TASK IF NOT EXISTS MyTask1 database = 'target' timezone = 'America/Los Angeles' AS VACUUM TABLE t 
+---------- AST ------------
+CreateTask(
+    CreateTaskStmt {
+        if_not_exists: true,
+        name: "MyTask1",
+        warehouse_opts: WarehouseOptions {
+            warehouse: None,
+        },
+        schedule_opts: None,
+        session_parameters: {
+            "database": "target",
+            "timezone": "America/Los Angeles",
+        },
+        suspend_task_after_num_failures: None,
+        comments: "",
+        after: [],
+        when_condition: None,
+        sql: "VACUUM TABLE t ",
+    },
+)
+
+
 ---------- Input ----------
 ALTER TASK MyTask1 RESUME
 ---------- Output ---------
@@ -14145,6 +14181,7 @@ AlterTask(
             comments: Some(
                 "serverless + cron",
             ),
+            session_parameters: None,
         },
     },
 )
@@ -14174,6 +14211,7 @@ AlterTask(
             comments: Some(
                 "serverless + cron",
             ),
+            session_parameters: None,
         },
     },
 )
@@ -14203,6 +14241,32 @@ AlterTask(
             comments: Some(
                 "serverless + cron",
             ),
+            session_parameters: None,
+        },
+    },
+)
+
+
+---------- Input ----------
+ALTER TASK MyTask1 SET DATABASE='newDB', TIMEZONE='America/Los_Angeles'
+---------- Output ---------
+ALTER TASK MyTask1 database = 'newDB' timezone = 'America/Los_Angeles'
+---------- AST ------------
+AlterTask(
+    AlterTaskStmt {
+        if_exists: false,
+        name: "MyTask1",
+        options: Set {
+            warehouse: None,
+            schedule: None,
+            suspend_task_after_num_failures: None,
+            comments: None,
+            session_parameters: Some(
+                {
+                    "database": "newDB",
+                    "timezone": "America/Los_Angeles",
+                },
+            ),
         },
     },
 )
diff --git a/src/query/catalog/src/plan/internal_column.rs b/src/query/catalog/src/plan/internal_column.rs
index 480c20b896d8..c473990ece51 100644
--- a/src/query/catalog/src/plan/internal_column.rs
+++ b/src/query/catalog/src/plan/internal_column.rs
@@ -254,12 +254,12 @@ impl InternalColumn {
                 let mut row_ids = Vec::with_capacity(num_rows);
                 if let Some(offsets) = &meta.offsets {
                     for i in offsets {
-                        let row_id = format!("{}{:06x}", uuid, *i).as_bytes().to_vec();
+                        let row_id = format!("{}{:06x}", uuid, *i);
                         row_ids.push(row_id);
                     }
                 } else {
                     for i in 0..num_rows {
-                        let row_id = format!("{}{:06x}", uuid, i).as_bytes().to_vec();
+                        let row_id = format!("{}{:06x}", uuid, i);
                         row_ids.push(row_id);
                     }
                 }
diff --git a/src/query/catalog/src/query_kind.rs b/src/query/catalog/src/query_kind.rs
index b2e27cd799d5..74e2aca44e84 100644
--- a/src/query/catalog/src/query_kind.rs
+++ b/src/query/catalog/src/query_kind.rs
@@ -24,6 +24,7 @@ pub enum QueryKind {
     Query,
     Explain,
     CopyIntoTable,
+    CopyIntoLocation,
     Update,
     Insert,
     Other,
diff --git a/src/query/catalog/src/table_args.rs b/src/query/catalog/src/table_args.rs
index 0397e6f65827..6878605e80f6 100644
--- a/src/query/catalog/src/table_args.rs
+++ b/src/query/catalog/src/table_args.rs
@@ -79,7 +79,7 @@ impl TableArgs {
                     .into_string()
                     .map_err(|_| ErrorCode::BadArguments("Expected string argument"))?;
 
-                Ok(String::from_utf8(arg)?)
+                Ok(arg)
             })
             .collect::<Result<Vec<_>>>()
     }
diff --git a/src/query/expression/src/aggregate/payload_row.rs b/src/query/expression/src/aggregate/payload_row.rs
index 54187c34a369..97ded71a4d58 100644
--- a/src/query/expression/src/aggregate/payload_row.rs
+++ b/src/query/expression/src/aggregate/payload_row.rs
@@ -108,7 +108,7 @@ pub unsafe fn serialize_column_to_rowformat(
         Column::String(v) => {
             for i in 0..rows {
                 let index = select_index.get_index(i);
-                let data = arena.alloc_slice_copy(v.index_unchecked(index));
+                let data = arena.alloc_slice_copy(v.index_unchecked(index).as_bytes());
 
                 store(&(data.len() as u32), address[index].add(offset) as *mut u8);
 
diff --git a/src/query/expression/src/converts/arrow2/from.rs b/src/query/expression/src/converts/arrow2/from.rs
index c51fe0bf4036..03f4e818a910 100644
--- a/src/query/expression/src/converts/arrow2/from.rs
+++ b/src/query/expression/src/converts/arrow2/from.rs
@@ -29,7 +29,6 @@ use crate::types::array::ArrayColumn;
 use crate::types::binary::BinaryColumn;
 use crate::types::decimal::DecimalColumn;
 use crate::types::nullable::NullableColumn;
-use crate::types::string::CheckUTF8;
 use crate::types::string::StringColumn;
 use crate::types::DataType;
 use crate::types::DecimalDataType;
@@ -423,7 +422,6 @@ impl Column {
                         .map(|x| *x as u64)
                         .collect::<Vec<_>>();
                     let column = StringColumn::new(arrow_col.values().clone(), offsets.into());
-                    column.check_utf8()?;
                     Column::String(column)
                 }
                 (DataType::String, ArrowDataType::LargeBinary) => {
@@ -437,7 +435,6 @@ impl Column {
                     let offsets =
                         unsafe { std::mem::transmute::<Buffer<i64>, Buffer<u64>>(offsets) };
                     let column = StringColumn::new(arrow_col.values().clone(), offsets);
-                    column.check_utf8()?;
                     Column::String(column)
                 }
                 (DataType::String, ArrowDataType::FixedSizeBinary(size)) => {
@@ -451,7 +448,6 @@ impl Column {
                         .map(|x| x * (*size) as u64)
                         .collect::<Vec<_>>();
                     let column = StringColumn::new(arrow_col.values().clone(), offsets.into());
-                    column.check_utf8()?;
                     Column::String(column)
                 }
                 (DataType::String, ArrowDataType::Utf8) => {
@@ -467,10 +463,12 @@ impl Column {
                         .iter()
                         .map(|x| *x as u64)
                         .collect::<Vec<_>>();
-                    Column::String(StringColumn::new(
-                        arrow_col.values().clone(),
-                        offsets.into(),
-                    ))
+                    unsafe {
+                        Column::String(StringColumn::new_unchecked(
+                            arrow_col.values().clone(),
+                            offsets.into(),
+                        ))
+                    }
                 }
                 (DataType::String, ArrowDataType::LargeUtf8) => {
                     let arrow_col = arrow_col
@@ -482,7 +480,12 @@ impl Column {
                     let offsets = arrow_col.offsets().clone().into_inner();
                     let offsets =
                         unsafe { std::mem::transmute::<Buffer<i64>, Buffer<u64>>(offsets) };
-                    Column::String(StringColumn::new(arrow_col.values().clone(), offsets))
+                    unsafe {
+                        Column::String(StringColumn::new_unchecked(
+                            arrow_col.values().clone(),
+                            offsets,
+                        ))
+                    }
                 }
                 (DataType::Timestamp, ArrowDataType::Timestamp(uint, _)) => {
                     let values = arrow_col
diff --git a/src/query/expression/src/converts/arrow2/to.rs b/src/query/expression/src/converts/arrow2/to.rs
index 1b77f9d738be..c331d9911718 100644
--- a/src/query/expression/src/converts/arrow2/to.rs
+++ b/src/query/expression/src/converts/arrow2/to.rs
@@ -25,7 +25,6 @@ use super::ARROW_EXT_TYPE_EMPTY_ARRAY;
 use super::ARROW_EXT_TYPE_EMPTY_MAP;
 use super::ARROW_EXT_TYPE_VARIANT;
 use crate::types::decimal::DecimalColumn;
-use crate::types::string::CheckUTF8;
 use crate::types::DecimalDataType;
 use crate::types::NumberColumn;
 use crate::types::NumberDataType;
@@ -301,11 +300,6 @@ impl Column {
                 )
             }
             Column::String(col) => {
-                // todo!("new string")
-                // always check utf8 until we can guarantee the correctness of data in string column
-                // #[cfg(debug_assertions)]
-                col.check_utf8().unwrap();
-
                 let offsets: Buffer<i64> =
                     col.offsets().iter().map(|offset| *offset as i64).collect();
 
diff --git a/src/query/expression/src/converts/datavalues/from.rs b/src/query/expression/src/converts/datavalues/from.rs
index 93a6b13508b8..c04a12af6105 100644
--- a/src/query/expression/src/converts/datavalues/from.rs
+++ b/src/query/expression/src/converts/datavalues/from.rs
@@ -126,7 +126,9 @@ pub fn from_scalar(datavalue: &DataValue, datatype: &DataType) -> Scalar {
         )),
         DataType::Timestamp => Scalar::Timestamp(*datavalue.as_int64().unwrap()),
         DataType::Date => Scalar::Date(*datavalue.as_int64().unwrap() as i32),
-        DataType::String => Scalar::String(datavalue.as_string().unwrap().to_vec()),
+        DataType::String => {
+            Scalar::String(String::from_utf8_lossy(datavalue.as_string().unwrap()).into_owned())
+        }
         DataType::Variant => match datavalue {
             DataValue::String(x) => Scalar::Variant(x.clone()),
             DataValue::Variant(x) => {
diff --git a/src/query/expression/src/converts/datavalues/to.rs b/src/query/expression/src/converts/datavalues/to.rs
index f9809ab201f4..50fd5af94916 100644
--- a/src/query/expression/src/converts/datavalues/to.rs
+++ b/src/query/expression/src/converts/datavalues/to.rs
@@ -39,7 +39,8 @@ pub fn scalar_to_datavalue(scalar: &Scalar) -> DataValue {
         Scalar::Timestamp(x) => DataValue::Int64(*x),
         Scalar::Date(x) => DataValue::Int64(*x as i64),
         Scalar::Boolean(x) => DataValue::Boolean(*x),
-        Scalar::String(x) | Scalar::Variant(x) => DataValue::String(x.clone()),
+        Scalar::Variant(x) => DataValue::String(x.clone()),
+        Scalar::String(x) => DataValue::String(x.as_bytes().to_vec()),
         Scalar::Array(x) => {
             let values = (0..x.len())
                 .map(|idx| scalar_to_datavalue(&x.index(idx).unwrap().to_owned()))
diff --git a/src/query/expression/src/converts/meta/bincode.rs b/src/query/expression/src/converts/meta/bincode.rs
index 76414f527e8c..491c98763424 100644
--- a/src/query/expression/src/converts/meta/bincode.rs
+++ b/src/query/expression/src/converts/meta/bincode.rs
@@ -95,7 +95,7 @@ impl From<LegacyScalar> for Scalar {
             LegacyScalar::Timestamp(ts) => Scalar::Timestamp(ts),
             LegacyScalar::Date(date) => Scalar::Date(date),
             LegacyScalar::Boolean(b) => Scalar::Boolean(b),
-            LegacyScalar::String(s) => Scalar::String(s),
+            LegacyScalar::String(s) => Scalar::String(String::from_utf8_lossy(&s).into_owned()),
             LegacyScalar::Array(col) => Scalar::Array(col.into()),
             LegacyScalar::Map(col) => Scalar::Map(col.into()),
             LegacyScalar::Bitmap(bmp) => Scalar::Bitmap(bmp),
@@ -152,7 +152,7 @@ impl From<Scalar> for LegacyScalar {
             Scalar::Date(date) => LegacyScalar::Date(date),
             Scalar::Boolean(b) => LegacyScalar::Boolean(b),
             Scalar::Binary(_) => unreachable!(),
-            Scalar::String(string) => LegacyScalar::String(string),
+            Scalar::String(string) => LegacyScalar::String(string.as_bytes().to_vec()),
             Scalar::Array(column) => LegacyScalar::Array(column.into()),
             Scalar::Map(column) => LegacyScalar::Map(column.into()),
             Scalar::Bitmap(bitmap) => LegacyScalar::Bitmap(bitmap),
diff --git a/src/query/expression/src/converts/meta/index_scalar.rs b/src/query/expression/src/converts/meta/index_scalar.rs
index c54ddcfe77d2..c4419955d97f 100644
--- a/src/query/expression/src/converts/meta/index_scalar.rs
+++ b/src/query/expression/src/converts/meta/index_scalar.rs
@@ -49,7 +49,7 @@ pub enum IndexScalar {
     Array(IndexColumn),
     Map(IndexColumn),
     Bitmap(Vec<u8>),
-    Tuple(Vec<Scalar>),
+    Tuple(Vec<IndexScalar>),
     Variant(Vec<u8>),
 }
 
@@ -91,11 +91,13 @@ impl From<IndexScalar> for Scalar {
             IndexScalar::Timestamp(ts) => Scalar::Timestamp(ts),
             IndexScalar::Date(date) => Scalar::Date(date),
             IndexScalar::Boolean(b) => Scalar::Boolean(b),
-            IndexScalar::String(s) => Scalar::String(s),
+            IndexScalar::String(s) => Scalar::String(unsafe { String::from_utf8_unchecked(s) }),
             IndexScalar::Array(col) => Scalar::Array(col.into()),
             IndexScalar::Map(col) => Scalar::Map(col.into()),
             IndexScalar::Bitmap(bmp) => Scalar::Bitmap(bmp),
-            IndexScalar::Tuple(tuple) => Scalar::Tuple(tuple),
+            IndexScalar::Tuple(tuple) => {
+                Scalar::Tuple(tuple.into_iter().map(|c| c.into()).collect())
+            }
             IndexScalar::Variant(variant) => Scalar::Variant(variant),
         }
     }
@@ -143,12 +145,14 @@ impl From<Scalar> for IndexScalar {
             Scalar::Timestamp(ts) => IndexScalar::Timestamp(ts),
             Scalar::Date(date) => IndexScalar::Date(date),
             Scalar::Boolean(b) => IndexScalar::Boolean(b),
-            Scalar::String(string) => IndexScalar::String(string),
+            Scalar::String(string) => IndexScalar::String(string.as_bytes().to_vec()),
             Scalar::Binary(s) => IndexScalar::String(s),
             Scalar::Array(column) => IndexScalar::Array(column.into()),
             Scalar::Map(column) => IndexScalar::Map(column.into()),
             Scalar::Bitmap(bitmap) => IndexScalar::Bitmap(bitmap),
-            Scalar::Tuple(tuple) => IndexScalar::Tuple(tuple),
+            Scalar::Tuple(tuple) => {
+                IndexScalar::Tuple(tuple.into_iter().map(|c| c.into()).collect())
+            }
             Scalar::Variant(variant) => IndexScalar::Variant(variant),
             Scalar::EmptyArray | Scalar::EmptyMap => unreachable!(),
         }
diff --git a/src/query/expression/src/kernels/group_by_hash/method_single_string.rs b/src/query/expression/src/kernels/group_by_hash/method_single_string.rs
index 16f7bdc0e186..3d5770dee7a3 100644
--- a/src/query/expression/src/kernels/group_by_hash/method_single_string.rs
+++ b/src/query/expression/src/kernels/group_by_hash/method_single_string.rs
@@ -93,6 +93,8 @@ impl KeyAccessor for BinaryKeyAccessor {
     /// # Safety
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*.
     unsafe fn key_unchecked(&self, index: usize) -> &Self::Key {
+        debug_assert!(index + 1 < self.offsets.len());
+
         &self.data[*self.offsets.get_unchecked(index) as usize
             ..*self.offsets.get_unchecked(index + 1) as usize]
     }
diff --git a/src/query/expression/src/property.rs b/src/query/expression/src/property.rs
index 4df8f598f8af..89fce25a59a1 100644
--- a/src/query/expression/src/property.rs
+++ b/src/query/expression/src/property.rs
@@ -237,12 +237,12 @@ impl Domain {
                 has_true: this.has_true || other.has_true,
             }),
             (Domain::String(this), Domain::String(other)) => Domain::String(StringDomain {
-                min: this.min.as_slice().min(&other.min).to_vec(),
+                min: this.min.as_str().min(&other.min).to_string(),
                 max: this
                     .max
                     .as_ref()
                     .zip(other.max.as_ref())
-                    .map(|(self_max, other_max)| self_max.max(other_max).to_vec()),
+                    .map(|(self_max, other_max)| self_max.max(other_max).to_string()),
             }),
             (Domain::Timestamp(this), Domain::Timestamp(other)) => {
                 Domain::Timestamp(SimpleDomain {
@@ -421,6 +421,7 @@ pub trait SimpleDomainCmp {
     fn domain_gte(&self, other: &Self) -> FunctionDomain<BooleanType>;
     fn domain_lt(&self, other: &Self) -> FunctionDomain<BooleanType>;
     fn domain_lte(&self, other: &Self) -> FunctionDomain<BooleanType>;
+    fn domain_contains(&self, other: &Self) -> FunctionDomain<BooleanType>;
 }
 
 const ALL_TRUE_DOMAIN: BooleanDomain = BooleanDomain {
@@ -489,36 +490,76 @@ impl<T: Ord + PartialOrd> SimpleDomainCmp for SimpleDomain<T> {
             FunctionDomain::Full
         }
     }
+
+    fn domain_contains(&self, other: &Self) -> FunctionDomain<BooleanType> {
+        if self.min > other.max || self.max < other.min {
+            FunctionDomain::Domain(ALL_FALSE_DOMAIN)
+        } else {
+            FunctionDomain::Full
+        }
+    }
 }
 
 impl SimpleDomainCmp for StringDomain {
     fn domain_eq(&self, other: &Self) -> FunctionDomain<BooleanType> {
-        let (d1, d2) = self.unify(other);
+        let (d1, d2) = unify_string(self, other);
         d1.domain_eq(&d2)
     }
 
     fn domain_noteq(&self, other: &Self) -> FunctionDomain<BooleanType> {
-        let (d1, d2) = self.unify(other);
+        let (d1, d2) = unify_string(self, other);
         d1.domain_noteq(&d2)
     }
 
     fn domain_gt(&self, other: &Self) -> FunctionDomain<BooleanType> {
-        let (d1, d2) = self.unify(other);
+        let (d1, d2) = unify_string(self, other);
         d1.domain_gt(&d2)
     }
 
     fn domain_gte(&self, other: &Self) -> FunctionDomain<BooleanType> {
-        let (d1, d2) = self.unify(other);
+        let (d1, d2) = unify_string(self, other);
         d1.domain_gte(&d2)
     }
 
     fn domain_lt(&self, other: &Self) -> FunctionDomain<BooleanType> {
-        let (d1, d2) = self.unify(other);
+        let (d1, d2) = unify_string(self, other);
         d1.domain_lt(&d2)
     }
 
     fn domain_lte(&self, other: &Self) -> FunctionDomain<BooleanType> {
-        let (d1, d2) = self.unify(other);
+        let (d1, d2) = unify_string(self, other);
         d1.domain_lte(&d2)
     }
+
+    fn domain_contains(&self, other: &Self) -> FunctionDomain<BooleanType> {
+        let (d1, d2) = unify_string(self, other);
+        d1.domain_contains(&d2)
+    }
+}
+
+fn unify_string(
+    lhs: &StringDomain,
+    rhs: &StringDomain,
+) -> (SimpleDomain<String>, SimpleDomain<String>) {
+    let mut max = lhs.min.as_str().max(&rhs.min);
+    if let Some(lhs_max) = &lhs.max {
+        max = max.max(lhs_max);
+    }
+    if let Some(rhs_max) = &rhs.max {
+        max = max.max(rhs_max);
+    }
+
+    let mut max = max.to_string();
+    max.push('\0');
+
+    (
+        SimpleDomain {
+            min: lhs.min.clone(),
+            max: lhs.max.clone().unwrap_or_else(|| max.clone()),
+        },
+        SimpleDomain {
+            min: rhs.min.clone(),
+            max: rhs.max.clone().unwrap_or_else(|| max.clone()),
+        },
+    )
 }
diff --git a/src/query/expression/src/row/row_converter.rs b/src/query/expression/src/row/row_converter.rs
index 3c26bcd93c2b..3c4d967ae5c0 100644
--- a/src/query/expression/src/row/row_converter.rs
+++ b/src/query/expression/src/row/row_converter.rs
@@ -156,16 +156,16 @@ impl RowConverter {
                             .iter()
                             .zip(validity.iter())
                             .zip(lengths.iter_mut())
-                            .for_each(|((bytes, v), length)| {
-                                *length += variable::encoded_len(bytes, !v) as u64
+                            .for_each(|((str, v), length)| {
+                                *length += variable::encoded_len(str.as_bytes(), !v) as u64
                             })
                     } else {
                         col.as_string()
                             .unwrap()
                             .iter()
                             .zip(lengths.iter_mut())
-                            .for_each(|(bytes, length)| {
-                                *length += variable::encoded_len(bytes, false) as u64
+                            .for_each(|(str, length)| {
+                                *length += variable::encoded_len(str.as_bytes(), false) as u64
                             })
                     }
                 }
@@ -255,7 +255,13 @@ fn encode_column(out: &mut BinaryColumnBuilder, column: &Column, asc: bool, null
         Column::Timestamp(col) => fixed::encode(out, col, validity, asc, nulls_first),
         Column::Date(col) => fixed::encode(out, col, validity, asc, nulls_first),
         Column::Binary(col) => variable::encode(out, col.iter(), validity, asc, nulls_first),
-        Column::String(col) => variable::encode(out, col.iter(), validity, asc, nulls_first),
+        Column::String(col) => variable::encode(
+            out,
+            col.iter().map(|s| s.as_bytes()),
+            validity,
+            asc,
+            nulls_first,
+        ),
         Column::Variant(col) => variable::encode(out, col.iter(), validity, asc, nulls_first),
         _ => unimplemented!(),
     }
diff --git a/src/query/expression/src/types/binary.rs b/src/query/expression/src/types/binary.rs
index f09aaa9de38a..c78459d1b75f 100644
--- a/src/query/expression/src/types/binary.rs
+++ b/src/query/expression/src/types/binary.rs
@@ -182,6 +182,7 @@ pub struct BinaryColumn {
 impl BinaryColumn {
     pub fn new(data: Buffer<u8>, offsets: Buffer<u64>) -> Self {
         debug_assert!({ offsets.windows(2).all(|w| w[0] <= w[1]) });
+
         BinaryColumn { data, offsets }
     }
 
@@ -216,7 +217,9 @@ impl BinaryColumn {
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
     #[inline]
     pub unsafe fn index_unchecked(&self, index: usize) -> &[u8] {
-        &self.data[(self.offsets[index] as usize)..(self.offsets[index + 1] as usize)]
+        let start = *self.offsets.get_unchecked(index) as usize;
+        let end = *self.offsets.get_unchecked(index + 1) as usize;
+        self.data.get_unchecked(start..end)
     }
 
     pub fn slice(&self, range: Range<usize>) -> Self {
@@ -314,6 +317,8 @@ impl BinaryColumnBuilder {
     }
 
     pub fn from_data(data: Vec<u8>, offsets: Vec<u64>) -> Self {
+        debug_assert!({ offsets.windows(2).all(|w| w[0] <= w[1]) });
+
         BinaryColumnBuilder {
             need_estimated: false,
             data,
@@ -419,6 +424,7 @@ impl BinaryColumnBuilder {
 
     pub fn build_scalar(self) -> Vec<u8> {
         assert_eq!(self.offsets.len(), 2);
+
         self.data[(self.offsets[0] as usize)..(self.offsets[1] as usize)].to_vec()
     }
 
@@ -431,9 +437,10 @@ impl BinaryColumnBuilder {
     ///
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
     pub unsafe fn index_unchecked(&self, row: usize) -> &[u8] {
+        debug_assert!(row + 1 < self.offsets.len());
+
         let start = *self.offsets.get_unchecked(row) as usize;
         let end = *self.offsets.get_unchecked(row + 1) as usize;
-        // soundness: the invariant of the struct
         self.data.get_unchecked(start..end)
     }
 
diff --git a/src/query/expression/src/types/boolean.rs b/src/query/expression/src/types/boolean.rs
index 78785968ecb7..c3c8ed751e46 100644
--- a/src/query/expression/src/types/boolean.rs
+++ b/src/query/expression/src/types/boolean.rs
@@ -110,6 +110,8 @@ impl ValueType for BooleanType {
 
     #[inline(always)]
     unsafe fn index_column_unchecked(col: &Self::Column, index: usize) -> Self::ScalarRef<'_> {
+        debug_assert!(index < col.len());
+
         col.get_bit_unchecked(index)
     }
 
diff --git a/src/query/expression/src/types/date.rs b/src/query/expression/src/types/date.rs
index 92ff4e3f3ddc..12951630afa3 100644
--- a/src/query/expression/src/types/date.rs
+++ b/src/query/expression/src/types/date.rs
@@ -134,6 +134,8 @@ impl ValueType for DateType {
 
     #[inline(always)]
     unsafe fn index_column_unchecked(col: &Self::Column, index: usize) -> Self::ScalarRef<'_> {
+        debug_assert!(index < col.len());
+
         *col.get_unchecked(index)
     }
 
diff --git a/src/query/expression/src/types/decimal.rs b/src/query/expression/src/types/decimal.rs
index 8219ba3cba79..814d6300288a 100644
--- a/src/query/expression/src/types/decimal.rs
+++ b/src/query/expression/src/types/decimal.rs
@@ -120,6 +120,8 @@ impl<Num: Decimal> ValueType for DecimalType<Num> {
 
     #[inline(always)]
     unsafe fn index_column_unchecked(col: &Self::Column, index: usize) -> Self::ScalarRef<'_> {
+        debug_assert!(index < col.len());
+
         *col.get_unchecked(index)
     }
 
@@ -944,6 +946,7 @@ impl DecimalColumn {
     ///
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
     pub unsafe fn index_unchecked(&self, index: usize) -> DecimalScalar {
+        debug_assert!(index < self.len());
         crate::with_decimal_type!(|DECIMAL_TYPE| match self {
             DecimalColumn::DECIMAL_TYPE(col, size) =>
                 DecimalScalar::DECIMAL_TYPE(*col.get_unchecked(index), *size),
diff --git a/src/query/expression/src/types/nullable.rs b/src/query/expression/src/types/nullable.rs
index 2503a6dee784..c6eca90939cd 100755
--- a/src/query/expression/src/types/nullable.rs
+++ b/src/query/expression/src/types/nullable.rs
@@ -253,6 +253,8 @@ impl<T: ValueType> NullableColumn<T> {
     ///
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
     pub unsafe fn index_unchecked(&self, index: usize) -> Option<T::ScalarRef<'_>> {
+        debug_assert!(index < self.validity.len());
+
         match self.validity.get_bit_unchecked(index) {
             true => Some(T::index_column(&self.column, index).unwrap()),
             false => None,
diff --git a/src/query/expression/src/types/number.rs b/src/query/expression/src/types/number.rs
index d10c57fdfe15..dbd794e15849 100644
--- a/src/query/expression/src/types/number.rs
+++ b/src/query/expression/src/types/number.rs
@@ -168,6 +168,8 @@ impl<Num: Number> ValueType for NumberType<Num> {
 
     #[inline(always)]
     unsafe fn index_column_unchecked(col: &Self::Column, index: usize) -> Self::ScalarRef<'_> {
+        debug_assert!(index < col.len());
+
         *col.get_unchecked(index)
     }
 
@@ -580,6 +582,8 @@ impl NumberColumn {
     ///
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
     pub unsafe fn index_unchecked(&self, index: usize) -> NumberScalar {
+        debug_assert!(index < self.len());
+
         crate::with_number_type!(|NUM_TYPE| match self {
             NumberColumn::NUM_TYPE(col) => NumberScalar::NUM_TYPE(*col.get_unchecked(index)),
         })
diff --git a/src/query/expression/src/types/string.rs b/src/query/expression/src/types/string.rs
index 861e5b362e8c..2b3063bd5bf2 100644
--- a/src/query/expression/src/types/string.rs
+++ b/src/query/expression/src/types/string.rs
@@ -25,7 +25,6 @@ use serde::Serialize;
 use super::binary::BinaryColumn;
 use super::binary::BinaryColumnBuilder;
 use super::binary::BinaryIterator;
-use super::SimpleDomain;
 use crate::property::Domain;
 use crate::types::ArgType;
 use crate::types::DataType;
@@ -41,20 +40,20 @@ use crate::ScalarRef;
 pub struct StringType;
 
 impl ValueType for StringType {
-    type Scalar = Vec<u8>;
-    type ScalarRef<'a> = &'a [u8];
+    type Scalar = String;
+    type ScalarRef<'a> = &'a str;
     type Column = StringColumn;
     type Domain = StringDomain;
     type ColumnIterator<'a> = StringIterator<'a>;
     type ColumnBuilder = StringColumnBuilder;
 
     #[inline]
-    fn upcast_gat<'short, 'long: 'short>(long: &'long [u8]) -> &'short [u8] {
+    fn upcast_gat<'short, 'long: 'short>(long: &'long str) -> &'short str {
         long
     }
 
     fn to_owned_scalar(scalar: Self::ScalarRef<'_>) -> Self::Scalar {
-        scalar.to_vec()
+        scalar.to_string()
     }
 
     fn to_scalar_ref(scalar: &Self::Scalar) -> Self::ScalarRef<'_> {
@@ -108,22 +107,12 @@ impl ValueType for StringType {
     }
 
     fn index_column(col: &Self::Column, index: usize) -> Option<Self::ScalarRef<'_>> {
-        let x = col.index(index)?;
-
-        #[cfg(debug_assertions)]
-        x.check_utf8().unwrap();
-
-        Some(x)
+        col.index(index)
     }
 
     #[inline(always)]
     unsafe fn index_column_unchecked(col: &Self::Column, index: usize) -> Self::ScalarRef<'_> {
-        let x = col.index_unchecked(index);
-
-        #[cfg(debug_assertions)]
-        x.check_utf8().unwrap();
-
-        x
+        col.index_unchecked(index)
     }
 
     fn slice_column(col: &Self::Column, range: Range<usize>) -> Self::Column {
@@ -143,7 +132,7 @@ impl ValueType for StringType {
     }
 
     fn push_item(builder: &mut Self::ColumnBuilder, item: Self::ScalarRef<'_>) {
-        builder.put_slice(item);
+        builder.put_str(item);
         builder.commit_row();
     }
 
@@ -209,7 +198,7 @@ impl ArgType for StringType {
 
     fn full_domain() -> Self::Domain {
         StringDomain {
-            min: vec![],
+            min: "".to_string(),
             max: None,
         }
     }
@@ -227,27 +216,34 @@ pub struct StringColumn {
 
 impl StringColumn {
     pub fn new(data: Buffer<u8>, offsets: Buffer<u64>) -> Self {
-        debug_assert!({ offsets.windows(2).all(|w| w[0] <= w[1]) });
-        let col = StringColumn { data, offsets };
-        // todo!("new string")
+        let col = BinaryColumn::new(data, offsets);
+
         col.check_utf8().unwrap();
-        col
+
+        unsafe { Self::from_binary_unchecked(col) }
     }
 
     /// # Safety
     /// This function is unsound iff:
     /// * the offsets are not monotonically increasing
-    /// * The `values` between two consecutive `offsets` are not valid utf8
+    /// * The `data` between two consecutive `offsets` are not valid utf8
     pub unsafe fn new_unchecked(data: Buffer<u8>, offsets: Buffer<u64>) -> Self {
-        debug_assert!({ offsets.windows(2).all(|w| w[0] <= w[1]) });
-        StringColumn { data, offsets }
+        let col = BinaryColumn::new(data, offsets);
+
+        #[cfg(debug_assertions)]
+        col.check_utf8().unwrap();
+
+        unsafe { Self::from_binary_unchecked(col) }
     }
 
     /// # Safety
     /// This function is unsound iff:
     /// * the offsets are not monotonically increasing
-    /// * The `values` between two consecutive `offsets` are not valid utf8
+    /// * The `data` between two consecutive `offsets` are not valid utf8
     pub unsafe fn from_binary_unchecked(col: BinaryColumn) -> Self {
+        #[cfg(debug_assertions)]
+        col.check_utf8().unwrap();
+
         StringColumn {
             data: col.data,
             offsets: col.offsets,
@@ -272,20 +268,34 @@ impl StringColumn {
         len * 8 + (offsets[len - 1] - offsets[0]) as usize
     }
 
-    pub fn index(&self, index: usize) -> Option<&[u8]> {
-        if index + 1 < self.offsets.len() {
-            Some(&self.data[(self.offsets[index] as usize)..(self.offsets[index + 1] as usize)])
-        } else {
-            None
+    pub fn index(&self, index: usize) -> Option<&str> {
+        if index + 1 >= self.offsets.len() {
+            return None;
         }
+
+        let bytes = &self.data[(self.offsets[index] as usize)..(self.offsets[index + 1] as usize)];
+
+        #[cfg(debug_assertions)]
+        bytes.check_utf8().unwrap();
+
+        unsafe { Some(std::str::from_utf8_unchecked(bytes)) }
     }
 
     /// # Safety
     ///
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
     #[inline]
-    pub unsafe fn index_unchecked(&self, index: usize) -> &[u8] {
-        &self.data[(self.offsets[index] as usize)..(self.offsets[index + 1] as usize)]
+    pub unsafe fn index_unchecked(&self, index: usize) -> &str {
+        debug_assert!(index + 1 < self.offsets.len());
+
+        let start = *self.offsets.get_unchecked(index) as usize;
+        let end = *self.offsets.get_unchecked(index + 1) as usize;
+        let bytes = &self.data.get_unchecked(start..end);
+
+        #[cfg(debug_assertions)]
+        bytes.check_utf8().unwrap();
+
+        std::str::from_utf8_unchecked(bytes)
     }
 
     pub fn slice(&self, range: Range<usize>) -> Self {
@@ -366,12 +376,18 @@ pub struct StringIterator<'a> {
 }
 
 impl<'a> Iterator for StringIterator<'a> {
-    type Item = &'a [u8];
+    type Item = &'a str;
 
     fn next(&mut self) -> Option<Self::Item> {
-        self.offsets
+        let bytes = self
+            .offsets
             .next()
-            .map(|range| &self.data[(range[0] as usize)..(range[1] as usize)])
+            .map(|range| &self.data[(range[0] as usize)..(range[1] as usize)])?;
+
+        #[cfg(debug_assertions)]
+        bytes.check_utf8().unwrap();
+
+        unsafe { Some(std::str::from_utf8_unchecked(bytes)) }
     }
 
     fn size_hint(&self) -> (usize, Option<usize>) {
@@ -411,18 +427,32 @@ impl StringColumnBuilder {
     }
 
     pub fn from_data(data: Vec<u8>, offsets: Vec<u64>) -> Self {
+        let builder = BinaryColumnBuilder::from_data(data, offsets);
+        builder.check_utf8().unwrap();
+        unsafe { StringColumnBuilder::from_binary_unchecked(builder) }
+    }
+
+    /// # Safety
+    /// This function is unsound iff:
+    /// * the offsets are not monotonically increasing
+    /// * The `data` between two consecutive `offsets` are not valid utf8
+    pub unsafe fn from_binary_unchecked(col: BinaryColumnBuilder) -> Self {
+        #[cfg(debug_assertions)]
+        col.check_utf8().unwrap();
+
         StringColumnBuilder {
-            need_estimated: false,
-            data,
-            offsets,
+            need_estimated: col.need_estimated,
+            data: col.data,
+            offsets: col.offsets,
         }
     }
 
-    pub fn repeat(scalar: &[u8], n: usize) -> Self {
-        let len = scalar.len();
+    pub fn repeat(scalar: &str, n: usize) -> Self {
+        let bytes = scalar.as_bytes();
+        let len = bytes.len();
         let mut data = Vec::with_capacity(len * n);
         for _ in 0..n {
-            data.extend_from_slice(scalar);
+            data.extend_from_slice(bytes);
         }
         let offsets = once(0)
             .chain((0..n).map(|i| (len * (i + 1)) as u64))
@@ -442,23 +472,23 @@ impl StringColumnBuilder {
         self.offsets.len() * 8 + self.data.len()
     }
 
-    pub fn put_u8(&mut self, item: u8) {
-        self.data.push(item);
-    }
-
     pub fn put_char(&mut self, item: char) {
         self.data
             .extend_from_slice(item.encode_utf8(&mut [0; 4]).as_bytes());
     }
 
     #[inline]
-    pub fn put_str(&mut self, item: &str) {
-        self.data.extend_from_slice(item.as_bytes());
+    #[deprecated]
+    pub fn put_slice(&mut self, item: &[u8]) {
+        #[cfg(debug_assertions)]
+        item.check_utf8().unwrap();
+
+        self.data.extend_from_slice(item);
     }
 
     #[inline]
-    pub fn put_slice(&mut self, item: &[u8]) {
-        self.data.extend_from_slice(item);
+    pub fn put_str(&mut self, item: &str) {
+        self.data.extend_from_slice(item.as_bytes());
     }
 
     pub fn put_char_iter(&mut self, iter: impl Iterator<Item = char>) {
@@ -469,10 +499,6 @@ impl StringColumnBuilder {
         }
     }
 
-    pub fn put(&mut self, item: &[u8]) {
-        self.data.extend_from_slice(item);
-    }
-
     #[inline]
     pub fn commit_row(&mut self) {
         self.offsets.push(self.data.len() as u64);
@@ -511,12 +537,18 @@ impl StringColumnBuilder {
     }
 
     pub fn build(self) -> StringColumn {
-        StringColumn::new(self.data.into(), self.offsets.into())
+        unsafe { StringColumn::new_unchecked(self.data.into(), self.offsets.into()) }
     }
 
-    pub fn build_scalar(self) -> Vec<u8> {
+    pub fn build_scalar(self) -> String {
         assert_eq!(self.offsets.len(), 2);
-        self.data[(self.offsets[0] as usize)..(self.offsets[1] as usize)].to_vec()
+
+        let bytes = self.data[(self.offsets[0] as usize)..(self.offsets[1] as usize)].to_vec();
+
+        #[cfg(debug_assertions)]
+        bytes.check_utf8().unwrap();
+
+        unsafe { String::from_utf8_unchecked(bytes) }
     }
 
     #[inline]
@@ -527,32 +559,42 @@ impl StringColumnBuilder {
     /// # Safety
     ///
     /// Calling this method with an out-of-bounds index is *[undefined behavior]*
-    pub unsafe fn index_unchecked(&self, row: usize) -> &[u8] {
+    pub unsafe fn index_unchecked(&self, row: usize) -> &str {
+        debug_assert!(row + 1 < self.offsets.len());
+
         let start = *self.offsets.get_unchecked(row) as usize;
         let end = *self.offsets.get_unchecked(row + 1) as usize;
-        // soundness: the invariant of the struct
-        self.data.get_unchecked(start..end)
+        let bytes = self.data.get_unchecked(start..end);
+
+        #[cfg(debug_assertions)]
+        bytes.check_utf8().unwrap();
+
+        std::str::from_utf8_unchecked(bytes)
     }
 
-    pub fn pop(&mut self) -> Option<Vec<u8>> {
+    pub fn pop(&mut self) -> Option<String> {
         if self.len() > 0 {
             let index = self.len() - 1;
             let start = unsafe { *self.offsets.get_unchecked(index) as usize };
             self.offsets.pop();
             let val = self.data.split_off(start);
-            Some(val)
+
+            #[cfg(debug_assertions)]
+            val.check_utf8().unwrap();
+
+            Some(unsafe { String::from_utf8_unchecked(val) })
         } else {
             None
         }
     }
 }
 
-impl<'a> FromIterator<&'a [u8]> for StringColumnBuilder {
-    fn from_iter<T: IntoIterator<Item = &'a [u8]>>(iter: T) -> Self {
+impl<'a> FromIterator<&'a str> for StringColumnBuilder {
+    fn from_iter<T: IntoIterator<Item = &'a str>>(iter: T) -> Self {
         let iter = iter.into_iter();
         let mut builder = StringColumnBuilder::with_capacity(iter.size_hint().0, 0);
         for item in iter {
-            builder.put_slice(item);
+            builder.put_str(item);
             builder.commit_row();
         }
         builder
@@ -584,34 +626,9 @@ impl TryFrom<BinaryColumnBuilder> for StringColumnBuilder {
 
 #[derive(Debug, Clone, PartialEq, Eq)]
 pub struct StringDomain {
-    pub min: Vec<u8>,
+    pub min: String,
     // max value is None for full domain
-    pub max: Option<Vec<u8>>,
-}
-
-impl StringDomain {
-    pub fn unify(&self, other: &Self) -> (SimpleDomain<Vec<u8>>, SimpleDomain<Vec<u8>>) {
-        let mut max_size = self.min.len().max(other.min.len());
-        if let Some(max) = &self.max {
-            max_size = max_size.max(max.len());
-        }
-        if let Some(max) = &other.max {
-            max_size = max_size.max(max.len());
-        }
-
-        let max_value = vec![255; max_size + 1];
-
-        (
-            SimpleDomain {
-                min: self.min.clone(),
-                max: self.max.clone().unwrap_or_else(|| max_value.clone()),
-            },
-            SimpleDomain {
-                min: other.min.clone(),
-                max: other.max.clone().unwrap_or_else(|| max_value.clone()),
-            },
-        )
-    }
+    pub max: Option<String>,
 }
 
 pub trait CheckUTF8 {
@@ -634,16 +651,13 @@ impl CheckUTF8 for &[u8] {
     }
 }
 
-impl CheckUTF8 for BinaryColumn {
+impl CheckUTF8 for Vec<u8> {
     fn check_utf8(&self) -> Result<()> {
-        for val in self.iter() {
-            val.check_utf8()?;
-        }
-        Ok(())
+        self.as_slice().check_utf8()
     }
 }
 
-impl CheckUTF8 for StringColumn {
+impl CheckUTF8 for BinaryColumn {
     fn check_utf8(&self) -> Result<()> {
         for val in self.iter() {
             val.check_utf8()?;
@@ -660,12 +674,3 @@ impl CheckUTF8 for BinaryColumnBuilder {
         Ok(())
     }
 }
-
-impl CheckUTF8 for StringColumnBuilder {
-    fn check_utf8(&self) -> Result<()> {
-        for row in 0..self.len() {
-            unsafe { self.index_unchecked(row) }.check_utf8()?;
-        }
-        Ok(())
-    }
-}
diff --git a/src/query/expression/src/types/timestamp.rs b/src/query/expression/src/types/timestamp.rs
index 8bdd9e3cb41c..4fbf8e0a8ef5 100644
--- a/src/query/expression/src/types/timestamp.rs
+++ b/src/query/expression/src/types/timestamp.rs
@@ -141,6 +141,8 @@ impl ValueType for TimestampType {
 
     #[inline(always)]
     unsafe fn index_column_unchecked(col: &Self::Column, index: usize) -> Self::ScalarRef<'_> {
+        debug_assert!(index < col.len());
+
         *col.get_unchecked(index)
     }
 
diff --git a/src/query/expression/src/types/variant.rs b/src/query/expression/src/types/variant.rs
index 64cea7623112..31a408ea52d8 100644
--- a/src/query/expression/src/types/variant.rs
+++ b/src/query/expression/src/types/variant.rs
@@ -207,7 +207,7 @@ pub fn cast_scalar_to_variant(scalar: ScalarRef, tz: TzLUT, buf: &mut Vec<u8>) {
         ScalarRef::Decimal(x) => x.to_float64().into(),
         ScalarRef::Boolean(b) => jsonb::Value::Bool(b),
         ScalarRef::Binary(s) => jsonb::Value::String(hex::encode_upper(s).into()),
-        ScalarRef::String(s) => jsonb::Value::String(String::from_utf8_lossy(s)),
+        ScalarRef::String(s) => jsonb::Value::String(s.into()),
         ScalarRef::Timestamp(ts) => timestamp_to_string(ts, inner_tz).to_string().into(),
         ScalarRef::Date(d) => date_to_string(d, inner_tz).to_string().into(),
         ScalarRef::Array(col) => {
@@ -221,7 +221,7 @@ pub fn cast_scalar_to_variant(scalar: ScalarRef, tz: TzLUT, buf: &mut Vec<u8>) {
                 .iter()
                 .map(|(k, v)| {
                     let key = match k {
-                        ScalarRef::String(v) => unsafe { String::from_utf8_unchecked(v.to_vec()) },
+                        ScalarRef::String(v) => v.to_string(),
                         ScalarRef::Number(v) => v.to_string(),
                         ScalarRef::Decimal(v) => v.to_string(),
                         ScalarRef::Boolean(v) => v.to_string(),
diff --git a/src/query/expression/src/utils/column_from.rs b/src/query/expression/src/utils/column_from.rs
index 0c180ba3508e..7c3dfef7faa0 100755
--- a/src/query/expression/src/utils/column_from.rs
+++ b/src/query/expression/src/utils/column_from.rs
@@ -84,13 +84,13 @@ impl<'a> FromData<&'a [u8]> for BinaryType {
 
 impl<'a> FromData<&'a str> for StringType {
     fn from_data(d: Vec<&'a str>) -> Column {
-        StringType::from_data(d.into_iter().map(|d| d.as_bytes().to_vec()).collect_vec())
+        StringType::from_data(d.into_iter().map(|d| d.to_string()).collect_vec())
     }
 
     fn from_opt_data(d: Vec<Option<&'a str>>) -> Column {
         StringType::from_opt_data(
             d.into_iter()
-                .map(|d| d.map(|d| d.as_bytes().to_vec()))
+                .map(|d| d.map(|d| d.to_string()))
                 .collect_vec(),
         )
     }
diff --git a/src/query/expression/src/utils/display.rs b/src/query/expression/src/utils/display.rs
index 5044de970612..a257c5ec92f4 100755
--- a/src/query/expression/src/utils/display.rs
+++ b/src/query/expression/src/utils/display.rs
@@ -121,15 +121,7 @@ impl<'a> Debug for ScalarRef<'a> {
                 }
                 Ok(())
             }
-            ScalarRef::String(s) => match std::str::from_utf8(s) {
-                Ok(v) => write!(f, "{:?}", v),
-                Err(_e) => {
-                    for c in *s {
-                        write!(f, "{c:02X}")?;
-                    }
-                    Ok(())
-                }
-            },
+            ScalarRef::String(s) => write!(f, "{s:?}"),
             ScalarRef::Timestamp(t) => write!(f, "{t:?}"),
             ScalarRef::Date(d) => write!(f, "{d:?}"),
             ScalarRef::Array(col) => write!(f, "[{}]", col.iter().join(", ")),
@@ -212,15 +204,7 @@ impl<'a> Display for ScalarRef<'a> {
                 }
                 Ok(())
             }
-            ScalarRef::String(s) => match std::str::from_utf8(s) {
-                Ok(v) => write!(f, "'{}'", v),
-                Err(_e) => {
-                    for c in *s {
-                        write!(f, "{c:02X}")?;
-                    }
-                    Ok(())
-                }
-            },
+            ScalarRef::String(s) => write!(f, "'{s}'"),
             ScalarRef::Timestamp(t) => write!(f, "'{}'", timestamp_to_string(*t, Tz::UTC)),
             ScalarRef::Date(d) => write!(f, "'{}'", date_to_string(*d as i64, Tz::UTC)),
             ScalarRef::Array(col) => write!(f, "[{}]", col.iter().join(", ")),
@@ -955,14 +939,9 @@ impl Display for BooleanDomain {
 impl Display for StringDomain {
     fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
         if let Some(max) = &self.max {
-            write!(
-                f,
-                "{{{:?}..={:?}}}",
-                String::from_utf8_lossy(&self.min),
-                String::from_utf8_lossy(max)
-            )
+            write!(f, "{{{:?}..={:?}}}", &self.min, max)
         } else {
-            write!(f, "{{{:?}..}}", String::from_utf8_lossy(&self.min))
+            write!(f, "{{{:?}..}}", &self.min)
         }
     }
 }
diff --git a/src/query/expression/src/values.rs b/src/query/expression/src/values.rs
index 9acdc5d1f10a..12c80fa4133a 100755
--- a/src/query/expression/src/values.rs
+++ b/src/query/expression/src/values.rs
@@ -66,7 +66,6 @@ use crate::types::number::NumberScalar;
 use crate::types::number::SimpleDomain;
 use crate::types::number::F32;
 use crate::types::number::F64;
-use crate::types::string::CheckUTF8;
 use crate::types::string::StringColumn;
 use crate::types::string::StringColumnBuilder;
 use crate::types::string::StringDomain;
@@ -113,7 +112,7 @@ pub enum Scalar {
     Date(i32),
     Boolean(bool),
     Binary(Vec<u8>),
-    String(Vec<u8>),
+    String(String),
     Array(Column),
     Map(Column),
     Bitmap(Vec<u8>),
@@ -131,7 +130,7 @@ pub enum ScalarRef<'a> {
     Decimal(DecimalScalar),
     Boolean(bool),
     Binary(&'a [u8]),
-    String(&'a [u8]),
+    String(&'a str),
     Timestamp(i64),
     Date(i32),
     Array(Column),
@@ -345,7 +344,7 @@ impl Scalar {
             Scalar::Decimal(d) => ScalarRef::Decimal(*d),
             Scalar::Boolean(b) => ScalarRef::Boolean(*b),
             Scalar::Binary(s) => ScalarRef::Binary(s.as_slice()),
-            Scalar::String(s) => ScalarRef::String(s.as_slice()),
+            Scalar::String(s) => ScalarRef::String(s.as_str()),
             Scalar::Timestamp(t) => ScalarRef::Timestamp(*t),
             Scalar::Date(d) => ScalarRef::Date(*d),
             Scalar::Array(col) => ScalarRef::Array(col.clone()),
@@ -362,8 +361,8 @@ impl Scalar {
             DataType::EmptyArray => Scalar::EmptyArray,
             DataType::EmptyMap => Scalar::EmptyMap,
             DataType::Boolean => Scalar::Boolean(false),
-            DataType::Binary => Scalar::Binary(vec![]),
-            DataType::String => Scalar::String(vec![]),
+            DataType::Binary => Scalar::Binary(Vec::new()),
+            DataType::String => Scalar::String(String::new()),
             DataType::Number(num_ty) => Scalar::Number(match num_ty {
                 NumberDataType::UInt8 => NumberScalar::UInt8(0),
                 NumberDataType::UInt16 => NumberScalar::UInt16(0),
@@ -441,7 +440,7 @@ impl<'a> ScalarRef<'a> {
             ScalarRef::Decimal(d) => Scalar::Decimal(*d),
             ScalarRef::Boolean(b) => Scalar::Boolean(*b),
             ScalarRef::Binary(s) => Scalar::Binary(s.to_vec()),
-            ScalarRef::String(s) => Scalar::String(s.to_vec()),
+            ScalarRef::String(s) => Scalar::String(s.to_string()),
             ScalarRef::Timestamp(t) => Scalar::Timestamp(*t),
             ScalarRef::Date(d) => Scalar::Date(*d),
             ScalarRef::Array(col) => Scalar::Array(col.clone()),
@@ -483,8 +482,8 @@ impl<'a> ScalarRef<'a> {
                 has_true: false,
             }),
             ScalarRef::String(s) => Domain::String(StringDomain {
-                min: s.to_vec(),
-                max: Some(s.to_vec()),
+                min: s.to_string(),
+                max: Some(s.to_string()),
             }),
             ScalarRef::Timestamp(t) => Domain::Timestamp(SimpleDomain { min: *t, max: *t }),
             ScalarRef::Date(d) => Domain::Date(SimpleDomain { min: *d, max: *d }),
@@ -927,8 +926,8 @@ impl Column {
             Column::String(col) => {
                 let (min, max) = StringType::iter_column(col).minmax().into_option().unwrap();
                 Domain::String(StringDomain {
-                    min: min.to_vec(),
-                    max: Some(max.to_vec()),
+                    min: min.to_string(),
+                    max: Some(max.to_string()),
                 })
             }
             Column::Timestamp(col) => {
@@ -1085,8 +1084,8 @@ impl Column {
                         rng.sample_iter(&Alphanumeric)
                             // randomly generate 5 characters.
                             .take(5)
-                            .map(u8::from)
-                            .collect::<Vec<_>>()
+                            .map(char::from)
+                            .collect::<String>()
                     })
                     .collect_vec(),
             ),
@@ -1620,7 +1619,7 @@ impl ColumnBuilder {
                 builder.commit_row();
             }
             (ColumnBuilder::String(builder), ScalarRef::String(value)) => {
-                builder.put_slice(value);
+                builder.put_str(value);
                 builder.commit_row();
             }
             (ColumnBuilder::Timestamp(builder), ScalarRef::Timestamp(value)) => {
@@ -1723,7 +1722,7 @@ impl ColumnBuilder {
                 builder.commit_row();
 
                 #[cfg(debug_assertions)]
-                (&builder.data[last..last + offset]).check_utf8().unwrap();
+                string::CheckUTF8::check_utf8(&(&builder.data[last..last + offset])).unwrap();
             }
             ColumnBuilder::Timestamp(builder) => {
                 let value: i64 = reader.read_scalar()?;
@@ -1813,11 +1812,13 @@ impl ColumnBuilder {
             }
             ColumnBuilder::String(builder) => {
                 for row in 0..rows {
+                    let bytes = &reader[step * row..];
+
                     #[cfg(debug_assertions)]
-                    (&reader[step * row..]).check_utf8().unwrap();
+                    string::CheckUTF8::check_utf8(&bytes).unwrap();
 
-                    let reader = &reader[step * row..];
-                    builder.put_slice(reader);
+                    let s = unsafe { std::str::from_utf8_unchecked(bytes) };
+                    builder.put_str(s);
                     builder.commit_row();
                 }
             }
diff --git a/src/query/expression/tests/it/block.rs b/src/query/expression/tests/it/block.rs
index d94ac3a93560..117b9ab15b0e 100644
--- a/src/query/expression/tests/it/block.rs
+++ b/src/query/expression/tests/it/block.rs
@@ -12,10 +12,10 @@ use crate::common::new_block;
 
 #[test]
 fn test_split_block() {
-    let value = b"abc";
+    let value = "abc";
     let n = 10;
     let block = new_block(&[Column::String(
-        StringColumnBuilder::repeat(&value[..], n).build(),
+        StringColumnBuilder::repeat(value, n).build(),
     )]);
     let sizes = block
         .split_by_rows_if_needed_no_tail(3)
@@ -27,11 +27,11 @@ fn test_split_block() {
 
 #[test]
 fn test_box_render_block() {
-    let value = b"abc";
+    let value = "abc";
     let n = 10;
     let block = new_block(&[
         Int32Type::from_data(vec![1, 2, 3, 4, 5, 6, 7, 8, 9, 10]),
-        Column::String(StringColumnBuilder::repeat(&value[..], n).build()),
+        Column::String(StringColumnBuilder::repeat(value, n).build()),
     ]);
 
     let schema = DataSchemaRefExt::create(vec![
diff --git a/src/query/expression/tests/it/meta_scalar.rs b/src/query/expression/tests/it/meta_scalar.rs
index 9499329598c0..e420ad0c1589 100644
--- a/src/query/expression/tests/it/meta_scalar.rs
+++ b/src/query/expression/tests/it/meta_scalar.rs
@@ -99,9 +99,9 @@ pub fn test_simple_converts() -> databend_common_exception::Result<()> {
             }
 
             // TODO: comment these when we swith string in scalar
-            let data = rmp_serde::to_vec(&scalars).unwrap();
-            let new_scalars: Vec<IndexScalar> = rmp_serde::from_slice(&data).unwrap();
-            assert_eq!(simple_scalars, new_scalars);
+            // let data = rmp_serde::to_vec(&scalars).unwrap();
+            // let new_scalars: Vec<IndexScalar> = rmp_serde::from_slice(&data).unwrap();
+            // assert_eq!(simple_scalars, new_scalars);
         }
     }
 
diff --git a/src/query/expression/tests/it/row.rs b/src/query/expression/tests/it/row.rs
index dd6a9dc18f2e..1e8106395c28 100644
--- a/src/query/expression/tests/it/row.rs
+++ b/src/query/expression/tests/it/row.rs
@@ -33,6 +33,7 @@ use itertools::Itertools;
 use jsonb::convert_to_comparable;
 use jsonb::parse_value;
 use ordered_float::OrderedFloat;
+use rand::distributions::Alphanumeric;
 use rand::distributions::Standard;
 use rand::prelude::Distribution;
 use rand::thread_rng;
@@ -303,13 +304,8 @@ fn test_binary() {
 
 #[test]
 fn test_string() {
-    let col = StringType::from_opt_data(vec![
-        Some("hello".as_bytes().to_vec()),
-        Some("he".as_bytes().to_vec()),
-        None,
-        Some("foo".as_bytes().to_vec()),
-        Some("".as_bytes().to_vec()),
-    ]);
+    let col =
+        StringType::from_opt_data(vec![Some("hello"), Some("he"), None, Some("foo"), Some("")]);
 
     let converter =
         RowConverter::new(vec![SortField::new(DataType::String.wrap_nullable())]).unwrap();
@@ -327,13 +323,13 @@ fn test_string() {
 
     let col = StringType::from_opt_data(vec![
         None,
-        Some(vec![0_u8; 0]),
-        Some(vec![0_u8; 6]),
-        Some(vec![0_u8; BLOCK_SIZE]),
-        Some(vec![0_u8; BLOCK_SIZE + 1]),
-        Some(vec![1_u8; 6]),
-        Some(vec![1_u8; BLOCK_SIZE]),
-        Some(vec![1_u8; BLOCK_SIZE + 1]),
+        Some(String::from_utf8(vec![0_u8; 0]).unwrap()),
+        Some(String::from_utf8(vec![0_u8; 6]).unwrap()),
+        Some(String::from_utf8(vec![0_u8; BLOCK_SIZE]).unwrap()),
+        Some(String::from_utf8(vec![0_u8; BLOCK_SIZE + 1]).unwrap()),
+        Some(String::from_utf8(vec![1_u8; 6]).unwrap()),
+        Some(String::from_utf8(vec![1_u8; BLOCK_SIZE]).unwrap()),
+        Some(String::from_utf8(vec![1_u8; BLOCK_SIZE + 1]).unwrap()),
     ]);
     let num_rows = col.len();
 
@@ -481,7 +477,12 @@ fn generate_string_column(len: usize, valid_percent: f64) -> Column {
         .map(|_| {
             rng.gen_bool(valid_percent).then(|| {
                 let len = rng.gen_range(0..100);
-                (0..len).map(|_| rng.gen_range(0..128)).collect_vec()
+                thread_rng()
+                    .sample_iter(&Alphanumeric)
+                    // randomly generate 5 characters.
+                    .take(len)
+                    .map(char::from)
+                    .collect::<String>()
             })
         })
         .collect::<Vec<_>>();
diff --git a/src/query/expression/tests/it/schema.rs b/src/query/expression/tests/it/schema.rs
index 27294100e7d2..241b426a030e 100644
--- a/src/query/expression/tests/it/schema.rs
+++ b/src/query/expression/tests/it/schema.rs
@@ -215,7 +215,7 @@ fn test_field_leaf_default_values() -> Result<()> {
         Scalar::Tuple(vec![
             Scalar::Tuple(vec![
                 Scalar::Boolean(true),
-                Scalar::String(['a', 'b'].iter().map(|c| *c as u8).collect::<Vec<_>>()),
+                Scalar::String("ab".to_string()),
             ]),
             Scalar::Number(databend_common_expression::types::number::NumberScalar::Int64(2)),
         ]),
@@ -229,10 +229,7 @@ fn test_field_leaf_default_values() -> Result<()> {
             Scalar::Number(databend_common_expression::types::number::NumberScalar::UInt64(1)),
         ),
         (1, Scalar::Boolean(true)),
-        (
-            2,
-            Scalar::String(['a', 'b'].iter().map(|c| *c as u8).collect::<Vec<_>>()),
-        ),
+        (2, Scalar::String("ab".to_string())),
         (
             3,
             Scalar::Number(databend_common_expression::types::number::NumberScalar::Int64(2)),
diff --git a/src/query/formats/src/field_encoder/csv.rs b/src/query/formats/src/field_encoder/csv.rs
index 5c151e4984eb..1f02dab517c1 100644
--- a/src/query/formats/src/field_encoder/csv.rs
+++ b/src/query/formats/src/field_encoder/csv.rs
@@ -128,7 +128,7 @@ impl FieldEncoderCSV {
             }
             Column::String(c) => {
                 let buf = unsafe { c.index_unchecked(row_index) };
-                self.string_formatter.write_string(buf, out_buf);
+                self.string_formatter.write_string(buf.as_bytes(), out_buf);
             }
 
             Column::Date(..) | Column::Timestamp(..) | Column::Bitmap(..) | Column::Variant(..) => {
diff --git a/src/query/formats/src/field_encoder/json.rs b/src/query/formats/src/field_encoder/json.rs
index 49c7473a9a13..b097b4209c94 100644
--- a/src/query/formats/src/field_encoder/json.rs
+++ b/src/query/formats/src/field_encoder/json.rs
@@ -63,7 +63,7 @@ impl FieldEncoderJSON {
             }
             Column::String(c) => {
                 let buf = unsafe { c.index_unchecked(row_index) };
-                self.write_string(buf, out_buf);
+                self.write_string(buf.as_bytes(), out_buf);
             }
 
             Column::Date(..) | Column::Timestamp(..) | Column::Bitmap(..) => {
diff --git a/src/query/formats/src/field_encoder/values.rs b/src/query/formats/src/field_encoder/values.rs
index 352eadb5a368..490c1e49df6a 100644
--- a/src/query/formats/src/field_encoder/values.rs
+++ b/src/query/formats/src/field_encoder/values.rs
@@ -232,7 +232,7 @@ impl FieldEncoderValues {
         in_nested: bool,
     ) {
         self.write_string_inner(
-            unsafe { column.index_unchecked(row_index) },
+            unsafe { column.index_unchecked(row_index).as_bytes() },
             out_buf,
             in_nested,
         );
diff --git a/src/query/formats/src/output_format/json.rs b/src/query/formats/src/output_format/json.rs
index 3edddf01f2a6..9e65f282f35e 100644
--- a/src/query/formats/src/output_format/json.rs
+++ b/src/query/formats/src/output_format/json.rs
@@ -99,7 +99,7 @@ fn scalar_to_json(s: ScalarRef<'_>, format: &FormatSettings) -> JsonValue {
         ScalarRef::EmptyArray => JsonValue::Array(vec![]),
         ScalarRef::EmptyMap => JsonValue::Object(JsonMap::new()),
         ScalarRef::Binary(x) => JsonValue::String(hex::encode_upper(x)),
-        ScalarRef::String(x) => JsonValue::String(String::from_utf8_lossy(x).to_string()),
+        ScalarRef::String(x) => JsonValue::String(x.to_string()),
         ScalarRef::Array(x) => {
             let vals = x
                 .iter()
diff --git a/src/query/functions/Cargo.toml b/src/query/functions/Cargo.toml
index 702507bf4271..fad69009319d 100644
--- a/src/query/functions/Cargo.toml
+++ b/src/query/functions/Cargo.toml
@@ -27,7 +27,6 @@ blake3 = "1.3.1"
 borsh = { workspace = true, features = ["derive"] }
 bstr = "1.0.1"
 bumpalo = { workspace = true }
-bytes = { workspace = true }
 chrono = { workspace = true }
 chrono-tz = { workspace = true }
 crc32fast = "1.3.2"
@@ -60,6 +59,7 @@ simdutf8 = "0.1.4"
 siphasher = "0.3"
 streaming_algorithms = { git = "https://github.com/ariesdevil/streaming_algorithms", rev = "2839d5d" }
 strength_reduce = "0.2.3"
+stringslice = "0.2.0"
 twox-hash = "1.6.3"
 
 [dev-dependencies]
diff --git a/src/query/functions/src/aggregates/aggregate_distinct_state.rs b/src/query/functions/src/aggregates/aggregate_distinct_state.rs
index 31f57a207104..b9be795b7376 100644
--- a/src/query/functions/src/aggregates/aggregate_distinct_state.rs
+++ b/src/query/functions/src/aggregates/aggregate_distinct_state.rs
@@ -194,7 +194,7 @@ impl DistinctStateFunc for AggregateDistinctStringState {
     fn add(&mut self, columns: &[Column], row: usize) -> Result<()> {
         let column = StringType::try_downcast_column(&columns[0]).unwrap();
         let data = unsafe { column.index_unchecked(row) };
-        let _ = self.set.set_insert(data);
+        let _ = self.set.set_insert(data.as_bytes());
         Ok(())
     }
 
@@ -211,14 +211,14 @@ impl DistinctStateFunc for AggregateDistinctStringState {
                 for row in 0..input_rows {
                     if v.get_bit(row) {
                         let data = unsafe { column.index_unchecked(row) };
-                        let _ = self.set.set_insert(data);
+                        let _ = self.set.set_insert(data.as_bytes());
                     }
                 }
             }
             None => {
                 for row in 0..input_rows {
                     let data = unsafe { column.index_unchecked(row) };
-                    let _ = self.set.set_insert(data);
+                    let _ = self.set.set_insert(data.as_bytes());
                 }
             }
         }
@@ -233,7 +233,7 @@ impl DistinctStateFunc for AggregateDistinctStringState {
     fn build_columns(&mut self, _types: &[DataType]) -> Result<Vec<Column>> {
         let mut builder = StringColumnBuilder::with_capacity(self.set.len(), self.set.len() * 2);
         for key in self.set.iter() {
-            builder.put_slice(key.key());
+            builder.put_str(unsafe { std::str::from_utf8_unchecked(key.key()) });
             builder.commit_row();
         }
         Ok(vec![Column::String(builder.build())])
@@ -360,7 +360,7 @@ impl DistinctStateFunc for AggregateUniqStringState {
         let column = columns[0].as_string().unwrap();
         let data = unsafe { column.index_unchecked(row) };
         let mut hasher = SipHasher24::new();
-        hasher.write(data);
+        hasher.write(data.as_bytes());
         let hash128 = hasher.finish128();
         let _ = self.set.set_insert(hash128.into()).is_ok();
         Ok(())
@@ -378,7 +378,7 @@ impl DistinctStateFunc for AggregateUniqStringState {
                 for (t, v) in column.iter().zip(v.iter()) {
                     if v {
                         let mut hasher = SipHasher24::new();
-                        hasher.write(t);
+                        hasher.write(t.as_bytes());
                         let hash128 = hasher.finish128();
                         let _ = self.set.set_insert(hash128.into()).is_ok();
                     }
@@ -388,7 +388,7 @@ impl DistinctStateFunc for AggregateUniqStringState {
                 for row in 0..input_rows {
                     let data = unsafe { column.index_unchecked(row) };
                     let mut hasher = SipHasher24::new();
-                    hasher.write(data);
+                    hasher.write(data.as_bytes());
                     let hash128 = hasher.finish128();
                     let _ = self.set.set_insert(hash128.into()).is_ok();
                 }
diff --git a/src/query/functions/src/aggregates/aggregate_string_agg.rs b/src/query/functions/src/aggregates/aggregate_string_agg.rs
index 29f024770a8b..354e1660b4d9 100644
--- a/src/query/functions/src/aggregates/aggregate_string_agg.rs
+++ b/src/query/functions/src/aggregates/aggregate_string_agg.rs
@@ -37,13 +37,13 @@ use crate::aggregates::AggregateFunction;
 
 #[derive(BorshSerialize, BorshDeserialize, Debug)]
 pub struct StringAggState {
-    values: Vec<u8>,
+    values: String,
 }
 
 #[derive(Clone)]
 pub struct AggregateStringAggFunction {
     display_name: String,
-    delimiter: Vec<u8>,
+    delimiter: String,
 }
 
 impl AggregateFunction for AggregateStringAggFunction {
@@ -56,7 +56,9 @@ impl AggregateFunction for AggregateStringAggFunction {
     }
 
     fn init_state(&self, place: StateAddr) {
-        place.write(|| StringAggState { values: Vec::new() });
+        place.write(|| StringAggState {
+            values: String::new(),
+        });
     }
 
     fn state_layout(&self) -> Layout {
@@ -76,15 +78,15 @@ impl AggregateFunction for AggregateStringAggFunction {
             Some(validity) => {
                 column.iter().zip(validity.iter()).for_each(|(v, b)| {
                     if b {
-                        state.values.extend_from_slice(v);
-                        state.values.extend_from_slice(self.delimiter.as_slice());
+                        state.values.push_str(v);
+                        state.values.push_str(&self.delimiter);
                     }
                 });
             }
             None => {
                 column.iter().for_each(|v| {
-                    state.values.extend_from_slice(v);
-                    state.values.extend_from_slice(self.delimiter.as_slice());
+                    state.values.push_str(v);
+                    state.values.push_str(&self.delimiter);
                 });
             }
         }
@@ -103,8 +105,8 @@ impl AggregateFunction for AggregateStringAggFunction {
         column_iter.zip(places.iter()).for_each(|(v, place)| {
             let addr = place.next(offset);
             let state = addr.get::<StringAggState>();
-            state.values.extend_from_slice(v);
-            state.values.extend_from_slice(self.delimiter.as_slice());
+            state.values.push_str(v);
+            state.values.push_str(&self.delimiter);
         });
         Ok(())
     }
@@ -114,8 +116,8 @@ impl AggregateFunction for AggregateStringAggFunction {
         let v = StringType::index_column(&column, row);
         if let Some(v) = v {
             let state = place.get::<StringAggState>();
-            state.values.extend_from_slice(v);
-            state.values.extend_from_slice(self.delimiter.as_slice());
+            state.values.push_str(v);
+            state.values.push_str(&self.delimiter);
         }
         Ok(())
     }
@@ -129,14 +131,14 @@ impl AggregateFunction for AggregateStringAggFunction {
     fn merge(&self, place: StateAddr, reader: &mut &[u8]) -> Result<()> {
         let state = place.get::<StringAggState>();
         let rhs: StringAggState = borsh_deserialize_state(reader)?;
-        state.values.extend_from_slice(rhs.values.as_slice());
+        state.values.push_str(&rhs.values);
         Ok(())
     }
 
     fn merge_states(&self, place: StateAddr, rhs: StateAddr) -> Result<()> {
         let state = place.get::<StringAggState>();
         let other = rhs.get::<StringAggState>();
-        state.values.extend_from_slice(other.values.as_slice());
+        state.values.push_str(&other.values);
         Ok(())
     }
 
@@ -145,7 +147,7 @@ impl AggregateFunction for AggregateStringAggFunction {
         let builder = StringType::try_downcast_builder(builder).unwrap();
         if !state.values.is_empty() {
             let len = state.values.len() - self.delimiter.len();
-            builder.put_slice(&state.values.as_slice()[..len]);
+            builder.put_str(&state.values[..len]);
         }
         builder.commit_row();
         Ok(())
@@ -168,7 +170,7 @@ impl fmt::Display for AggregateStringAggFunction {
 }
 
 impl AggregateStringAggFunction {
-    fn try_create(display_name: &str, delimiter: Vec<u8>) -> Result<Arc<dyn AggregateFunction>> {
+    fn try_create(display_name: &str, delimiter: String) -> Result<Arc<dyn AggregateFunction>> {
         let func = AggregateStringAggFunction {
             display_name: display_name.to_string(),
             delimiter,
@@ -193,7 +195,7 @@ pub fn try_create_aggregate_string_agg_function(
     let delimiter = if params.len() == 1 {
         params[0].as_string().unwrap().clone()
     } else {
-        vec![]
+        String::new()
     };
     AggregateStringAggFunction::try_create(display_name, delimiter)
 }
diff --git a/src/query/functions/src/lib.rs b/src/query/functions/src/lib.rs
index 12b59465f6df..0a6661974fc1 100644
--- a/src/query/functions/src/lib.rs
+++ b/src/query/functions/src/lib.rs
@@ -20,6 +20,7 @@
 #![feature(try_blocks)]
 #![feature(downcast_unchecked)]
 #![feature(lazy_cell)]
+#![feature(str_internals)]
 
 use aggregates::AggregateFunctionFactory;
 use ctor::ctor;
diff --git a/src/query/functions/src/scalars/arithmetic.rs b/src/query/functions/src/scalars/arithmetic.rs
index 190e5bc37425..8db8dcc0f217 100644
--- a/src/query/functions/src/scalars/arithmetic.rs
+++ b/src/query/functions/src/scalars/arithmetic.rs
@@ -833,8 +833,7 @@ fn register_string_to_number(registry: &mut FunctionRegistry) {
                         |_, _| FunctionDomain::MayThrow,
                         vectorize_with_builder_1_arg::<StringType, NumberType<DEST_TYPE>>(
                             move |val, output, ctx| {
-                                let str_val = String::from_utf8_lossy(val);
-                                match str_val.parse::<DEST_TYPE>() {
+                                match val.parse::<DEST_TYPE>() {
                                     Ok(new_val) => output.push(new_val),
                                     Err(e) => {
                                         ctx.set_error(output.len(), e.to_string());
@@ -854,8 +853,7 @@ fn register_string_to_number(registry: &mut FunctionRegistry) {
                             StringType,
                             NullableType<NumberType<DEST_TYPE>>,
                         >(|val, output, _| {
-                            let str_val = String::from_utf8_lossy(val);
-                            if let Ok(new_val) = str_val.parse::<DEST_TYPE>() {
+                            if let Ok(new_val) = val.parse::<DEST_TYPE>() {
                                 output.push(new_val);
                             } else {
                                 output.push_null();
@@ -876,7 +874,7 @@ pub fn register_number_to_string(registry: &mut FunctionRegistry) {
                         "to_string",
                         |_, _| FunctionDomain::Full,
                         |from, _| match from {
-                            ValueRef::Scalar(s) => Value::Scalar(s.to_string().into_bytes()),
+                            ValueRef::Scalar(s) => Value::Scalar(s.to_string()),
                             ValueRef::Column(from) => {
                                 let options = NUM_TYPE::lexical_options();
                                 const FORMAT: u128 = lexical_core::format::STANDARD;
@@ -913,7 +911,7 @@ pub fn register_number_to_string(registry: &mut FunctionRegistry) {
                     "try_to_string",
                     |_, _| FunctionDomain::Full,
                     |from, _| match from {
-                        ValueRef::Scalar(s) => Value::Scalar(Some(s.to_string().into_bytes())),
+                        ValueRef::Scalar(s) => Value::Scalar(Some(s.to_string())),
                         ValueRef::Column(from) => {
                             let options = NUM_TYPE::lexical_options();
                             const FORMAT: u128 = lexical_core::format::STANDARD;
diff --git a/src/query/functions/src/scalars/array.rs b/src/query/functions/src/scalars/array.rs
index 4d29a9c1972e..e96bfdb68f1c 100644
--- a/src/query/functions/src/scalars/array.rs
+++ b/src/query/functions/src/scalars/array.rs
@@ -57,6 +57,7 @@ use databend_common_expression::FunctionRegistry;
 use databend_common_expression::FunctionSignature;
 use databend_common_expression::Scalar;
 use databend_common_expression::ScalarRef;
+use databend_common_expression::SimpleDomainCmp;
 use databend_common_expression::SortColumnDescription;
 use databend_common_expression::Value;
 use databend_common_expression::ValueRef;
@@ -280,9 +281,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                 |lhs, rhs, output, _| {
                     for (i, d) in lhs.iter().enumerate() {
                         if i != 0 {
-                            output.put_slice(rhs);
+                            output.put_str(rhs);
                         }
-                        output.put_slice(d);
+                        output.put_str(d);
                     }
                     output.commit_row();
                 },
@@ -481,11 +482,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                 registry.register_passthrough_nullable_2_arg::<ArrayType<NumberType<NUM_TYPE>>, NumberType<NUM_TYPE>, BooleanType, _, _>(
                     "contains",
                     |_, lhs, rhs| {
-                        let has_true = lhs.is_some_and(|lhs| !(lhs.min > rhs.max || lhs.max < rhs.min));
-                        FunctionDomain::Domain(BooleanDomain {
-                            has_false: true,
-                            has_true,
-                        })
+                        lhs.as_ref().map(|lhs| {
+                            lhs.domain_contains(rhs)
+                        }).unwrap_or(FunctionDomain::Full)
                     },
                     |lhs, rhs, _| eval_contains::<NumberType<NUM_TYPE>>(lhs, rhs)
                 );
@@ -495,9 +494,11 @@ pub fn register(registry: &mut FunctionRegistry) {
 
     registry.register_passthrough_nullable_2_arg::<ArrayType<StringType>, StringType, BooleanType, _, _>(
         "contains",
-        |_, _, _| {
-            FunctionDomain::Full
-        },
+         |_, lhs, rhs| {
+                        lhs.as_ref().map(|lhs| {
+                            lhs.domain_contains(rhs)
+                        }).unwrap_or(FunctionDomain::Full)
+                    },
         |lhs, rhs, _| {
             match lhs {
                 ValueRef::Scalar(array) => {
diff --git a/src/query/functions/src/scalars/binary.rs b/src/query/functions/src/scalars/binary.rs
index dae984d1038d..58113ec7a9e2 100644
--- a/src/query/functions/src/scalars/binary.rs
+++ b/src/query/functions/src/scalars/binary.rs
@@ -13,20 +13,32 @@
 // limitations under the License.
 
 use std::io::Write;
+use std::sync::Arc;
 
 use databend_common_arrow::arrow::bitmap::Bitmap;
 use databend_common_expression::error_to_null;
+use databend_common_expression::passthrough_nullable;
 use databend_common_expression::types::binary::BinaryColumn;
 use databend_common_expression::types::binary::BinaryColumnBuilder;
 use databend_common_expression::types::nullable::NullableColumn;
 use databend_common_expression::types::string::StringColumn;
 use databend_common_expression::types::string::StringColumnBuilder;
+use databend_common_expression::types::AnyType;
 use databend_common_expression::types::BinaryType;
+use databend_common_expression::types::DataType;
+use databend_common_expression::types::NumberDataType;
 use databend_common_expression::types::NumberType;
 use databend_common_expression::types::StringType;
+use databend_common_expression::types::UInt8Type;
+use databend_common_expression::types::ValueType;
+use databend_common_expression::Column;
 use databend_common_expression::EvalContext;
+use databend_common_expression::Function;
 use databend_common_expression::FunctionDomain;
+use databend_common_expression::FunctionEval;
 use databend_common_expression::FunctionRegistry;
+use databend_common_expression::FunctionSignature;
+use databend_common_expression::Scalar;
 use databend_common_expression::Value;
 use databend_common_expression::ValueRef;
 
@@ -68,7 +80,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         "to_binary",
         |_, _| FunctionDomain::Full,
         |val, _| match val {
-            ValueRef::Scalar(val) => Value::Scalar(val.to_vec()),
+            ValueRef::Scalar(val) => Value::Scalar(val.as_bytes().to_vec()),
             ValueRef::Column(col) => Value::Column(col.into()),
         },
     );
@@ -77,7 +89,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         "try_to_binary",
         |_, _| FunctionDomain::Full,
         |val, _| match val {
-            ValueRef::Scalar(val) => Value::Scalar(Some(val.to_vec())),
+            ValueRef::Scalar(val) => Value::Scalar(Some(val.as_bytes().to_vec())),
             ValueRef::Column(col) => Value::Column(NullableColumn {
                 validity: Bitmap::new_constant(true, col.len()),
                 column: col.into(),
@@ -140,14 +152,61 @@ pub fn register(registry: &mut FunctionRegistry) {
         |_, _| FunctionDomain::Full,
         error_to_null(eval_from_base64),
     );
+
+    registry.register_function_factory("char", |_, args_type| {
+        if args_type.is_empty() {
+            return None;
+        }
+        let has_null = args_type.iter().any(|t| t.is_nullable_or_null());
+        let f = Function {
+            signature: FunctionSignature {
+                name: "char".to_string(),
+                args_type: vec![DataType::Number(NumberDataType::UInt8); args_type.len()],
+                return_type: DataType::Binary,
+            },
+            eval: FunctionEval::Scalar {
+                calc_domain: Box::new(|_, _| FunctionDomain::Full),
+                eval: Box::new(char_fn),
+            },
+        };
+
+        if has_null {
+            Some(Arc::new(f.passthrough_nullable()))
+        } else {
+            Some(Arc::new(f))
+        }
+    });
+
+    // nullable char
+    registry.register_function_factory("char", |_, args_type| {
+        if args_type.is_empty() {
+            return None;
+        }
+        Some(Arc::new(Function {
+            signature: FunctionSignature {
+                name: "char".to_string(),
+                args_type: vec![
+                    DataType::Nullable(Box::new(DataType::Number(
+                        NumberDataType::UInt8
+                    )));
+                    args_type.len()
+                ],
+                return_type: DataType::Nullable(Box::new(DataType::Binary)),
+            },
+            eval: FunctionEval::Scalar {
+                calc_domain: Box::new(|_, _| FunctionDomain::MayThrow),
+                eval: Box::new(passthrough_nullable(char_fn)),
+            },
+        }))
+    });
 }
 
 fn eval_binary_to_string(val: ValueRef<BinaryType>, ctx: &mut EvalContext) -> Value<StringType> {
     vectorize_binary_to_string(
         |col| col.data().len(),
         |val, output, ctx| {
-            if simdutf8::basic::from_utf8(val).is_ok() {
-                output.put_slice(val);
+            if let Ok(val) = simdutf8::basic::from_utf8(val) {
+                output.put_str(val);
             } else {
                 ctx.set_error(output.len(), "invalid utf8 sequence");
             }
@@ -213,7 +272,7 @@ pub fn vectorize_binary_to_string(
 /// String to Binary scalar function with estimated output column capacity.
 pub fn vectorize_string_to_binary(
     estimate_bytes: impl Fn(&StringColumn) -> usize + Copy,
-    func: impl Fn(&[u8], &mut BinaryColumnBuilder, &mut EvalContext) + Copy,
+    func: impl Fn(&str, &mut BinaryColumnBuilder, &mut EvalContext) + Copy,
 ) -> impl Fn(ValueRef<StringType>, &mut EvalContext) -> Value<BinaryType> + Copy {
     move |arg1, ctx| match arg1 {
         ValueRef::Scalar(val) => {
@@ -232,3 +291,47 @@ pub fn vectorize_string_to_binary(
         }
     }
 }
+
+fn char_fn(args: &[ValueRef<AnyType>], _: &mut EvalContext) -> Value<AnyType> {
+    let args = args
+        .iter()
+        .map(|arg| arg.try_downcast::<UInt8Type>().unwrap())
+        .collect::<Vec<_>>();
+
+    let len = args.iter().find_map(|arg| match arg {
+        ValueRef::Column(col) => Some(col.len()),
+        _ => None,
+    });
+    let input_rows = len.unwrap_or(1);
+
+    let mut values: Vec<u8> = vec![0; input_rows * args.len()];
+    let values_ptr = values.as_mut_ptr();
+
+    for (i, arg) in args.iter().enumerate() {
+        match arg {
+            ValueRef::Scalar(v) => {
+                for j in 0..input_rows {
+                    unsafe {
+                        *values_ptr.add(args.len() * j + i) = *v;
+                    }
+                }
+            }
+            ValueRef::Column(c) => {
+                for (j, ch) in UInt8Type::iter_column(c).enumerate() {
+                    unsafe {
+                        *values_ptr.add(args.len() * j + i) = ch;
+                    }
+                }
+            }
+        }
+    }
+    let offsets = (0..(input_rows + 1) as u64 * args.len() as u64)
+        .step_by(args.len())
+        .collect::<Vec<_>>();
+    let result = BinaryColumn::new(values.into(), offsets.into());
+
+    match len {
+        Some(_) => Value::Column(Column::Binary(result)),
+        _ => Value::Scalar(Scalar::Binary(result.index(0).unwrap().to_vec())),
+    }
+}
diff --git a/src/query/functions/src/scalars/bitmap.rs b/src/query/functions/src/scalars/bitmap.rs
index 46927c8fbeb8..6adc7eb2245a 100644
--- a/src/query/functions/src/scalars/bitmap.rs
+++ b/src/query/functions/src/scalars/bitmap.rs
@@ -45,7 +45,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         "to_bitmap",
         |_, _| FunctionDomain::MayThrow,
         vectorize_with_builder_1_arg::<StringType, BitmapType>(|s, builder, ctx| {
-            match parse_bitmap(s) {
+            match parse_bitmap(s.as_bytes()) {
                 Ok(rb) => {
                     rb.serialize_into(&mut builder.data).unwrap();
                 }
diff --git a/src/query/functions/src/scalars/boolean.rs b/src/query/functions/src/scalars/boolean.rs
index e364bdddd2ab..8f28f5d8b5fc 100644
--- a/src/query/functions/src/scalars/boolean.rs
+++ b/src/query/functions/src/scalars/boolean.rs
@@ -462,9 +462,9 @@ fn eval_boolean_to_string(val: ValueRef<BooleanType>, ctx: &mut EvalContext) ->
 
 fn eval_string_to_boolean(val: ValueRef<StringType>, ctx: &mut EvalContext) -> Value<BooleanType> {
     vectorize_with_builder_1_arg::<StringType, BooleanType>(|val, output, ctx| {
-        if val.eq_ignore_ascii_case(b"true") {
+        if val.eq_ignore_ascii_case("true") {
             output.push(true);
-        } else if val.eq_ignore_ascii_case(b"false") {
+        } else if val.eq_ignore_ascii_case("false") {
             output.push(false);
         } else {
             ctx.set_error(output.len(), "cannot parse to type `BOOLEAN`");
diff --git a/src/query/functions/src/scalars/comparison.rs b/src/query/functions/src/scalars/comparison.rs
index e4d04f7f3661..adfb70674759 100644
--- a/src/query/functions/src/scalars/comparison.rs
+++ b/src/query/functions/src/scalars/comparison.rs
@@ -48,7 +48,7 @@ use databend_common_expression::SimpleDomainCmp;
 use databend_common_expression::ValueRef;
 use memchr::memchr;
 use memchr::memmem;
-use regex::bytes::Regex;
+use regex::Regex;
 
 use crate::scalars::decimal::register_decimal_compare_op;
 use crate::scalars::string_multi_args::regexp;
@@ -495,38 +495,29 @@ fn register_like(registry: &mut FunctionRegistry) {
         "like",
         |_, lhs, rhs| {
             if rhs.max.as_ref() == Some(&rhs.min) {
-                let pattern_type = check_pattern_type(&rhs.min, false);
-                if pattern_type == PatternType::EndOfPercent
-                    || pattern_type == PatternType::OrdinalStr
-                {
-                    let (min, max) = if pattern_type == PatternType::EndOfPercent {
-                        let min = rhs.min[..rhs.min.len() - 1].to_vec();
-                        let mut max = min.clone();
+                let pattern_type = check_pattern_type(rhs.min.as_bytes(), false);
 
-                        let l = max.len();
-                        if max[l - 1] != u8::MAX {
-                            max[l - 1] += 1;
-                        } else {
-                            return FunctionDomain::Full;
-                        }
-                        (min, max)
-                    } else {
-                        (rhs.min.clone(), rhs.min.clone())
-                    };
+                if pattern_type == PatternType::OrdinalStr {
+                    return lhs.domain_eq(rhs);
+                }
 
+                if pattern_type == PatternType::EndOfPercent {
+                    let mut pat_str = rhs.min.clone();
+                    // remove the last char '%'
+                    pat_str.pop();
+                    let pat_len = pat_str.chars().count();
                     let other = StringDomain {
-                        min,
-                        max: Some(max),
+                        min: pat_str.clone(),
+                        max: Some(pat_str),
                     };
-                    let gte = lhs.domain_gte(&other);
-                    let lt = lhs.domain_lt(&other);
-
-                    if let (FunctionDomain::Domain(lhs), FunctionDomain::Domain(rhs)) = (lt, gte) {
-                        return FunctionDomain::Domain(BooleanDomain {
-                            has_false: lhs.has_false || rhs.has_false,
-                            has_true: lhs.has_true && rhs.has_true,
-                        });
-                    }
+                    let lhs = StringDomain {
+                        min: lhs.min.chars().take(pat_len).collect(),
+                        max: lhs
+                            .max
+                            .as_ref()
+                            .map(|max| max.chars().take(pat_len).collect()),
+                    };
+                    return lhs.domain_eq(&other);
                 }
             }
             FunctionDomain::Full
@@ -572,7 +563,7 @@ fn register_like(registry: &mut FunctionRegistry) {
                 match regexp::build_regexp_from_pattern("regexp", pat, None) {
                     Ok(re) => {
                         builder.push(re.is_match(str));
-                        map.insert(pat.to_vec(), re);
+                        map.insert(pat.to_string(), re);
                     }
                     Err(e) => {
                         ctx.set_error(builder.len(), e);
@@ -590,24 +581,24 @@ fn vectorize_like(
 {
     move |arg1, arg2, ctx| match (arg1, arg2) {
         (ValueRef::Scalar(arg1), ValueRef::Scalar(arg2)) => {
-            let pattern_type = check_pattern_type(arg2, false);
-            Value::Scalar(func(arg1, arg2, ctx, &pattern_type))
+            let pattern_type = check_pattern_type(arg2.as_bytes(), false);
+            Value::Scalar(func(arg1.as_bytes(), arg2.as_bytes(), ctx, &pattern_type))
         }
         (ValueRef::Column(arg1), ValueRef::Scalar(arg2)) => {
             let arg1_iter = StringType::iter_column(&arg1);
 
-            let pattern_type = check_pattern_type(arg2, false);
+            let pattern_type = check_pattern_type(arg2.as_bytes(), false);
             // faster path for memmem to have a single instance of Finder
             if pattern_type == PatternType::SurroundByPercent && arg2.len() > 2 {
                 let finder = memmem::Finder::new(&arg2[1..arg2.len() - 1]);
-                let it = arg1_iter.map(|arg1| finder.find(arg1).is_some());
+                let it = arg1_iter.map(|arg1| finder.find(arg1.as_bytes()).is_some());
                 let bitmap = BooleanType::column_from_iter(it, &[]);
                 return Value::Column(bitmap);
             }
 
             let mut builder = MutableBitmap::with_capacity(arg1.len());
             for arg1 in arg1_iter {
-                builder.push(func(arg1, arg2, ctx, &pattern_type));
+                builder.push(func(arg1.as_bytes(), arg2.as_bytes(), ctx, &pattern_type));
             }
             Value::Column(builder.into())
         }
@@ -615,8 +606,8 @@ fn vectorize_like(
             let arg2_iter = StringType::iter_column(&arg2);
             let mut builder = MutableBitmap::with_capacity(arg2.len());
             for arg2 in arg2_iter {
-                let pattern_type = check_pattern_type(arg2, false);
-                builder.push(func(arg1, arg2, ctx, &pattern_type));
+                let pattern_type = check_pattern_type(arg2.as_bytes(), false);
+                builder.push(func(arg1.as_bytes(), arg2.as_bytes(), ctx, &pattern_type));
             }
             Value::Column(builder.into())
         }
@@ -625,8 +616,8 @@ fn vectorize_like(
             let arg2_iter = StringType::iter_column(&arg2);
             let mut builder = MutableBitmap::with_capacity(arg2.len());
             for (arg1, arg2) in arg1_iter.zip(arg2_iter) {
-                let pattern_type = check_pattern_type(arg2, false);
-                builder.push(func(arg1, arg2, ctx, &pattern_type));
+                let pattern_type = check_pattern_type(arg2.as_bytes(), false);
+                builder.push(func(arg1.as_bytes(), arg2.as_bytes(), ctx, &pattern_type));
             }
             Value::Column(builder.into())
         }
@@ -639,13 +630,13 @@ fn variant_vectorize_like(
 {
     move |arg1, arg2, ctx| match (arg1, arg2) {
         (ValueRef::Scalar(arg1), ValueRef::Scalar(arg2)) => {
-            let pattern_type = check_pattern_type(arg2, false);
-            Value::Scalar(func(arg1, arg2, ctx, &pattern_type))
+            let pattern_type = check_pattern_type(arg2.as_bytes(), false);
+            Value::Scalar(func(arg1, arg2.as_bytes(), ctx, &pattern_type))
         }
         (ValueRef::Column(arg1), ValueRef::Scalar(arg2)) => {
             let arg1_iter = VariantType::iter_column(&arg1);
 
-            let pattern_type = check_pattern_type(arg2, false);
+            let pattern_type = check_pattern_type(arg2.as_bytes(), false);
             // faster path for memmem to have a single instance of Finder
             if pattern_type == PatternType::SurroundByPercent && arg2.len() > 2 {
                 let finder = memmem::Finder::new(&arg2[1..arg2.len() - 1]);
@@ -656,7 +647,7 @@ fn variant_vectorize_like(
 
             let mut builder = MutableBitmap::with_capacity(arg1.len());
             for arg1 in arg1_iter {
-                builder.push(func(arg1, arg2, ctx, &pattern_type));
+                builder.push(func(arg1, arg2.as_bytes(), ctx, &pattern_type));
             }
             Value::Column(builder.into())
         }
@@ -664,8 +655,8 @@ fn variant_vectorize_like(
             let arg2_iter = StringType::iter_column(&arg2);
             let mut builder = MutableBitmap::with_capacity(arg2.len());
             for arg2 in arg2_iter {
-                let pattern_type = check_pattern_type(arg2, false);
-                builder.push(func(arg1, arg2, ctx, &pattern_type));
+                let pattern_type = check_pattern_type(arg2.as_bytes(), false);
+                builder.push(func(arg1, arg2.as_bytes(), ctx, &pattern_type));
             }
             Value::Column(builder.into())
         }
@@ -674,8 +665,8 @@ fn variant_vectorize_like(
             let arg2_iter = StringType::iter_column(&arg2);
             let mut builder = MutableBitmap::with_capacity(arg2.len());
             for (arg1, arg2) in arg1_iter.zip(arg2_iter) {
-                let pattern_type = check_pattern_type(arg2, false);
-                builder.push(func(arg1, arg2, ctx, &pattern_type));
+                let pattern_type = check_pattern_type(arg2.as_bytes(), false);
+                builder.push(func(arg1, arg2.as_bytes(), ctx, &pattern_type));
             }
             Value::Column(builder.into())
         }
@@ -684,11 +675,11 @@ fn variant_vectorize_like(
 
 fn vectorize_regexp(
     func: impl Fn(
-        &[u8],
-        &[u8],
+        &str,
+        &str,
         &mut MutableBitmap,
         &mut EvalContext,
-        &mut HashMap<Vec<u8>, Regex>,
+        &mut HashMap<String, Regex>,
         &mut HashMap<Vec<u8>, String>,
     ) + Copy,
 ) -> impl Fn(ValueRef<StringType>, ValueRef<StringType>, &mut EvalContext) -> Value<BooleanType> + Copy
diff --git a/src/query/functions/src/scalars/datetime.rs b/src/query/functions/src/scalars/datetime.rs
index c7e5b3de9fd4..9116b7478fc7 100644
--- a/src/query/functions/src/scalars/datetime.rs
+++ b/src/query/functions/src/scalars/datetime.rs
@@ -162,20 +162,11 @@ fn register_string_to_timestamp(registry: &mut FunctionRegistry) {
                 if format.is_empty() {
                     output.push_null();
                 } else {
-                    match (std::str::from_utf8(timestamp), std::str::from_utf8(format)) {
-                        (Ok(date), Ok(format)) => {
-                            // date need has timezone info.
-                            if let Ok(res) = DateTime::parse_from_str(date, format) {
-                                output.push(
-                                    res.with_timezone(&ctx.func_ctx.tz.tz).timestamp_micros(),
-                                );
-                            } else {
-                                output.push_null();
-                            }
-                        }
-                        _ => {
-                            output.push_null();
-                        }
+                    // date need has timezone info.
+                    if let Ok(res) = DateTime::parse_from_str(timestamp, format) {
+                        output.push(res.with_timezone(&ctx.func_ctx.tz.tz).timestamp_micros());
+                    } else {
+                        output.push_null();
                     }
                 }
             },
@@ -190,17 +181,11 @@ fn register_string_to_timestamp(registry: &mut FunctionRegistry) {
                 if format.is_empty() {
                     output.push_null();
                 } else {
-                    match (std::str::from_utf8(date), std::str::from_utf8(format)) {
-                        (Ok(date), Ok(format)) => match NaiveDate::parse_from_str(date, format) {
-                            Ok(res) => {
-                                output.push(res.num_days_from_ce() - EPOCH_DAYS_FROM_CE);
-                            }
-                            Err(e) => {
-                                ctx.set_error(output.len(), e.to_string());
-                                output.push_null();
-                            }
-                        },
-                        (Err(e), _) | (_, Err(e)) => {
+                    match NaiveDate::parse_from_str(date, format) {
+                        Ok(res) => {
+                            output.push(res.num_days_from_ce() - EPOCH_DAYS_FROM_CE);
+                        }
+                        Err(e) => {
                             ctx.set_error(output.len(), e.to_string());
                             output.push_null();
                         }
@@ -420,16 +405,8 @@ fn register_to_string(registry: &mut FunctionRegistry) {
                     output.push_null();
                 } else {
                     let ts = date.to_timestamp(ctx.func_ctx.tz.tz);
-                    match std::str::from_utf8(format) {
-                        Ok(format) => {
-                            let res = ts.format(format).to_string();
-                            output.push(res.as_bytes());
-                        }
-                        Err(e) => {
-                            ctx.set_error(output.len(), e.to_string());
-                            output.push_null();
-                        }
-                    }
+                    let res = ts.format(format).to_string();
+                    output.push(&res);
                 }
             },
         ),
@@ -464,7 +441,7 @@ fn register_to_string(registry: &mut FunctionRegistry) {
             FunctionDomain::Domain(NullableDomain {
                 has_null: false,
                 value: Some(Box::new(StringDomain {
-                    min: vec![],
+                    min: "".to_string(),
                     max: None,
                 })),
             })
@@ -487,7 +464,7 @@ fn register_to_string(registry: &mut FunctionRegistry) {
             FunctionDomain::Domain(NullableDomain {
                 has_null: false,
                 value: Some(Box::new(StringDomain {
-                    min: vec![],
+                    min: "".to_string(),
                     max: None,
                 })),
             })
diff --git a/src/query/functions/src/scalars/decimal/arithmetic.rs b/src/query/functions/src/scalars/decimal/arithmetic.rs
index ba9dfb7ff1c8..7bc9b63b20b9 100644
--- a/src/query/functions/src/scalars/decimal/arithmetic.rs
+++ b/src/query/functions/src/scalars/decimal/arithmetic.rs
@@ -28,6 +28,9 @@ use databend_common_expression::FunctionRegistry;
 use databend_common_expression::FunctionSignature;
 use ethnum::i256;
 
+use super::convert_to_decimal;
+use super::convert_to_decimal_domain;
+
 #[derive(Copy, Clone, Debug)]
 enum ArithmeticOp {
     Plus,
@@ -267,16 +270,13 @@ macro_rules! register_decimal_binary_op {
             let function = Function {
                 signature: FunctionSignature {
                     name: format!("{:?}", $arithmetic_op).to_lowercase(),
-                    args_type: vec![
-                        DataType::Decimal(left.clone()),
-                        DataType::Decimal(right.clone()),
-                    ],
+                    args_type: args_type.clone(),
                     return_type: DataType::Decimal(return_decimal_type),
                 },
                 eval: FunctionEval::Scalar {
-                    calc_domain: Box::new(move |_ctx, d| {
-                        let lhs = d[0].as_decimal();
-                        let rhs = d[1].as_decimal();
+                    calc_domain: Box::new(move |ctx, d| {
+                        let lhs = convert_to_decimal_domain(ctx, d[0].clone(), left.clone());
+                        let rhs = convert_to_decimal_domain(ctx, d[1].clone(), right.clone());
 
                         if lhs.is_none() || rhs.is_none() {
                             return FunctionDomain::Full;
@@ -308,9 +308,15 @@ macro_rules! register_decimal_binary_op {
                         .unwrap_or($default_domain)
                     }),
                     eval: Box::new(move |args, ctx| {
+                        let a = convert_to_decimal(&args[0], ctx, &args_type[0], left);
+                        let b = convert_to_decimal(&args[1], ctx, &args_type[1], right);
+
+                        let a = a.as_ref();
+                        let b = b.as_ref();
+
                         let res = op_decimal!(
-                            &args[0],
-                            &args[1],
+                            &a,
+                            &b,
                             ctx,
                             left,
                             right,
diff --git a/src/query/functions/src/scalars/decimal/cast.rs b/src/query/functions/src/scalars/decimal/cast.rs
index e1c4e9b3168a..53e6a20fd215 100644
--- a/src/query/functions/src/scalars/decimal/cast.rs
+++ b/src/query/functions/src/scalars/decimal/cast.rs
@@ -328,7 +328,7 @@ fn decimal_to_string(
     })
 }
 
-fn convert_to_decimal(
+pub fn convert_to_decimal(
     arg: &ValueRef<AnyType>,
     ctx: &mut EvalContext,
     from_type: &DataType,
@@ -387,7 +387,7 @@ fn convert_to_decimal(
     })
 }
 
-fn convert_to_decimal_domain(
+pub fn convert_to_decimal_domain(
     func_ctx: &FunctionContext,
     domain: Domain,
     dest_type: DecimalDataType,
@@ -468,14 +468,16 @@ fn string_to_decimal<T: Decimal>(
 where
     T: Decimal + Mul<Output = T>,
 {
-    let f = |x: &[u8], builder: &mut Vec<T>, ctx: &mut EvalContext| {
-        let value = match read_decimal_with_size::<T>(x, size, true, ctx.func_ctx.rounding_mode) {
-            Ok((d, _)) => d,
-            Err(e) => {
-                ctx.set_error(builder.len(), e.message());
-                T::zero()
-            }
-        };
+    let f = |x: &str, builder: &mut Vec<T>, ctx: &mut EvalContext| {
+        let value =
+            match read_decimal_with_size::<T>(x.as_bytes(), size, true, ctx.func_ctx.rounding_mode)
+            {
+                Ok((d, _)) => d,
+                Err(e) => {
+                    ctx.set_error(builder.len(), e.message());
+                    T::zero()
+                }
+            };
 
         builder.push(value);
     };
diff --git a/src/query/functions/src/scalars/decimal/mod.rs b/src/query/functions/src/scalars/decimal/mod.rs
index c55ea2d36f56..ec533ecd828d 100644
--- a/src/query/functions/src/scalars/decimal/mod.rs
+++ b/src/query/functions/src/scalars/decimal/mod.rs
@@ -18,6 +18,8 @@ mod comparison;
 mod math;
 
 pub(crate) use arithmetic::register_decimal_arithmetic;
+pub(crate) use cast::convert_to_decimal;
+pub(crate) use cast::convert_to_decimal_domain;
 pub(crate) use cast::register_decimal_to_float;
 pub(crate) use cast::register_decimal_to_int;
 pub(crate) use cast::register_decimal_to_string;
diff --git a/src/query/functions/src/scalars/geo.rs b/src/query/functions/src/scalars/geo.rs
index 3a2b5c084610..83ecdd1464f0 100644
--- a/src/query/functions/src/scalars/geo.rs
+++ b/src/query/functions/src/scalars/geo.rs
@@ -183,13 +183,10 @@ pub fn register(registry: &mut FunctionRegistry) {
             "geohash_decode",
             |_, _| FunctionDomain::Full,
             vectorize_with_builder_1_arg::<StringType, KvPair<Float64Type, Float64Type>>(
-                |encoded, builder, ctx| match std::str::from_utf8(encoded)
-                    .map_err(|e| e.to_string())
-                    .and_then(|s| geohash::decode(s).map_err(|e| e.to_string()))
-                {
+                |encoded, builder, ctx| match geohash::decode(encoded) {
                     Ok((c, _, _)) => builder.push((c.x.into(), c.y.into())),
                     Err(e) => {
-                        ctx.set_error(builder.len(), e);
+                        ctx.set_error(builder.len(), e.to_string());
                         builder.push((F64::from(0.0), F64::from(0.0)))
                     }
                 },
diff --git a/src/query/functions/src/scalars/geo_h3.rs b/src/query/functions/src/scalars/geo_h3.rs
index b99d7b258b35..045327555552 100644
--- a/src/query/functions/src/scalars/geo_h3.rs
+++ b/src/query/functions/src/scalars/geo_h3.rs
@@ -12,8 +12,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::str;
-
 use databend_common_expression::types::map::KvPair;
 use databend_common_expression::types::ArrayType;
 use databend_common_expression::types::BooleanType;
@@ -288,13 +286,10 @@ pub fn register(registry: &mut FunctionRegistry) {
         "string_to_h3",
         |_, _| FunctionDomain::Full,
         vectorize_with_builder_1_arg::<StringType, UInt64Type>(|h3_str, builder, ctx| {
-            match str::from_utf8(h3_str)
-                .map_err(|e| e.to_string())
-                .and_then(|h3_str| str::parse::<CellIndex>(h3_str).map_err(|e| e.to_string()))
-            {
+            match h3_str.parse::<CellIndex>() {
                 Ok(index) => builder.push(index.into()),
                 Err(err) => {
-                    ctx.set_error(builder.len(), err);
+                    ctx.set_error(builder.len(), err.to_string());
                     builder.push(0);
                 }
             }
diff --git a/src/query/functions/src/scalars/hash.rs b/src/query/functions/src/scalars/hash.rs
index 66034e6d746a..152ee21b4f27 100644
--- a/src/query/functions/src/scalars/hash.rs
+++ b/src/query/functions/src/scalars/hash.rs
@@ -104,7 +104,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                 output.data.resize(old_len + 40, 0);
                 // TODO sha1 lib doesn't allow encode into buffer...
                 let mut m = ::sha1::Sha1::new();
-                sha1::digest::Update::update(&mut m, val);
+                sha1::digest::Update::update(&mut m, val.as_bytes());
 
                 if let Err(err) =
                     hex::encode_to_slice(m.finalize().as_slice(), &mut output.data[old_len..])
@@ -124,9 +124,10 @@ pub fn register(registry: &mut FunctionRegistry) {
             |val, output, ctx| {
                 let old_len = output.data.len();
                 output.data.resize(old_len + 64, 0);
-                if let Err(err) =
-                    hex::encode_to_slice(blake3::hash(val).as_bytes(), &mut output.data[old_len..])
-                {
+                if let Err(err) = hex::encode_to_slice(
+                    blake3::hash(val.as_bytes()).as_bytes(),
+                    &mut output.data[old_len..],
+                ) {
                     ctx.set_error(output.len(), err.to_string());
                 }
                 output.commit_row();
@@ -143,22 +144,22 @@ pub fn register(registry: &mut FunctionRegistry) {
                 let res = match l {
                     224 => {
                         let mut h = sha2::Sha224::new();
-                        sha2::digest::Update::update(&mut h, val);
+                        sha2::digest::Update::update(&mut h, val.as_bytes());
                         format!("{:x}", h.finalize())
                     }
                     256 | 0 => {
                         let mut h = sha2::Sha256::new();
-                        sha2::digest::Update::update(&mut h, val);
+                        sha2::digest::Update::update(&mut h, val.as_bytes());
                         format!("{:x}", h.finalize())
                     }
                     384 => {
                         let mut h = sha2::Sha384::new();
-                        sha2::digest::Update::update(&mut h, val);
+                        sha2::digest::Update::update(&mut h, val.as_bytes());
                         format!("{:x}", h.finalize())
                     }
                     512 => {
                         let mut h = sha2::Sha512::new();
-                        sha2::digest::Update::update(&mut h, val);
+                        sha2::digest::Update::update(&mut h, val.as_bytes());
                         format!("{:x}", h.finalize())
                     }
                     v => {
@@ -172,7 +173,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         String::new()
                     },
                 };
-                output.put_slice(res.as_bytes());
+                output.put_str(&res);
                 output.commit_row();
             },
         ),
@@ -340,6 +341,27 @@ impl<'a> DFHash for &'a [u8] {
     }
 }
 
+impl<'a> DFHash for &'a str {
+    #[inline]
+    fn hash<H: Hasher>(&self, state: &mut H) {
+        Hash::hash_slice(self.as_bytes(), state);
+    }
+}
+
+impl DFHash for [u8] {
+    #[inline]
+    fn hash<H: Hasher>(&self, state: &mut H) {
+        Hash::hash_slice(self, state);
+    }
+}
+
+impl DFHash for str {
+    #[inline]
+    fn hash<H: Hasher>(&self, state: &mut H) {
+        Hash::hash_slice(self.as_bytes(), state);
+    }
+}
+
 impl DFHash for bool {
     #[inline]
     fn hash<H: Hasher>(&self, state: &mut H) {
@@ -357,10 +379,11 @@ impl DFHash for Scalar {
                     DFHash::hash(v, state);
                 }
             }),
-            Scalar::Binary(vals) | Scalar::String(vals) | Scalar::Variant(vals) => {
-                for v in vals {
-                    DFHash::hash(v, state);
-                }
+            Scalar::Binary(vals) | Scalar::Variant(vals) => {
+                DFHash::hash(vals.as_slice(), state);
+            }
+            Scalar::String(vals) => {
+                DFHash::hash(vals.as_str(), state);
             }
             _ => {}
         }
diff --git a/src/query/functions/src/scalars/math.rs b/src/query/functions/src/scalars/math.rs
index a89953a91724..a81c592307ef 100644
--- a/src/query/functions/src/scalars/math.rs
+++ b/src/query/functions/src/scalars/math.rs
@@ -209,7 +209,7 @@ pub fn register(registry: &mut FunctionRegistry) {
     registry.register_1_arg::<StringType, NumberType<u32>, _, _>(
         "crc32",
         |_, _| FunctionDomain::Full,
-        |val, _| crc32fast::hash(val),
+        |val, _| crc32fast::hash(val.as_bytes()),
     );
 
     registry.register_1_arg::<NumberType<F64>, NumberType<F64>, _, _>(
diff --git a/src/query/functions/src/scalars/other.rs b/src/query/functions/src/scalars/other.rs
index 46c02e2ca2fa..e3abc326c4e1 100644
--- a/src/query/functions/src/scalars/other.rs
+++ b/src/query/functions/src/scalars/other.rs
@@ -164,7 +164,7 @@ pub fn register(registry: &mut FunctionRegistry) {
     registry.register_1_arg_core::<GenericType<0>, StringType, _, _>(
         "typeof",
         |_, _| FunctionDomain::Full,
-        |_, ctx| Value::Scalar(ctx.generics[0].sql_name().into_bytes()),
+        |_, ctx| Value::Scalar(ctx.generics[0].sql_name()),
     );
 
     registry.register_function_factory("ignore", |_, args_type| {
@@ -249,8 +249,7 @@ fn register_inet_aton(registry: &mut FunctionRegistry) {
     );
 
     fn eval_inet_aton(val: ValueRef<StringType>, ctx: &mut EvalContext) -> Value<UInt32Type> {
-        vectorize_with_builder_1_arg::<StringType, UInt32Type>(|v, output, ctx| {
-            let addr_str = String::from_utf8_lossy(v);
+        vectorize_with_builder_1_arg::<StringType, UInt32Type>(|addr_str, output, ctx| {
             match addr_str.parse::<Ipv4Addr>() {
                 Ok(addr) => {
                     let addr_binary = u32::from(addr);
diff --git a/src/query/functions/src/scalars/string.rs b/src/query/functions/src/scalars/string.rs
index d23c80839521..7a4c59586533 100644
--- a/src/query/functions/src/scalars/string.rs
+++ b/src/query/functions/src/scalars/string.rs
@@ -15,7 +15,6 @@
 use std::cmp::Ordering;
 use std::io::Write;
 
-use bstr::ByteSlice;
 use databend_common_base::base::uuid::Uuid;
 use databend_common_expression::types::decimal::Decimal128Type;
 use databend_common_expression::types::number::SimpleDomain;
@@ -34,16 +33,23 @@ use databend_common_expression::FunctionDomain;
 use databend_common_expression::FunctionRegistry;
 use databend_common_expression::Value;
 use databend_common_expression::ValueRef;
-use itertools::izip;
+use stringslice::StringSlice;
 
 pub fn register(registry: &mut FunctionRegistry) {
     registry.register_aliases("to_string", &["to_varchar", "to_text"]);
     registry.register_aliases("upper", &["ucase"]);
     registry.register_aliases("lower", &["lcase"]);
-    registry.register_aliases("length", &["octet_length"]);
-    registry.register_aliases("char_length", &["character_length", "length_utf8"]);
-    registry.register_aliases("substr", &["substring", "mid"]);
-    registry.register_aliases("substr_utf8", &["substring_utf8"]);
+    registry.register_aliases("length", &[
+        "char_length",
+        "character_length",
+        "length_utf8",
+    ]);
+    registry.register_aliases("substr", &[
+        "substring",
+        "mid",
+        "substr_utf8",
+        "substring_utf8",
+    ]);
 
     registry.register_passthrough_nullable_1_arg::<StringType, StringType, _, _>(
         "upper",
@@ -51,12 +57,9 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len(),
             |val, output, _| {
-                for (start, end, ch) in val.char_indices() {
-                    if ch == '\u{FFFD}' {
-                        // If char is invalid, just copy it.
-                        output.put_slice(&val.as_bytes()[start..end]);
-                    } else if ch.is_ascii() {
-                        output.put_u8(ch.to_ascii_uppercase() as u8);
+                for ch in val.chars() {
+                    if ch.is_ascii() {
+                        output.put_char(ch.to_ascii_uppercase());
                     } else {
                         for x in ch.to_uppercase() {
                             output.put_char(x);
@@ -74,12 +77,9 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len(),
             |val, output, _| {
-                for (start, end, ch) in val.char_indices() {
-                    if ch == '\u{FFFD}' {
-                        // If char is invalid, just copy it.
-                        output.put_slice(&val.as_bytes()[start..end]);
-                    } else if ch.is_ascii() {
-                        output.put_u8(ch.to_ascii_lowercase() as u8);
+                for ch in val.chars() {
+                    if ch.is_ascii() {
+                        output.put_char(ch.to_ascii_lowercase());
                     } else {
                         for x in ch.to_lowercase() {
                             output.put_char(x);
@@ -98,7 +98,7 @@ pub fn register(registry: &mut FunctionRegistry) {
     );
 
     registry.register_passthrough_nullable_1_arg::<StringType, NumberType<u64>, _, _>(
-        "length",
+        "octet_length",
         |_, _| FunctionDomain::Full,
         |val, _| match val {
             ValueRef::Scalar(s) => Value::Scalar(s.len() as u64),
@@ -115,20 +115,10 @@ pub fn register(registry: &mut FunctionRegistry) {
         },
     );
 
-    registry.register_passthrough_nullable_1_arg::<StringType, NumberType<u64>, _, _>(
-        "char_length",
-        |_, _| FunctionDomain::MayThrow,
-        vectorize_with_builder_1_arg::<StringType, NumberType<u64>>(|s, output, ctx| {
-            match std::str::from_utf8(s) {
-                Ok(s) => {
-                    output.push(s.chars().count() as u64);
-                }
-                Err(err) => {
-                    ctx.set_error(output.len(), err.to_string());
-                    output.push(0);
-                }
-            }
-        }),
+    registry.register_1_arg::<StringType, NumberType<u64>, _, _>(
+        "length",
+        |_, _| FunctionDomain::Full,
+        |val, _ctx| val.chars().count() as u64,
     );
 
     const MAX_PADDING_LENGTH: usize = 1000000;
@@ -138,24 +128,26 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_with_builder_3_arg::<StringType, NumberType<u64>, StringType, StringType>(
             |s, pad_len, pad, output, ctx| {
                 let pad_len = pad_len as usize;
+                let s_len = s.chars().count();
                 if pad_len > MAX_PADDING_LENGTH {
                     ctx.set_error(output.len(), format!("padding length '{}' is too big, max is: '{}'", pad_len, MAX_PADDING_LENGTH));
-                } else if pad_len <= s.len() {
-                    output.put_slice(&s[..pad_len]);
+                } else if pad_len <= s_len {
+                    output.put_str(s.slice(..pad_len));
                 } else if pad.is_empty() {
-                    ctx.set_error(output.len(), format!("can't fill the '{}' length to '{}' with an empty pad string", String::from_utf8_lossy(s), pad_len));
+                    ctx.set_error(output.len(), format!("can't fill the '{}' length to '{}' with an empty pad string", s, pad_len));
                 } else {
-                    let mut remain_pad_len = pad_len - s.len();
+                    let mut remain_pad_len = pad_len - s_len;
+                    let p_len = pad.chars().count();
                     while remain_pad_len > 0 {
-                        if remain_pad_len < pad.len() {
-                            output.put_slice(&pad[..remain_pad_len]);
+                        if remain_pad_len < p_len {
+                            output.put_str(pad.slice(..remain_pad_len));
                             remain_pad_len = 0;
                         } else {
-                            output.put_slice(pad);
-                            remain_pad_len -= pad.len();
+                            output.put_str(pad);
+                            remain_pad_len -= p_len;
                         }
                     }
-                    output.put_slice(s);
+                    output.put_str(s);
                 }
                 output.commit_row();
             }
@@ -169,14 +161,15 @@ pub fn register(registry: &mut FunctionRegistry) {
             |srcstr, pos, len, substr, output, _| {
                 let pos = pos as usize;
                 let len = len as usize;
-                if pos < 1 || pos > srcstr.len() {
-                    output.put_slice(srcstr);
+                let srcstr_len = srcstr.chars().count();
+                if pos < 1 || pos > srcstr_len {
+                    output.put_str(srcstr);
                 } else {
                     let pos = pos - 1;
-                    output.put_slice(&srcstr[0..pos]);
-                    output.put_slice(substr);
-                    if pos + len < srcstr.len() {
-                        output.put_slice(&srcstr[(pos + len)..]);
+                    output.put_str(srcstr.slice(0..pos));
+                    output.put_str(substr);
+                    if pos + len < srcstr_len {
+                        output.put_str(srcstr.slice(pos + len .. ));
                     }
                 }
                 output.commit_row();
@@ -187,29 +180,31 @@ pub fn register(registry: &mut FunctionRegistry) {
         "rpad",
         |_, _, _, _| FunctionDomain::MayThrow,
         vectorize_with_builder_3_arg::<StringType, NumberType<u64>, StringType, StringType>(
-        |s: &[u8], pad_len: u64, pad: &[u8], output, ctx| {
-            let pad_len = pad_len as usize;
-            if pad_len > MAX_PADDING_LENGTH {
-                ctx.set_error(output.len(), format!("padding length '{}' is too big, max is: '{}'", pad_len, MAX_PADDING_LENGTH));
-            } else if pad_len <= s.len() {
-                output.put_slice(&s[..pad_len])
-            } else if pad.is_empty() {
-                ctx.set_error(output.len(), format!("can't fill the '{}' length to '{}' with an empty pad string", String::from_utf8_lossy(s), pad_len));
-            } else {
-                output.put_slice(s);
-                let mut remain_pad_len = pad_len - s.len();
-                while remain_pad_len > 0 {
-                    if remain_pad_len < pad.len() {
-                        output.put_slice(&pad[..remain_pad_len]);
-                        remain_pad_len = 0;
-                    } else {
-                        output.put_slice(pad);
-                        remain_pad_len -= pad.len();
+            |s, pad_len, pad, output, ctx| {
+                let pad_len = pad_len as usize;
+                let s_len = s.chars().count();
+                if pad_len > MAX_PADDING_LENGTH {
+                    ctx.set_error(output.len(), format!("padding length '{}' is too big, max is: '{}'", pad_len, MAX_PADDING_LENGTH));
+                } else if pad_len <= s_len {
+                    output.put_str(s.slice(..pad_len));
+                } else if pad.is_empty() {
+                    ctx.set_error(output.len(), format!("can't fill the '{}' length to '{}' with an empty pad string", s, pad_len));
+                } else {
+                    output.put_str(s);
+                    let mut remain_pad_len = pad_len - s_len;
+                    let p_len = pad.chars().count();
+                    while remain_pad_len > 0 {
+                        if remain_pad_len < p_len {
+                            output.put_str(pad.slice(..remain_pad_len));
+                            remain_pad_len = 0;
+                        } else {
+                            output.put_str(pad);
+                            remain_pad_len -= p_len;
+                        }
                     }
                 }
-            }
-            output.commit_row();
-        }),
+                output.commit_row();
+            }),
     );
 
     registry.register_passthrough_nullable_3_arg::<StringType, StringType, StringType, StringType, _, _>(
@@ -217,26 +212,22 @@ pub fn register(registry: &mut FunctionRegistry) {
         |_, _, _, _| FunctionDomain::Full,
         vectorize_with_builder_3_arg::<StringType, StringType, StringType, StringType>(
             |str, from, to, output, _| {
-            if from.is_empty() || from == to {
-                output.put_slice(str);
-                output.commit_row();
-                return;
-            }
-            let mut skip = 0;
-            for (p, w) in str.windows(from.len()).enumerate() {
-                if w == from {
-                    output.put_slice(to);
-                    skip = from.len();
-                } else if p + w.len() == str.len() {
-                    output.put_slice(w);
-                } else if skip > 1 {
-                    skip -= 1;
-                } else {
-                    output.put_slice(&w[0..1]);
+                if from.is_empty() || from == to {
+                    output.put_str(str);
+                    output.commit_row();
+                    return;
                 }
-            }
-            output.commit_row();
-        }),
+
+                let mut last_end = 0;
+                for (start, _) in str.match_indices(from) {
+                    output.put_str(&str[last_end..start]);
+                    output.put_str(to);
+                    last_end = start + from.len();
+                }
+                output.put_str(&str[last_end..]);
+
+                output.commit_row();
+            }),
     );
 
     registry.register_passthrough_nullable_3_arg::<StringType, StringType, StringType, StringType, _, _>(
@@ -244,23 +235,23 @@ pub fn register(registry: &mut FunctionRegistry) {
         |_, _, _, _| FunctionDomain::Full,
         vectorize_with_builder_3_arg::<StringType, StringType, StringType, StringType>(
             |str, from, to, output, _| {
-            if from.is_empty() || from == to {
-                output.put_slice(str);
-                output.commit_row();
-                return;
-            }
-            let to_len = to.len();
-            str.iter().for_each(|x| {
-                if let Some(index) = from.find([*x]) {
-                    if index < to_len {
-                        output.put_u8(to[index]);
-                    }
-                } else {
-                    output.put_u8(*x);
+                if from.is_empty() || from == to {
+                    output.put_str(str);
+                    output.commit_row();
+                    return;
                 }
-            });
-            output.commit_row();
-        }),
+                let to_len = to.chars().count();
+                str.chars().for_each(|x| {
+                    if let Some(index) = from.chars().position(|c| c == x) {
+                        if index < to_len {
+                            output.put_char(to.chars().nth(index).unwrap());
+                        }
+                    } else {
+                        output.put_char(x);
+                    }
+                });
+                output.commit_row();
+            }),
     );
 
     registry.register_passthrough_nullable_1_arg::<Decimal128Type, StringType, _, _>(
@@ -269,7 +260,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_with_builder_1_arg::<Decimal128Type, StringType>(|arg, output, _| {
             let uuid = Uuid::from_u128(arg as u128);
             let str = uuid.as_simple().to_string();
-            output.put_slice(str.as_bytes());
+            output.put_str(str.as_str());
             output.commit_row();
         }),
     );
@@ -277,32 +268,14 @@ pub fn register(registry: &mut FunctionRegistry) {
     registry.register_2_arg::<StringType, StringType, NumberType<i8>, _, _>(
         "strcmp",
         |_, _, _| FunctionDomain::Full,
-        |s1, s2, _| {
-            let res = match s1.len().cmp(&s2.len()) {
-                Ordering::Equal => {
-                    let mut res = Ordering::Equal;
-                    for (s1i, s2i) in izip!(s1, s2) {
-                        match s1i.cmp(s2i) {
-                            Ordering::Equal => continue,
-                            ord => {
-                                res = ord;
-                                break;
-                            }
-                        }
-                    }
-                    res
-                }
-                ord => ord,
-            };
-            match res {
-                Ordering::Equal => 0,
-                Ordering::Greater => 1,
-                Ordering::Less => -1,
-            }
+        |s1, s2, _| match s1.cmp(s2) {
+            Ordering::Equal => 0,
+            Ordering::Greater => 1,
+            Ordering::Less => -1,
         },
     );
 
-    let find_at = |str: &[u8], substr: &[u8], pos: u64| {
+    let find_at = |s: &str, substr: &str, pos: u64| {
         if substr.is_empty() {
             // the same behavior as MySQL, Postgres and Clickhouse
             return if pos == 0 { 1_u64 } else { pos };
@@ -313,37 +286,37 @@ pub fn register(registry: &mut FunctionRegistry) {
             return 0_u64;
         }
         let p = pos - 1;
-        if p + substr.len() <= str.len() {
-            str[p..]
-                .windows(substr.len())
-                .position(|w| w == substr)
-                .map_or(0, |i| i + 1 + p) as u64
+
+        let src = s.slice(p..);
+        if let Some(find_at) = src.find(substr) {
+            (src[..find_at].chars().count() + p + 1) as u64
         } else {
             0_u64
         }
     };
+
     registry.register_2_arg::<StringType, StringType, NumberType<u64>, _, _>(
         "instr",
         |_, _, _| FunctionDomain::Full,
-        move |str: &[u8], substr: &[u8], _| find_at(str, substr, 1),
+        move |s: &str, substr: &str, _| find_at(s, substr, 1),
     );
 
     registry.register_2_arg::<StringType, StringType, NumberType<u64>, _, _>(
         "position",
         |_, _, _| FunctionDomain::Full,
-        move |substr: &[u8], str: &[u8], _| find_at(str, substr, 1),
+        move |substr: &str, s: &str, _| find_at(s, substr, 1),
     );
 
     registry.register_2_arg::<StringType, StringType, NumberType<u64>, _, _>(
         "locate",
         |_, _, _| FunctionDomain::Full,
-        move |substr: &[u8], str: &[u8], _| find_at(str, substr, 1),
+        move |substr: &str, s: &str, _| find_at(s, substr, 1),
     );
 
     registry.register_3_arg::<StringType, StringType, NumberType<u64>, NumberType<u64>, _, _>(
         "locate",
         |_, _, _, _| FunctionDomain::Full,
-        move |substr: &[u8], str: &[u8], pos: u64, _| find_at(str, substr, pos),
+        move |substr: &str, s: &str, pos: u64, _| find_at(s, substr, pos),
     );
 
     registry.register_passthrough_nullable_1_arg::<StringType, StringType, _, _>(
@@ -352,17 +325,17 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len() * 2,
             |val, output, _| {
-                for ch in val {
+                for ch in val.chars() {
                     match ch {
-                        0 => output.put_slice(&[b'\\', b'0']),
-                        b'\'' => output.put_slice(&[b'\\', b'\'']),
-                        b'\"' => output.put_slice(&[b'\\', b'\"']),
-                        8 => output.put_slice(&[b'\\', b'b']),
-                        b'\n' => output.put_slice(&[b'\\', b'n']),
-                        b'\r' => output.put_slice(&[b'\\', b'r']),
-                        b'\t' => output.put_slice(&[b'\\', b't']),
-                        b'\\' => output.put_slice(&[b'\\', b'\\']),
-                        c => output.put_u8(*c),
+                        '\0' => output.put_str("\\0"),
+                        '\'' => output.put_str("\\\'"),
+                        '\"' => output.put_str("\\\""),
+                        '\u{8}' => output.put_str("\\b"),
+                        '\n' => output.put_str("\\n"),
+                        '\r' => output.put_str("\\r"),
+                        '\t' => output.put_str("\\t"),
+                        '\\' => output.put_str("\\\\"),
+                        c => output.put_char(c),
                     }
                 }
                 output.commit_row();
@@ -376,10 +349,9 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len(),
             |val, output, _| {
-                let start = output.data.len();
-                output.put_slice(val);
-                let buf = &mut output.data[start..];
-                buf.reverse();
+                for char in val.chars().rev() {
+                    output.put_char(char);
+                }
                 output.commit_row();
             },
         ),
@@ -389,14 +361,15 @@ pub fn register(registry: &mut FunctionRegistry) {
         "ascii",
         |_, domain| {
             FunctionDomain::Domain(SimpleDomain {
-                min: domain.min.first().cloned().unwrap_or(0),
+                min: domain.min.as_bytes().first().map_or(0, |v| *v),
                 max: domain
                     .max
                     .as_ref()
-                    .map_or(u8::MAX, |v| v.first().cloned().unwrap_or_default()),
+                    .and_then(|x| x.as_bytes().first())
+                    .map_or(u8::MAX, |v| *v),
             })
         },
-        |val, _| val.first().cloned().unwrap_or_default(),
+        |val, _| val.as_bytes().first().map_or(0, |v| *v),
     );
 
     // Trim functions
@@ -406,10 +379,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len(),
             |val, output, _| {
-                let pos = val.iter().position(|ch| *ch != b' ' && *ch != b'\t');
-                if let Some(idx) = pos {
-                    output.put_slice(&val.as_bytes()[idx..]);
-                }
+                output.put_str(val.trim_start());
                 output.commit_row();
             },
         ),
@@ -421,10 +391,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len(),
             |val, output, _| {
-                let pos = val.iter().rev().position(|ch| *ch != b' ' && *ch != b'\t');
-                if let Some(idx) = pos {
-                    output.put_slice(&val.as_bytes()[..val.len() - idx]);
-                }
+                output.put_str(val.trim_end());
                 output.commit_row();
             },
         ),
@@ -436,11 +403,7 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string(
             |col| col.data().len(),
             |val, output, _| {
-                let start_pos = val.iter().position(|ch| *ch != b' ' && *ch != b'\t');
-                let end_pos = val.iter().rev().position(|ch| *ch != b' ' && *ch != b'\t');
-                if let (Some(start_idx), Some(end_idx)) = (start_pos, end_pos) {
-                    output.put_slice(&val.as_bytes()[start_idx..val.len() - end_idx]);
-                }
+                output.put_str(val.trim());
                 output.commit_row();
             },
         ),
@@ -452,17 +415,13 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string_2_arg(
             |col, _| col.data().len(),
             |val, trim_str, _, output| {
-                let chunk_size = trim_str.len();
-                if chunk_size == 0 {
-                    output.put_slice(val);
+                if trim_str.is_empty() {
+                    output.put_str(val);
                     output.commit_row();
                     return;
                 }
 
-                let pos = val.chunks(chunk_size).position(|chunk| chunk != trim_str);
-                if let Some(idx) = pos {
-                    output.put_slice(&val.as_bytes()[idx * chunk_size..]);
-                }
+                output.put_str(val.trim_start_matches(trim_str));
                 output.commit_row();
             },
         ),
@@ -474,17 +433,13 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string_2_arg(
             |col, _| col.data().len(),
             |val, trim_str, _, output| {
-                let chunk_size = trim_str.len();
-                if chunk_size == 0 {
-                    output.put_slice(val);
+                if trim_str.is_empty() {
+                    output.put_str(val);
                     output.commit_row();
                     return;
                 }
 
-                let pos = val.rchunks(chunk_size).position(|chunk| chunk != trim_str);
-                if let Some(idx) = pos {
-                    output.put_slice(&val.as_bytes()[..val.len() - idx * chunk_size]);
-                }
+                output.put_str(val.trim_end_matches(trim_str));
                 output.commit_row();
             },
         ),
@@ -496,29 +451,22 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_string_to_string_2_arg(
             |col, _| col.data().len(),
             |val, trim_str, _, output| {
-                let chunk_size = trim_str.len();
-                if chunk_size == 0 {
-                    output.put_slice(val);
+                if trim_str.is_empty() {
+                    output.put_str(val);
                     output.commit_row();
                     return;
                 }
 
-                let start_pos = val.chunks(chunk_size).position(|chunk| chunk != trim_str);
+                let mut res = val;
 
-                // Trim all
-                if start_pos.is_none() {
-                    output.commit_row();
-                    return;
+                while res.starts_with(trim_str) {
+                    res = &res[trim_str.len()..];
                 }
-
-                let end_pos = val.rchunks(chunk_size).position(|chunk| chunk != trim_str);
-
-                if let (Some(start_idx), Some(end_idx)) = (start_pos, end_pos) {
-                    output.put_slice(
-                        &val.as_bytes()[start_idx * chunk_size..val.len() - end_idx * chunk_size],
-                    );
+                while res.ends_with(trim_str) {
+                    res = &res[..res.len() - trim_str.len()];
                 }
 
+                output.put_str(res);
                 output.commit_row();
             },
         ),
@@ -581,7 +529,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         ),
                     );
                 } else {
-                    (0..times).for_each(|_| output.put_slice(a));
+                    (0..times).for_each(|_| output.put_str(a));
                 }
                 output.commit_row();
             },
@@ -591,7 +539,8 @@ pub fn register(registry: &mut FunctionRegistry) {
     registry.register_1_arg::<StringType, UInt64Type, _, _>(
         "ord",
         |_, _| FunctionDomain::Full,
-        |str: &[u8], _| {
+        |str: &str, _| {
+            let str = str.as_bytes();
             let mut res: u64 = 0;
             if !str.is_empty() {
                 if str[0].is_ascii() {
@@ -621,7 +570,6 @@ pub fn register(registry: &mut FunctionRegistry) {
         ),
     );
 
-    const SPACE: u8 = 0x20;
     const MAX_SPACE_LENGTH: u64 = 1000000;
     registry.register_passthrough_nullable_1_arg::<NumberType<u64>, StringType, _, _>(
         "space",
@@ -633,9 +581,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                         0,
                         format!("space length is too big, max is: {}", MAX_SPACE_LENGTH),
                     );
-                    Value::Scalar(vec![])
+                    Value::Scalar("".to_string())
                 } else {
-                    Value::Scalar(vec![SPACE; times as usize])
+                    Value::Scalar(" ".repeat(times as usize))
                 }
             }
             ValueRef::Column(col) => {
@@ -658,7 +606,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                     total_space = 0;
                 }
                 let col = StringColumnBuilder {
-                    data: vec![SPACE; total_space as usize],
+                    data: " ".repeat(total_space as usize).into_bytes(),
                     offsets,
                     need_estimated: false,
                 }
@@ -674,10 +622,11 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_with_builder_2_arg::<StringType, NumberType<u64>, StringType>(
             |s, n, output, _| {
                 let n = n as usize;
-                if n < s.len() {
-                    output.put_slice(&s[0..n]);
+                let s_len = s.chars().count();
+                if n < s_len {
+                    output.put_str(s.slice(0..n));
                 } else {
-                    output.put_slice(s);
+                    output.put_str(s);
                 }
                 output.commit_row();
             },
@@ -690,10 +639,11 @@ pub fn register(registry: &mut FunctionRegistry) {
         vectorize_with_builder_2_arg::<StringType, NumberType<u64>, StringType>(
             |s, n, output, _| {
                 let n = n as usize;
-                if n < s.len() {
-                    output.put_slice(&s[s.len() - n..]);
+                let s_len = s.chars().count();
+                if n < s_len {
+                    output.put_str(s.slice(s_len - n..));
                 } else {
-                    output.put_slice(s);
+                    output.put_str(s);
                 }
                 output.commit_row();
             },
@@ -704,82 +654,38 @@ pub fn register(registry: &mut FunctionRegistry) {
         "substr",
         |_, _, _| FunctionDomain::Full,
         vectorize_with_builder_2_arg::<StringType, NumberType<i64>, StringType>(
-            |s, pos, output, _| {
-                output.put_slice(substr(s, pos, s.len() as u64));
-                output.commit_row();
+            |s, pos, output, _ctx| {
+                substr(output, s, pos, s.len() as u64);
             },
         ),
     );
 
     registry.register_passthrough_nullable_3_arg::<StringType, NumberType<i64>, NumberType<u64>, StringType, _, _>(
         "substr",
-        |_, _, _, _| FunctionDomain::Full,
-        vectorize_with_builder_3_arg::<StringType, NumberType<i64>, NumberType<u64>, StringType>(|s, pos, len, output, _| {
-            output.put_slice(substr(s, pos, len));
-            output.commit_row();
-        }),
-    );
-
-    registry.register_passthrough_nullable_2_arg::<StringType, NumberType<i64>, StringType, _, _>(
-        "substr_utf8",
-        |_, _, _| FunctionDomain::MayThrow,
-        vectorize_with_builder_2_arg::<StringType, NumberType<i64>, StringType>(
-            |s, pos, output, ctx| match std::str::from_utf8(s) {
-                Ok(s) => substr_utf8(output, s, pos, s.len() as u64),
-                Err(e) => {
-                    ctx.set_error(output.len(), e.to_string());
-                    output.commit_row();
-                }
-            },
-        ),
-    );
-
-    registry.register_passthrough_nullable_3_arg::<StringType, NumberType<i64>, NumberType<u64>, StringType, _, _>(
-        "substr_utf8",
-        |_, _, _, _| FunctionDomain::MayThrow,
-        vectorize_with_builder_3_arg::<StringType, NumberType<i64>, NumberType<u64>, StringType>(|s, pos, len, output, ctx| {
-            match std::str::from_utf8(s)  {
-                Ok(s) => substr_utf8(output, s, pos, len),
-                Err(e) =>  {
-                    ctx.set_error(output.len(), e.to_string());
-                    output.commit_row();
-                },
-            }
-        }),
-    );
+             |_, _, _, _| FunctionDomain::Full,
+             vectorize_with_builder_3_arg::<StringType, NumberType<i64>, NumberType<u64>, StringType>(|s, pos, len, output, _ctx| {
+                substr(output, s, pos, len);
+             }),
+         );
 
     registry
         .register_passthrough_nullable_2_arg::<StringType, StringType, ArrayType<StringType>, _, _>(
             "split",
             |_, _, _| FunctionDomain::Full,
             vectorize_with_builder_2_arg::<StringType, StringType, ArrayType<StringType>>(
-                |str, sep, output, ctx| match std::str::from_utf8(str) {
-                    Ok(s) => match std::str::from_utf8(sep) {
-                        Ok(sep) => {
-                            if s == sep {
-                                output.builder.put_slice(&[]);
-                                output.builder.commit_row();
-                            } else if sep.is_empty() {
-                                output.builder.put_slice(str);
-                                output.builder.commit_row();
-                            } else {
-                                let split = s.split(&sep);
-                                for i in split {
-                                    output.builder.put_slice(i.as_bytes());
-                                    output.builder.commit_row();
-                                }
-                            }
-                            output.commit_row()
-                        }
-                        Err(e) => {
-                            ctx.set_error(output.len(), e.to_string());
-                            output.commit_row();
+                |s, sep, output, _ctx| {
+                    if s == sep {
+                        output.builder.commit_row();
+                    } else if sep.is_empty() {
+                        output.builder.put_str(s);
+                        output.builder.commit_row();
+                    } else {
+                        for v in s.split(sep) {
+                            output.builder.put_str(v);
+                            output.builder.commit_row();
                         }
-                    },
-                    Err(e) => {
-                        ctx.set_error(output.len(), e.to_string());
-                        output.commit_row();
                     }
+                    output.commit_row();
                 },
             ),
         );
@@ -789,52 +695,35 @@ pub fn register(registry: &mut FunctionRegistry) {
             "split_part",
             |_, _, _, _| FunctionDomain::Full,
             vectorize_with_builder_3_arg::<StringType, StringType, NumberType<i64>, StringType>(
-                |str, sep, part, output, ctx| match std::str::from_utf8(str) {
-                    Ok(s) => match std::str::from_utf8(sep) {
-                        Ok(sep) => {
-                            if s == sep {
-                                output.commit_row()
-                            } else if sep.is_empty() {
-                                if part == 0 || part == 1 || part == -1 {
-                                    output.put_slice(str);
+                |s, sep, part, output, _| {
+                    if sep.is_empty() {
+                        if part == 0 || part == 1 || part == -1 {
+                            output.put_str(s);
+                        }
+                    } else if s != sep {
+                        if part < 0 {
+                            let idx = (-part-1) as usize;
+                            for (i, v) in s.rsplit(sep).enumerate() {
+                                if i == idx {
+                                    output.put_str(v);
+                                    break;
                                 }
-                                output.commit_row()
+                            }
+                        } else {
+                            let idx = if part == 0 {
+                                0usize
                             } else {
-                                if part < 0 {
-                                    let split = s.rsplit(&sep);
-                                    let idx = (-part-1) as usize;
-                                    for (count, i) in split.enumerate() {
-                                        if idx == count {
-                                            output.put_slice(i.as_bytes());
-                                            break
-                                        }
-                                    }
-                                } else {
-                                    let split = s.split(&sep);
-                                    let idx = if part == 0 {
-                                        0usize
-                                    } else {
-                                        (part - 1) as usize
-                                    };
-                                    for (count, i) in split.enumerate() {
-                                        if idx == count {
-                                            output.put_slice(i.as_bytes());
-                                            break
-                                        }
-                                    }
+                                (part - 1) as usize
+                            };
+                            for (i, v) in s.split(sep).enumerate() {
+                                if i == idx {
+                                    output.put_str(v);
+                                    break;
                                 }
-                                output.commit_row();
                             }
                         }
-                        Err(e) => {
-                            ctx.set_error(output.len(), e.to_string());
-                            output.commit_row();
-                        }
-                    },
-                    Err(e) => {
-                        ctx.set_error(output.len(), e.to_string());
-                        output.commit_row();
                     }
+                    output.commit_row();
                 },
             ),
         )
@@ -844,11 +733,11 @@ pub(crate) mod soundex {
     use databend_common_expression::types::string::StringColumnBuilder;
     use databend_common_expression::EvalContext;
 
-    pub fn soundex(val: &[u8], output: &mut StringColumnBuilder, _eval_context: &mut EvalContext) {
+    pub fn soundex(val: &str, output: &mut StringColumnBuilder, _eval_context: &mut EvalContext) {
         let mut last = None;
         let mut count = 0;
 
-        for ch in String::from_utf8_lossy(val).chars() {
+        for ch in val.chars() {
             let score = number_map(ch);
             if last.is_none() {
                 if !is_uni_alphabetic(ch) {
@@ -903,30 +792,7 @@ pub(crate) mod soundex {
 }
 
 #[inline]
-fn substr(str: &[u8], pos: i64, len: u64) -> &[u8] {
-    if pos > 0 && pos <= str.len() as i64 {
-        let l = str.len();
-        let s = (pos - 1) as usize;
-        let mut e = len as usize + s;
-        if e > l {
-            e = l;
-        }
-        return &str[s..e];
-    }
-    if pos < 0 && -(pos) <= str.len() as i64 {
-        let l = str.len();
-        let s = l - -pos as usize;
-        let mut e = len as usize + s;
-        if e > l {
-            e = l;
-        }
-        return &str[s..e];
-    }
-    &str[0..0]
-}
-
-#[inline]
-fn substr_utf8(builder: &mut StringColumnBuilder, str: &str, pos: i64, len: u64) {
+fn substr(builder: &mut StringColumnBuilder, str: &str, pos: i64, len: u64) {
     if pos == 0 || len == 0 {
         builder.commit_row();
         return;
@@ -948,7 +814,7 @@ fn substr_utf8(builder: &mut StringColumnBuilder, str: &str, pos: i64, len: u64)
 /// String to String scalar function with estimated output column capacity.
 pub fn vectorize_string_to_string(
     estimate_bytes: impl Fn(&StringColumn) -> usize + Copy,
-    func: impl Fn(&[u8], &mut StringColumnBuilder, &mut EvalContext) + Copy,
+    func: impl Fn(&str, &mut StringColumnBuilder, &mut EvalContext) + Copy,
 ) -> impl Fn(ValueRef<StringType>, &mut EvalContext) -> Value<StringType> + Copy {
     move |arg1, ctx| match arg1 {
         ValueRef::Scalar(val) => {
@@ -971,7 +837,7 @@ pub fn vectorize_string_to_string(
 /// (String, String) to String scalar function with estimated output column capacity.
 fn vectorize_string_to_string_2_arg(
     estimate_bytes: impl Fn(&StringColumn, &StringColumn) -> usize + Copy,
-    func: impl Fn(&[u8], &[u8], &mut EvalContext, &mut StringColumnBuilder) + Copy,
+    func: impl Fn(&str, &str, &mut EvalContext, &mut StringColumnBuilder) + Copy,
 ) -> impl Fn(ValueRef<StringType>, ValueRef<StringType>, &mut EvalContext) -> Value<StringType> + Copy
 {
     move |arg1, arg2, ctx| match (arg1, arg2) {
diff --git a/src/query/functions/src/scalars/string_multi_args.rs b/src/query/functions/src/scalars/string_multi_args.rs
index e78020417a92..f8135f3745c8 100644
--- a/src/query/functions/src/scalars/string_multi_args.rs
+++ b/src/query/functions/src/scalars/string_multi_args.rs
@@ -19,8 +19,6 @@ use databend_common_expression::passthrough_nullable;
 use databend_common_expression::types::nullable::NullableColumn;
 use databend_common_expression::types::number::Int64Type;
 use databend_common_expression::types::number::NumberScalar;
-use databend_common_expression::types::number::UInt8Type;
-use databend_common_expression::types::string::StringColumn;
 use databend_common_expression::types::string::StringColumnBuilder;
 use databend_common_expression::types::string::StringDomain;
 use databend_common_expression::types::NumberColumn;
@@ -123,9 +121,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                             for idx in 0..size {
                                 for (arg_index, arg) in args.iter().skip(1).enumerate() {
                                     if arg_index != 0 {
-                                        builder.put_slice(sep);
+                                        builder.put_str(sep);
                                     }
-                                    unsafe { builder.put_slice(arg.index_unchecked(idx)) }
+                                    builder.put_str(unsafe { arg.index_unchecked(idx) });
                                 }
                                 builder.commit_row();
                             }
@@ -134,11 +132,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                             for idx in 0..size {
                                 for (arg_index, arg) in args.iter().skip(1).enumerate() {
                                     if arg_index != 0 {
-                                        unsafe {
-                                            builder.put_slice(c.index_unchecked(idx));
-                                        }
+                                        builder.put_str(unsafe { c.index_unchecked(idx) });
                                     }
-                                    unsafe { builder.put_slice(arg.index_unchecked(idx)) }
+                                    builder.put_str(unsafe { arg.index_unchecked(idx) });
                                 }
                                 builder.commit_row();
                             }
@@ -197,9 +193,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                                     .enumerate()
                                 {
                                     if i != 0 {
-                                        builder.put_slice(v);
+                                        builder.put_str(v);
                                     }
-                                    builder.put_slice(s);
+                                    builder.put_str(s);
                                 }
                                 builder.commit_row();
                             }
@@ -212,16 +208,16 @@ pub fn register(registry: &mut FunctionRegistry) {
                                 unsafe {
                                     match new_args[0].index_unchecked(idx) {
                                         Some(sep) => {
-                                            for (i, str) in new_args
+                                            for (i, s) in new_args
                                                 .iter()
                                                 .skip(1)
                                                 .filter_map(|arg| arg.index_unchecked(idx))
                                                 .enumerate()
                                             {
                                                 if i != 0 {
-                                                    builder.put_slice(sep);
+                                                    builder.put_str(sep);
                                                 }
-                                                builder.put_slice(str);
+                                                builder.put_str(s);
                                             }
                                             builder.commit_row();
                                             validity.push(true);
@@ -247,53 +243,6 @@ pub fn register(registry: &mut FunctionRegistry) {
         }))
     });
 
-    registry.register_function_factory("char", |_, args_type| {
-        if args_type.is_empty() {
-            return None;
-        }
-        let has_null = args_type.iter().any(|t| t.is_nullable_or_null());
-        let f = Function {
-            signature: FunctionSignature {
-                name: "char".to_string(),
-                args_type: vec![DataType::Number(NumberDataType::UInt8); args_type.len()],
-                return_type: DataType::String,
-            },
-            eval: FunctionEval::Scalar {
-                calc_domain: Box::new(|_, _| FunctionDomain::Full),
-                eval: Box::new(char_fn),
-            },
-        };
-
-        if has_null {
-            Some(Arc::new(f.passthrough_nullable()))
-        } else {
-            Some(Arc::new(f))
-        }
-    });
-
-    // nullable char
-    registry.register_function_factory("char", |_, args_type| {
-        if args_type.is_empty() {
-            return None;
-        }
-        Some(Arc::new(Function {
-            signature: FunctionSignature {
-                name: "char".to_string(),
-                args_type: vec![
-                    DataType::Nullable(Box::new(DataType::Number(
-                        NumberDataType::UInt8
-                    )));
-                    args_type.len()
-                ],
-                return_type: DataType::Nullable(Box::new(DataType::String)),
-            },
-            eval: FunctionEval::Scalar {
-                calc_domain: Box::new(|_, _| FunctionDomain::MayThrow),
-                eval: Box::new(passthrough_nullable(char_fn)),
-            },
-        }))
-    });
-
     // Notes: https://dev.mysql.com/doc/refman/8.0/en/regexp.html#function_regexp-instr
     registry.register_function_factory("regexp_instr", |_, args_type| {
         let has_null = args_type.iter().any(|t| t.is_nullable_or_null());
@@ -484,7 +433,7 @@ fn concat_fn(args: &[ValueRef<AnyType>], _: &mut EvalContext) -> Value<AnyType>
     let mut builder = StringColumnBuilder::with_capacity(size, 0);
     for idx in 0..size {
         for arg in &args {
-            unsafe { builder.put_slice(arg.index_unchecked(idx)) }
+            builder.put_str(unsafe { arg.index_unchecked(idx) });
         }
         builder.commit_row();
     }
@@ -495,51 +444,6 @@ fn concat_fn(args: &[ValueRef<AnyType>], _: &mut EvalContext) -> Value<AnyType>
     }
 }
 
-fn char_fn(args: &[ValueRef<AnyType>], _: &mut EvalContext) -> Value<AnyType> {
-    let args = args
-        .iter()
-        .map(|arg| arg.try_downcast::<UInt8Type>().unwrap())
-        .collect::<Vec<_>>();
-
-    let len = args.iter().find_map(|arg| match arg {
-        ValueRef::Column(col) => Some(col.len()),
-        _ => None,
-    });
-    let input_rows = len.unwrap_or(1);
-
-    let mut values: Vec<u8> = vec![0; input_rows * args.len()];
-    let values_ptr = values.as_mut_ptr();
-
-    for (i, arg) in args.iter().enumerate() {
-        match arg {
-            ValueRef::Scalar(v) => {
-                for j in 0..input_rows {
-                    unsafe {
-                        *values_ptr.add(args.len() * j + i) = *v;
-                    }
-                }
-            }
-            ValueRef::Column(c) => {
-                for (j, ch) in UInt8Type::iter_column(c).enumerate() {
-                    unsafe {
-                        *values_ptr.add(args.len() * j + i) = ch;
-                    }
-                }
-            }
-        }
-    }
-    let offsets = (0..(input_rows + 1) as u64 * args.len() as u64)
-        .step_by(args.len())
-        .collect::<Vec<_>>();
-    let result = StringColumn::new(values.into(), offsets.into());
-
-    let col = Column::String(result);
-    match len {
-        Some(_) => Value::Column(col),
-        _ => Value::Scalar(AnyType::index_column(&col, 0).unwrap().to_owned()),
-    }
-}
-
 fn regexp_instr_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value<AnyType> {
     let len = args.iter().find_map(|arg| match arg {
         ValueRef::Column(col) => Some(col.len()),
@@ -589,7 +493,7 @@ fn regexp_instr_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value<A
     };
 
     for idx in 0..size {
-        let source = unsafe { source_arg.index_unchecked(idx) };
+        let mut source = unsafe { source_arg.index_unchecked(idx) };
         let pat = unsafe { pat_arg.index_unchecked(idx) };
         let pos = pos_arg
             .as_ref()
@@ -636,6 +540,9 @@ fn regexp_instr_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value<A
         let occur = occur.unwrap_or(1);
         let ro = ro.unwrap_or(0);
 
+        if let Some((idx, _)) = source.char_indices().nth((pos - 1) as usize) {
+            source = &source[idx..];
+        }
         let instr = regexp::regexp_instr(source, re, pos, occur, ro);
         builder.push(instr);
     }
@@ -784,7 +691,7 @@ fn regexp_replace_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value
         }
 
         if source.is_empty() || pat.is_empty() {
-            builder.data.extend_from_slice(source);
+            builder.put_str(source);
             builder.commit_row();
             continue;
         }
@@ -809,7 +716,7 @@ fn regexp_replace_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value
         let pos = pos.unwrap_or(1);
         let occur = occur.unwrap_or(0);
 
-        regexp::regexp_replace(source, re, repl, pos, occur, &mut builder.data);
+        regexp::regexp_replace(source, re, repl, pos, occur, &mut builder);
         builder.commit_row();
     }
     match len {
@@ -911,7 +818,7 @@ fn regexp_substr_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value<
         let substr = regexp::regexp_substr(source, re, pos, occur);
         match substr {
             Some(substr) => {
-                builder.put_slice(substr);
+                builder.put_str(substr);
                 validity.push(true);
             }
             None => {
@@ -942,33 +849,31 @@ fn regexp_substr_fn(args: &[ValueRef<AnyType>], ctx: &mut EvalContext) -> Value<
 }
 
 pub mod regexp {
-    use bstr::ByteSlice;
-    use regex::bytes::Match;
-    use regex::bytes::Regex;
-    use regex::bytes::RegexBuilder;
+    use databend_common_expression::types::string::StringColumnBuilder;
+    use regex::Regex;
+    use regex::RegexBuilder;
 
     #[inline]
     pub fn build_regexp_from_pattern(
         fn_name: &str,
-        pat: &[u8],
-        mt: Option<&[u8]>,
+        pat: &str,
+        mt: Option<&str>,
     ) -> Result<Regex, String> {
         let pattern = match pat.is_empty() {
             true => "^$",
-            false => simdutf8::basic::from_utf8(pat).map_err(|e| {
-                format!("Unable to convert the {} pattern to string: {}", fn_name, e)
-            })?,
+            false => pat,
         };
+
         // the default match type value is 'i', if it is empty
         let mt = match mt {
             Some(mt) => {
                 if mt.is_empty() {
-                    "i".as_bytes()
+                    "i"
                 } else {
                     mt
                 }
             }
-            None => "i".as_bytes(),
+            None => "i",
         };
 
         let mut builder = RegexBuilder::new(pattern);
@@ -1040,115 +945,76 @@ pub mod regexp {
     }
 
     #[inline]
-    pub fn regexp_instr(s: &[u8], re: &Regex, pos: i64, occur: i64, ro: i64) -> u64 {
-        let pos = (pos - 1) as usize; // set the index start from 0
-
-        // the 'pos' position is the character index,
-        // so we should iterate the character to find the byte index.
-        let mut pos = match s.char_indices().nth(pos) {
-            Some((start, _, _)) => start,
-            None => return 0,
-        };
+    pub fn regexp_instr(
+        expr: &str,
+        regex: &Regex,
+        pos: i64,
+        occurrence: i64,
+        return_option: i64,
+    ) -> u64 {
+        if let Some(m) = regex.find_iter(expr).nth((occurrence - 1) as usize) {
+            let find_pos = if return_option == 0 {
+                m.start()
+            } else {
+                m.end()
+            };
 
-        let m = regexp_match_result(s, re, &mut pos, &occur);
-        if m.is_none() {
-            return 0;
+            let count = expr[..find_pos].chars().count() as i64;
+            return (count + pos) as _;
         }
-
-        // the matched result is the byte index, but the 'regexp_instr' function returns the character index,
-        // so we should iterate the character to find the character index.
-        let mut instr = 0_usize;
-        for (p, (start, end, _)) in s.char_indices().enumerate() {
-            if ro == 0 {
-                if start == m.unwrap().start() {
-                    instr = p + 1;
-                    break;
-                }
-            } else if end == m.unwrap().end() {
-                instr = p + 2;
-                break;
-            }
-        }
-
-        instr as u64
+        0
     }
 
     #[inline]
     pub fn regexp_replace(
-        s: &[u8],
+        s: &str,
         re: &Regex,
-        repl: &[u8],
+        repl: &str,
         pos: i64,
         occur: i64,
-        buf: &mut Vec<u8>,
+        builder: &mut StringColumnBuilder,
     ) {
         let pos = (pos - 1) as usize; // set the index start from 0
-
         // the 'pos' position is the character index,
         // so we should iterate the character to find the byte index.
-        let mut pos = match s.char_indices().nth(pos) {
-            Some((start, _, _)) => start,
+        let char_pos = match s.char_indices().nth(pos) {
+            Some((start, _)) => start,
             None => {
-                buf.extend_from_slice(s);
+                builder.put_str(s);
                 return;
             }
         };
 
-        let m = regexp_match_result(s, re, &mut pos, &occur);
-        if m.is_none() {
-            buf.extend_from_slice(s);
-            return;
-        }
-
-        buf.extend_from_slice(&s[..m.unwrap().start()]);
+        let (before_trimmed, trimmed) = (&s[..char_pos], &s[char_pos..]);
+        builder.put_str(before_trimmed);
 
+        // means we should replace all matched strings
         if occur == 0 {
-            let s = &s[m.unwrap().start()..];
-            buf.extend_from_slice(&re.replace_all(s, repl));
+            builder.put_str(&re.replace_all(trimmed, repl));
+        } else if let Some(capature) = re.captures_iter(trimmed).nth((occur - 1) as _) {
+            // unwrap on 0 is OK because captures only reports matches.
+            let m = capature.get(0).unwrap();
+            builder.put_str(&trimmed[0..m.start()]);
+            builder.put_str(repl);
+            builder.put_str(&trimmed[m.end()..]);
         } else {
-            buf.extend_from_slice(repl);
-            buf.extend_from_slice(&s[m.unwrap().end()..])
+            builder.put_str(trimmed);
         }
     }
 
     #[inline]
-    pub fn regexp_substr<'a>(s: &'a [u8], re: &Regex, pos: i64, occur: i64) -> Option<&'a [u8]> {
+    pub fn regexp_substr<'a>(s: &'a str, re: &Regex, pos: i64, occur: i64) -> Option<&'a str> {
         let occur = if occur < 1 { 1 } else { occur };
         let pos = if pos < 1 { 0 } else { (pos - 1) as usize };
 
         // the 'pos' position is the character index,
         // so we should iterate the character to find the byte index.
-        let mut pos = match s.char_indices().nth(pos) {
-            Some((start, _, _)) => start,
+        let char_pos = match s.char_indices().nth(pos) {
+            Some((start, _)) => start,
             None => return None,
         };
 
-        let m = regexp_match_result(s, re, &mut pos, &occur);
-
-        m.map(|m| m.as_bytes())
-    }
-
-    #[inline]
-    fn regexp_match_result<'a>(
-        s: &'a [u8],
-        re: &Regex,
-        pos: &mut usize,
-        occur: &i64,
-    ) -> Option<Match<'a>> {
-        let mut i = 1_i64;
-        let m = loop {
-            let m = re.find_at(s, *pos);
-            if i >= *occur || m.is_none() {
-                break m;
-            }
-
-            i += 1;
-            if let Some(m) = m {
-                // set the start position of 'find_at' function to the position following the matched substring
-                *pos = m.end();
-            }
-        };
-
-        m
+        let m = re.find_iter(&s[char_pos..]).nth((occur - 1) as _);
+        m.map(|m| m.as_str())
     }
 }
diff --git a/src/query/functions/src/scalars/variant.rs b/src/query/functions/src/scalars/variant.rs
index bf2bd559026a..850616b3782f 100644
--- a/src/query/functions/src/scalars/variant.rs
+++ b/src/query/functions/src/scalars/variant.rs
@@ -135,7 +135,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                     return;
                 }
             }
-            match parse_value(s) {
+            match parse_value(s.as_bytes()) {
                 Ok(value) => {
                     value.write_to_vec(&mut output.data);
                 }
@@ -180,7 +180,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                     return;
                 }
             }
-            match parse_value(s) {
+            match parse_value(s.as_bytes()) {
                 Ok(value) => {
                     output.validity.push(true);
                     value.write_to_vec(&mut output.builder.data);
@@ -205,7 +205,7 @@ pub fn register(registry: &mut FunctionRegistry) {
             let val = to_string(s);
             match parse_value(val.as_bytes()) {
                 Ok(_) => output.push_null(),
-                Err(e) => output.push(e.to_string().as_bytes()),
+                Err(e) => output.push(&e.to_string()),
             }
         }),
     );
@@ -220,9 +220,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                     return;
                 }
             }
-            match parse_value(s) {
+            match parse_value(s.as_bytes()) {
                 Ok(_) => output.push_null(),
-                Err(e) => output.push(e.to_string().as_bytes()),
+                Err(e) => output.push(&e.to_string()),
             }
         }),
     );
@@ -291,7 +291,7 @@ pub fn register(registry: &mut FunctionRegistry) {
 
     registry.register_combine_nullable_2_arg::<VariantType, StringType, VariantType, _, _>(
         "get",
-        |_, _, _| FunctionDomain::MayThrow,
+        |_, _, _| FunctionDomain::Full,
         vectorize_with_builder_2_arg::<VariantType, StringType, NullableType<VariantType>>(
             |val, name, output, ctx| {
                 if let Some(validity) = &ctx.validity {
@@ -300,24 +300,11 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match std::str::from_utf8(name) {
-                    Ok(name) => match get_by_name(val, name, false) {
-                        Some(v) => {
-                            output.push(&v);
-                        }
-                        None => {
-                            output.push_null();
-                        }
-                    },
-                    Err(err) => {
-                        ctx.set_error(
-                            output.len(),
-                            format!(
-                                "Unable convert name '{}' to string: {}",
-                                &String::from_utf8_lossy(name),
-                                err
-                            ),
-                        );
+                match get_by_name(val, name, false) {
+                    Some(v) => {
+                        output.push(&v);
+                    }
+                    None => {
                         output.push_null();
                     }
                 }
@@ -354,7 +341,7 @@ pub fn register(registry: &mut FunctionRegistry) {
 
     registry.register_combine_nullable_2_arg::<VariantType, StringType, VariantType, _, _>(
         "get_ignore_case",
-        |_, _, _| FunctionDomain::MayThrow,
+        |_, _, _| FunctionDomain::Full,
         vectorize_with_builder_2_arg::<VariantType, StringType, NullableType<VariantType>>(
             |val, name, output, ctx| {
                 if let Some(validity) = &ctx.validity {
@@ -363,22 +350,9 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match std::str::from_utf8(name) {
-                    Ok(name) => match get_by_name(val, name, true) {
-                        Some(v) => output.push(&v),
-                        None => output.push_null(),
-                    },
-                    Err(err) => {
-                        ctx.set_error(
-                            output.len(),
-                            format!(
-                                "Unable convert name '{}' to string: {}",
-                                &String::from_utf8_lossy(name),
-                                err
-                            ),
-                        );
-                        output.push_null();
-                    }
+                match get_by_name(val, name, true) {
+                    Some(v) => output.push(&v),
+                    None => output.push_null(),
                 }
             },
         ),
@@ -386,7 +360,7 @@ pub fn register(registry: &mut FunctionRegistry) {
 
     registry.register_combine_nullable_2_arg::<VariantType, StringType, StringType, _, _>(
         "get_string",
-        |_, _, _| FunctionDomain::MayThrow,
+        |_, _, _| FunctionDomain::Full,
         vectorize_with_builder_2_arg::<VariantType, StringType, NullableType<StringType>>(
             |val, name, output, ctx| {
                 if let Some(validity) = &ctx.validity {
@@ -395,27 +369,12 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match std::str::from_utf8(name) {
-                    Ok(name) => match get_by_name(val, name, false) {
-                        Some(v) => {
-                            let json_str = cast_to_string(&v);
-                            output.builder.put_str(&json_str);
-                            output.builder.commit_row();
-                            output.validity.push(true);
-                        }
-                        None => output.push_null(),
-                    },
-                    Err(err) => {
-                        ctx.set_error(
-                            output.len(),
-                            format!(
-                                "Unable convert name '{}' to string: {}",
-                                &String::from_utf8_lossy(name),
-                                err
-                            ),
-                        );
-                        output.push_null();
+                match get_by_name(val, name, false) {
+                    Some(v) => {
+                        let json_str = cast_to_string(&v);
+                        output.push(&json_str);
                     }
+                    None => output.push_null(),
                 }
             },
         ),
@@ -438,9 +397,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                     match get_by_index(val, idx as usize) {
                         Some(v) => {
                             let json_str = cast_to_string(&v);
-                            output.builder.put_str(&json_str);
-                            output.builder.commit_row();
-                            output.validity.push(true);
+                            output.push(&json_str);
                         }
                         None => {
                             output.push_null();
@@ -462,7 +419,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match parse_json_path(path) {
+                match parse_json_path(path.as_bytes()) {
                     Ok(json_path) => {
                         get_by_path_array(
                             val,
@@ -477,10 +434,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         }
                     }
                     Err(_) => {
-                        ctx.set_error(
-                            output.len(),
-                            format!("Invalid JSON Path '{}'", &String::from_utf8_lossy(path),),
-                        );
+                        ctx.set_error(output.len(), format!("Invalid JSON Path '{path}'"));
                         output.push_null();
                     }
                 }
@@ -499,7 +453,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match parse_json_path(path) {
+                match parse_json_path(path.as_bytes()) {
                     Ok(json_path) => {
                         get_by_path_first(
                             val,
@@ -514,10 +468,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         }
                     }
                     Err(_) => {
-                        ctx.set_error(
-                            output.len(),
-                            format!("Invalid JSON Path '{}'", &String::from_utf8_lossy(path),),
-                        );
+                        ctx.set_error(output.len(), format!("Invalid JSON Path '{path}'"));
                         output.push_null();
                     }
                 }
@@ -584,7 +535,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match parse_json_path(path) {
+                match parse_json_path(path.as_bytes()) {
                     Ok(json_path) => {
                         get_by_path(
                             val,
@@ -599,10 +550,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         }
                     }
                     Err(_) => {
-                        ctx.set_error(
-                            output.len(),
-                            format!("Invalid JSON Path '{}'", &String::from_utf8_lossy(path),),
-                        );
+                        ctx.set_error(output.len(), format!("Invalid JSON Path '{path}'"));
                         output.push_null();
                     }
                 }
@@ -621,11 +569,11 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                match parse_value(s) {
+                match parse_value(s.as_bytes()) {
                     Ok(val) => {
                         let mut buf = Vec::new();
                         val.write_to_vec(&mut buf);
-                        match parse_json_path(path) {
+                        match parse_json_path(path.as_bytes()) {
                             Ok(json_path) => {
                                 let mut out_buf = Vec::new();
                                 let mut out_offsets = Vec::new();
@@ -634,19 +582,11 @@ pub fn register(registry: &mut FunctionRegistry) {
                                     output.push_null();
                                 } else {
                                     let json_str = cast_to_string(&out_buf);
-                                    output.builder.put_str(&json_str);
-                                    output.builder.commit_row();
-                                    output.validity.push(true);
+                                    output.push(&json_str);
                                 }
                             }
                             Err(_) => {
-                                ctx.set_error(
-                                    output.len(),
-                                    format!(
-                                        "Invalid JSON Path '{}'",
-                                        &String::from_utf8_lossy(path),
-                                    ),
-                                );
+                                ctx.set_error(output.len(), format!("Invalid JSON Path '{path}'"));
                                 output.push_null();
                             }
                         }
@@ -722,7 +662,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                 }
             }
             match as_str(v) {
-                Some(val) => output.push(val.as_bytes()),
+                Some(val) => output.push(&val),
                 None => output.push_null(),
             }
         }),
@@ -873,9 +813,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                     }
                 }
                 let json_str = cast_to_string(val);
-                output.builder.put_str(&json_str);
-                output.builder.commit_row();
-                output.validity.push(true);
+                output.push(&json_str);
             },
         ),
     );
@@ -1069,7 +1007,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                 }
             }
             let s = to_pretty_string(val);
-            output.put_slice(s.as_bytes());
+            output.put_str(&s);
             output.commit_row();
         }),
     );
@@ -1247,7 +1185,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                let result = exists_any_keys(val, keys.iter());
+                let result = exists_any_keys(val, keys.iter().map(|k| k.as_bytes()));
                 output.push(result);
             },
         ),
@@ -1264,7 +1202,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                let result = exists_all_keys(val, keys.iter());
+                let result = exists_all_keys(val, keys.iter().map(|k| k.as_bytes()));
                 output.push(result);
             },
         ),
@@ -1281,7 +1219,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                         return;
                     }
                 }
-                let result = exists_all_keys(val, once(key));
+                let result = exists_all_keys(val, once(key.as_bytes()));
                 output.push(result);
             },
         ),
@@ -1351,7 +1289,7 @@ fn json_object_impl_fn(
                     continue;
                 }
                 let key = match k {
-                    ScalarRef::String(v) => unsafe { String::from_utf8_unchecked(v.to_vec()) },
+                    ScalarRef::String(v) => v,
                     _ => {
                         has_err = true;
                         ctx.set_error(builder.len(), "Key must be a string value");
@@ -1363,7 +1301,7 @@ fn json_object_impl_fn(
                     ctx.set_error(builder.len(), "Keys have to be unique");
                     break;
                 }
-                set.insert(key.clone());
+                set.insert(key);
                 let mut val = vec![];
                 cast_scalar_to_variant(v, ctx.func_ctx.tz, &mut val);
                 kvs.push((key, val));
@@ -1413,7 +1351,7 @@ fn get_by_keypath_fn(
     string_res: bool,
 ) -> Value<AnyType> {
     let scalar_keypath = match &args[1] {
-        ValueRef::Scalar(ScalarRef::String(v)) => Some(parse_key_paths(v)),
+        ValueRef::Scalar(ScalarRef::String(v)) => Some(parse_key_paths(v.as_bytes())),
         _ => None,
     };
     let len_opt = args.iter().find_map(|arg| match arg {
@@ -1436,7 +1374,7 @@ fn get_by_keypath_fn(
             ValueRef::Column(col) => {
                 let scalar = unsafe { col.index_unchecked(idx) };
                 let path = match scalar {
-                    ScalarRef::String(buf) => Some(parse_key_paths(buf)),
+                    ScalarRef::String(buf) => Some(parse_key_paths(buf.as_bytes())),
                     _ => None,
                 };
                 Cow::Owned(path)
@@ -1507,8 +1445,7 @@ where
 {
     let scalar_jsonpath = match &args[1] {
         ValueRef::Scalar(ScalarRef::String(v)) => {
-            let res = parse_json_path(v)
-                .map_err(|_| format!("Invalid JSON Path '{}'", &String::from_utf8_lossy(v)));
+            let res = parse_json_path(v.as_bytes()).map_err(|_| format!("Invalid JSON Path '{v}'"));
             Some(res)
         }
         _ => None,
@@ -1530,9 +1467,8 @@ where
                 let scalar = unsafe { col.index_unchecked(idx) };
                 match scalar {
                     ScalarRef::String(buf) => {
-                        let res = parse_json_path(buf).map_err(|_| {
-                            format!("Invalid JSON Path '{}'", &String::from_utf8_lossy(buf))
-                        });
+                        let res = parse_json_path(buf.as_bytes())
+                            .map_err(|_| format!("Invalid JSON Path '{buf}'"));
                         Some(res)
                     }
                     _ => None,
diff --git a/src/query/functions/src/scalars/vector.rs b/src/query/functions/src/scalars/vector.rs
index aa3d65a19fed..b551f2539bf3 100644
--- a/src/query/functions/src/scalars/vector.rs
+++ b/src/query/functions/src/scalars/vector.rs
@@ -141,17 +141,6 @@ pub fn register(registry: &mut FunctionRegistry) {
                 }
             }
 
-            let data = match std::str::from_utf8(data) {
-                Ok(data) => data,
-                Err(_) => {
-                    ctx.set_error(
-                        output.len(),
-                        format!("Invalid data: {:?}", String::from_utf8_lossy(data)),
-                    );
-                    output.push(vec![F32::from(0.0)].into());
-                    return;
-                }
-            };
             if ctx.func_ctx.openai_api_key.is_empty() {
                 ctx.set_error(output.len(), "openai_api_key is empty".to_string());
                 output.push(vec![F32::from(0.0)].into());
@@ -205,18 +194,6 @@ pub fn register(registry: &mut FunctionRegistry) {
                 }
             }
 
-            let data = match std::str::from_utf8(data) {
-                Ok(data) => data,
-                Err(_) => {
-                    ctx.set_error(
-                        output.len(),
-                        format!("Invalid data: {:?}", String::from_utf8_lossy(data)),
-                    );
-                    output.put_str("");
-                    output.commit_row();
-                    return;
-                }
-            };
             if ctx.func_ctx.openai_api_key.is_empty() {
                 ctx.set_error(output.len(), "openai_api_key is empty".to_string());
                 output.put_str("");
diff --git a/src/query/functions/src/srfs/variant.rs b/src/query/functions/src/srfs/variant.rs
index 6b6c5c3b1ccf..38a9e2cb261e 100644
--- a/src/query/functions/src/srfs/variant.rs
+++ b/src/query/functions/src/srfs/variant.rs
@@ -78,41 +78,38 @@ pub fn register(registry: &mut FunctionRegistry) {
                     let path_arg = args[1].clone().to_owned();
                     let mut results = Vec::with_capacity(ctx.num_rows);
                     match path_arg {
-                        Value::Scalar(Scalar::String(path)) => match parse_json_path(&path) {
-                            Ok(json_path) => {
-                                let selector = Selector::new(json_path, SelectorMode::All);
-                                for (row, max_nums_per_row) in
-                                    max_nums_per_row.iter_mut().enumerate().take(ctx.num_rows)
-                                {
-                                    let val = unsafe { val_arg.index_unchecked(row) };
-                                    let mut builder = BinaryColumnBuilder::with_capacity(0, 0);
-                                    if let ScalarRef::Variant(val) = val {
-                                        selector.select(
-                                            val,
-                                            &mut builder.data,
-                                            &mut builder.offsets,
-                                        );
+                        Value::Scalar(Scalar::String(path)) => {
+                            match parse_json_path(path.as_bytes()) {
+                                Ok(json_path) => {
+                                    let selector = Selector::new(json_path, SelectorMode::All);
+                                    for (row, max_nums_per_row) in
+                                        max_nums_per_row.iter_mut().enumerate().take(ctx.num_rows)
+                                    {
+                                        let val = unsafe { val_arg.index_unchecked(row) };
+                                        let mut builder = BinaryColumnBuilder::with_capacity(0, 0);
+                                        if let ScalarRef::Variant(val) = val {
+                                            selector.select(
+                                                val,
+                                                &mut builder.data,
+                                                &mut builder.offsets,
+                                            );
+                                        }
+                                        let array =
+                                            Column::Variant(builder.build()).wrap_nullable(None);
+                                        let array_len = array.len();
+                                        *max_nums_per_row =
+                                            std::cmp::max(*max_nums_per_row, array_len);
+                                        results.push((
+                                            Value::Column(Column::Tuple(vec![array])),
+                                            array_len,
+                                        ));
                                     }
-                                    let array =
-                                        Column::Variant(builder.build()).wrap_nullable(None);
-                                    let array_len = array.len();
-                                    *max_nums_per_row = std::cmp::max(*max_nums_per_row, array_len);
-                                    results.push((
-                                        Value::Column(Column::Tuple(vec![array])),
-                                        array_len,
-                                    ));
+                                }
+                                Err(_) => {
+                                    ctx.set_error(0, format!("Invalid JSON Path '{}'", &path,));
                                 }
                             }
-                            Err(_) => {
-                                ctx.set_error(
-                                    0,
-                                    format!(
-                                        "Invalid JSON Path '{}'",
-                                        &String::from_utf8_lossy(&path),
-                                    ),
-                                );
-                            }
-                        },
+                        }
                         _ => {
                             for (row, max_nums_per_row) in
                                 max_nums_per_row.iter_mut().enumerate().take(ctx.num_rows)
@@ -121,7 +118,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                                 let path = unsafe { path_arg.index_unchecked(row) };
                                 let mut builder = BinaryColumnBuilder::with_capacity(0, 0);
                                 if let ScalarRef::String(path) = path {
-                                    match parse_json_path(path) {
+                                    match parse_json_path(path.as_bytes()) {
                                         Ok(json_path) => {
                                             if let ScalarRef::Variant(val) = val {
                                                 let selector =
@@ -136,10 +133,7 @@ pub fn register(registry: &mut FunctionRegistry) {
                                         Err(_) => {
                                             ctx.set_error(
                                                 row,
-                                                format!(
-                                                    "Invalid JSON Path '{}'",
-                                                    &String::from_utf8_lossy(path),
-                                                ),
+                                                format!("Invalid JSON Path '{}'", &path,),
                                             );
                                             break;
                                         }
@@ -297,23 +291,18 @@ pub fn register(registry: &mut FunctionRegistry) {
 
                     if args.len() >= 2 {
                         match &args[1] {
-                            ValueRef::Scalar(ScalarRef::String(v)) => match parse_json_path(v) {
-                                Ok(jsonpath) => {
-                                    let path = unsafe { std::str::from_utf8_unchecked(v) };
-                                    let selector = Selector::new(jsonpath, SelectorMode::First);
-                                    json_path = Some((path, selector));
-                                }
-                                Err(_) => {
-                                    ctx.set_error(
-                                        0,
-                                        format!(
-                                            "Invalid JSON Path {:?}",
-                                            String::from_utf8_lossy(v)
-                                        ),
-                                    );
-                                    return results;
+                            ValueRef::Scalar(ScalarRef::String(v)) => {
+                                match parse_json_path(v.as_bytes()) {
+                                    Ok(jsonpath) => {
+                                        let selector = Selector::new(jsonpath, SelectorMode::First);
+                                        json_path = Some((v, selector));
+                                    }
+                                    Err(_) => {
+                                        ctx.set_error(0, format!("Invalid JSON Path {v:?}",));
+                                        return results;
+                                    }
                                 }
-                            },
+                            }
                             ValueRef::Column(_) => {
                                 ctx.set_error(
                                     0,
@@ -360,30 +349,18 @@ pub fn register(registry: &mut FunctionRegistry) {
                     if args.len() >= 5 {
                         match args[4] {
                             ValueRef::Scalar(ScalarRef::String(v)) => {
-                                match String::from_utf8(v.to_vec()) {
-                                    Ok(val) => match val.to_lowercase().as_str() {
-                                        "object" => {
-                                            mode = FlattenMode::Object;
-                                        }
-                                        "array" => {
-                                            mode = FlattenMode::Array;
-                                        }
-                                        "both" => {
-                                            mode = FlattenMode::Both;
-                                        }
-                                        _ => {
-                                            ctx.set_error(0, format!("Invalid mode {:?}", val));
-                                            return results;
-                                        }
-                                    },
-                                    Err(_) => {
-                                        ctx.set_error(
-                                            0,
-                                            format!(
-                                                "Invalid mode string {:?}",
-                                                String::from_utf8_lossy(v)
-                                            ),
-                                        );
+                                match v.to_lowercase().as_str() {
+                                    "object" => {
+                                        mode = FlattenMode::Object;
+                                    }
+                                    "array" => {
+                                        mode = FlattenMode::Array;
+                                    }
+                                    "both" => {
+                                        mode = FlattenMode::Both;
+                                    }
+                                    _ => {
+                                        ctx.set_error(0, format!("Invalid mode {v:?}"));
                                         return results;
                                     }
                                 }
@@ -491,7 +468,7 @@ fn unnest_variant_obj(
             max_nums_per_row[row] = std::cmp::max(max_nums_per_row[row], len);
 
             for (key, val) in vals {
-                key_builder.put_slice(&key);
+                key_builder.put_str(&String::from_utf8_lossy(&key));
                 key_builder.commit_row();
                 val_builder.put_slice(&val);
                 val_builder.commit_row();
@@ -615,7 +592,7 @@ impl FlattenGenerator {
                     key_builder.push_null();
                 }
                 if let Some(path_builder) = path_builder {
-                    path_builder.put_slice(inner_path.as_bytes());
+                    path_builder.put_str(&inner_path);
                     path_builder.commit_row();
                 }
                 if let Some(index_builder) = index_builder {
@@ -672,10 +649,10 @@ impl FlattenGenerator {
                     };
 
                     if let Some(key_builder) = key_builder {
-                        key_builder.push(name.as_bytes());
+                        key_builder.push(name.as_ref());
                     }
                     if let Some(path_builder) = path_builder {
-                        path_builder.put_slice(inner_path.as_bytes());
+                        path_builder.put_str(&inner_path);
                         path_builder.commit_row();
                     }
                     if let Some(index_builder) = index_builder {
@@ -768,13 +745,13 @@ impl FlattenGenerator {
         let key_column = if let Some(key_builder) = key_builder {
             NullableType::<StringType>::upcast_column(key_builder.build())
         } else {
-            StringType::upcast_column(StringColumnBuilder::repeat(&[], rows).build())
+            StringType::upcast_column(StringColumnBuilder::repeat("", rows).build())
                 .wrap_nullable(None)
         };
         let path_column = if let Some(path_builder) = path_builder {
             StringType::upcast_column(path_builder.build()).wrap_nullable(None)
         } else {
-            StringType::upcast_column(StringColumnBuilder::repeat(&[], rows).build())
+            StringType::upcast_column(StringColumnBuilder::repeat("", rows).build())
                 .wrap_nullable(None)
         };
         let index_column = if let Some(index_builder) = index_builder {
diff --git a/src/query/functions/tests/it/aggregates/agg_hashtable.rs b/src/query/functions/tests/it/aggregates/agg_hashtable.rs
index 3c1d98807e9d..4e92c0ef2f8b 100644
--- a/src/query/functions/tests/it/aggregates/agg_hashtable.rs
+++ b/src/query/functions/tests/it/aggregates/agg_hashtable.rs
@@ -57,11 +57,7 @@ fn test_agg_hashtable() {
     let m: usize = 4;
     for n in [100, 1000, 10_000, 100_000] {
         let columns = vec![
-            StringType::from_data(
-                (0..n)
-                    .map(|x| format!("{}", x % m).as_bytes().to_vec())
-                    .collect_vec(),
-            ),
+            StringType::from_data((0..n).map(|x| format!("{}", x % m)).collect_vec()),
             Int64Type::from_data((0..n).map(|x| (x % m) as i64).collect_vec()),
             Int32Type::from_data((0..n).map(|x| (x % m) as i32).collect_vec()),
             Int16Type::from_data((0..n).map(|x| (x % m) as i16).collect_vec()),
diff --git a/src/query/functions/tests/it/scalars/array.rs b/src/query/functions/tests/it/scalars/array.rs
index e83608c4511b..10fbc8f5daf9 100644
--- a/src/query/functions/tests/it/scalars/array.rs
+++ b/src/query/functions/tests/it/scalars/array.rs
@@ -155,6 +155,10 @@ fn test_contains(file: &mut impl Write) {
 
     let columns = [
         ("int8_col", Int8Type::from_data(vec![1i8, 2, 7, 8])),
+        (
+            "string_col",
+            StringType::from_data(vec![r#"1"#, r#"2"#, r#"5"#, r#"1234"#]),
+        ),
         (
             "nullable_col",
             Int64Type::from_data_with_validity(vec![9i64, 10, 11, 12], vec![
@@ -164,6 +168,20 @@ fn test_contains(file: &mut impl Write) {
     ];
 
     run_ast(file, "int8_col not in (1, 2, 3, 4, 5, null)", &columns);
+    run_ast(
+        file,
+        "contains(['5000', '6000', '7000'], string_col)",
+        &columns,
+    );
+
+    run_ast(file, "contains(['1', '5'], string_col)", &columns);
+
+    run_ast(
+        file,
+        "contains(['15000', '6000', '7000'], string_col)",
+        &columns,
+    );
+
     run_ast(file, "contains([1,2,null], nullable_col)", &columns);
     run_ast(
         file,
diff --git a/src/query/functions/tests/it/scalars/comparison.rs b/src/query/functions/tests/it/scalars/comparison.rs
index c93cc6046858..bb0f85c3ff17 100644
--- a/src/query/functions/tests/it/scalars/comparison.rs
+++ b/src/query/functions/tests/it/scalars/comparison.rs
@@ -295,6 +295,8 @@ fn test_gt(file: &mut impl Write) {
     ];
     run_ast(file, "parse_json(lhs) > parse_json(rhs)", &table);
     run_ast(file, "lhs > rhs", &table);
+    let table = [("col", StringType::from_data(vec![r#"bcd"#, r#"efg"#]))];
+    run_ast(file, "col > 'efg'", &table);
 }
 
 fn test_gte(file: &mut impl Write) {
diff --git a/src/query/functions/tests/it/scalars/parser.rs b/src/query/functions/tests/it/scalars/parser.rs
index 2f0c898b5800..db8934468c20 100644
--- a/src/query/functions/tests/it/scalars/parser.rs
+++ b/src/query/functions/tests/it/scalars/parser.rs
@@ -354,7 +354,7 @@ pub fn transform_expr(ast: AExpr, columns: &[(&str, DataType)]) -> RawExpr {
                     transform_expr(*expr, columns),
                     RawExpr::Constant {
                         span,
-                        scalar: Scalar::String(key.name.into_bytes()),
+                        scalar: Scalar::String(key.name),
                     },
                 ]),
                 MapAccessor::DotNumber { key } => {
@@ -567,7 +567,7 @@ pub fn transform_literal(lit: ASTLiteral) -> Scalar {
             precision,
             scale,
         })),
-        ASTLiteral::String(s) => Scalar::String(s.as_bytes().to_vec()),
+        ASTLiteral::String(s) => Scalar::String(s),
         ASTLiteral::Boolean(b) => Scalar::Boolean(b),
         ASTLiteral::Null => Scalar::Null,
         ASTLiteral::Float64(f) => Scalar::Number(NumberScalar::Float64(OrderedFloat(f))),
diff --git a/src/query/functions/tests/it/scalars/string.rs b/src/query/functions/tests/it/scalars/string.rs
index 089e77e093bc..e5909df54f09 100644
--- a/src/query/functions/tests/it/scalars/string.rs
+++ b/src/query/functions/tests/it/scalars/string.rs
@@ -86,6 +86,10 @@ fn test_lower(file: &mut impl Write) {
 
 fn test_bit_length(file: &mut impl Write) {
     run_ast(file, "bit_length('latin')", &[]);
+    run_ast(file, "bit_length('CAFÉ')", &[]);
+    run_ast(file, "bit_length('数据库')", &[]);
+    run_ast(file, "bit_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')", &[]);
+    run_ast(file, "bit_length('قاعدة البيانات')", &[]);
     run_ast(file, "bit_length(NULL)", &[]);
     run_ast(file, "bit_length(a)", &[(
         "a",
@@ -95,8 +99,12 @@ fn test_bit_length(file: &mut impl Write) {
 
 fn test_octet_length(file: &mut impl Write) {
     run_ast(file, "octet_length('latin')", &[]);
+    run_ast(file, "octet_length('CAFÉ')", &[]);
+    run_ast(file, "octet_length('数据库')", &[]);
+    run_ast(file, "octet_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')", &[]);
+    run_ast(file, "octet_length('قاعدة البيانات')", &[]);
     run_ast(file, "octet_length(NULL)", &[]);
-    run_ast(file, "length(a)", &[(
+    run_ast(file, "octet_length(a)", &[(
         "a",
         StringType::from_data(vec!["latin", "кириллица", "кириллица and latin"]),
     )]);
@@ -104,6 +112,10 @@ fn test_octet_length(file: &mut impl Write) {
 
 fn test_char_length(file: &mut impl Write) {
     run_ast(file, "char_length('latin')", &[]);
+    run_ast(file, "char_length('CAFÉ')", &[]);
+    run_ast(file, "char_length('数据库')", &[]);
+    run_ast(file, "char_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')", &[]);
+    run_ast(file, "char_length('قاعدة البيانات')", &[]);
     run_ast(file, "char_length(NULL)", &[]);
     run_ast(file, "character_length(a)", &[(
         "a",
@@ -162,6 +174,10 @@ fn test_ascii(file: &mut impl Write) {
 fn test_ltrim(file: &mut impl Write) {
     run_ast(file, "ltrim('   abc   ')", &[]);
     run_ast(file, "ltrim('  ')", &[]);
+    run_ast(file, "ltrim('  你  好  ')", &[]);
+    run_ast(file, "ltrim('  분산 데이터베이스    ')", &[]);
+    run_ast(file, "ltrim('   あなたのことが好きです   ')", &[
+    ]);
     run_ast(file, "ltrim(NULL)", &[]);
     run_ast(file, "ltrim(a)", &[(
         "a",
@@ -172,6 +188,10 @@ fn test_ltrim(file: &mut impl Write) {
 fn test_rtrim(file: &mut impl Write) {
     run_ast(file, "rtrim('   abc   ')", &[]);
     run_ast(file, "rtrim('  ')", &[]);
+    run_ast(file, "rtrim('  你  好  ')", &[]);
+    run_ast(file, "rtrim('  분산 데이터베이스    ')", &[]);
+    run_ast(file, "rtrim('   あなたのことが好きです   ')", &[
+    ]);
     run_ast(file, "rtrim(NULL)", &[]);
     run_ast(file, "rtrim(a)", &[(
         "a",
@@ -187,6 +207,12 @@ fn test_trim_leading(file: &mut impl Write) {
     run_ast(file, "trim_leading(NULL, 'a')", &[]);
     run_ast(file, "trim_leading('aaaaaaaa', NULL)", &[]);
     run_ast(file, "trim_leading('aaaaaaaa', '')", &[]);
+    run_ast(file, "trim_leading('분산 데이터베이스', '분산 ')", &[]);
+    run_ast(
+        file,
+        "trim_leading('あなたのことが好きです', 'あなたの')",
+        &[],
+    );
 
     let table = [
         (
@@ -209,6 +235,12 @@ fn test_trim_trailing(file: &mut impl Write) {
     run_ast(file, "trim_trailing(NULL, 'a')", &[]);
     run_ast(file, "trim_trailing('aaaaaaaa', NULL)", &[]);
     run_ast(file, "trim_trailing('aaaaaaaa', '')", &[]);
+    run_ast(file, "trim_trailing('분산 데이터베이스', '베이스')", &[]);
+    run_ast(
+        file,
+        "trim_trailing('あなたのことが好きです', '好きです')",
+        &[],
+    );
 
     let table = [
         (
@@ -231,6 +263,10 @@ fn test_trim_both(file: &mut impl Write) {
     run_ast(file, "trim_both(NULL, 'a')", &[]);
     run_ast(file, "trim_both('aaaaaaaa', NULL)", &[]);
     run_ast(file, "trim_both('aaaaaaaa', '')", &[]);
+    run_ast(file, "trim_both('  你  好  ', ' ')", &[]);
+    run_ast(file, "trim_both('  분산 데이터베이스    ', ' ')", &[
+    ]);
+    run_ast(file, "trim_both('   あなたのことが好きです   ', ' ')", &[]);
 
     let table = [
         (
@@ -313,6 +349,9 @@ fn test_trim(file: &mut impl Write) {
     // TRIM(<expr>)
     run_ast(file, "trim('   abc   ')", &[]);
     run_ast(file, "trim('  ')", &[]);
+    run_ast(file, "trim('  你  好  ')", &[]);
+    run_ast(file, "trim('  분산 데이터베이스    ')", &[]);
+    run_ast(file, "trim('   あなたのことが好きです   ')", &[]);
     run_ast(file, "trim(NULL)", &[]);
     run_ast(file, "trim(a)", &[(
         "a",
@@ -328,6 +367,11 @@ fn test_trim(file: &mut impl Write) {
 fn test_concat(file: &mut impl Write) {
     run_ast(file, "concat('5', '3', '4')", &[]);
     run_ast(file, "concat(NULL, '3', '4')", &[]);
+    run_ast(
+        file,
+        "concat('忠犬ハチ公', 'CAFÉ', '数据库', 'قاعدة البيانات', 'НОЧЬ НА ОКРАИНЕ МОСКВЫ')",
+        &[],
+    );
     run_ast(file, "concat(a, '3', '4', '5')", &[(
         "a",
         StringType::from_data(vec!["abc", "   abc", "   abc   ", "abc   "]),
@@ -342,6 +386,11 @@ fn test_concat(file: &mut impl Write) {
 
     run_ast(file, "concat_ws('-', '3', null, '4', null, '5')", &[]);
     run_ast(file, "concat_ws(NULL, '3', '4')", &[]);
+    run_ast(
+        file,
+        "concat_ws(',', '忠犬ハチ公', 'CAFÉ', '数据库', 'قاعدة البيانات', 'НОЧЬ НА ОКРАИНЕ МОСКВЫ')",
+        &[],
+    );
     run_ast(file, "concat_ws(a, '3', '4', '5')", &[(
         "a",
         StringType::from_data(vec![",", "-", ",", "-"]),
@@ -498,8 +547,8 @@ fn test_strcmp(file: &mut impl Write) {
     run_ast(file, "strcmp('hii', 'hii')", &[]);
 
     let table = [
-        ("a", StringType::from_data(vec!["hi", "test", "cc"])),
-        ("b", StringType::from_data(vec!["i", "test", "ccb"])),
+        ("a", StringType::from_data(vec!["i", "h", "test", "cc"])),
+        ("b", StringType::from_data(vec!["hi", "hi", "test", "ccb"])),
     ];
     run_ast(file, "strcmp(a, b)", &table);
 }
@@ -508,6 +557,7 @@ fn test_locate(file: &mut impl Write) {
     run_ast(file, "locate('bar', 'foobarbar')", &[]);
     run_ast(file, "locate('', 'foobarbar')", &[]);
     run_ast(file, "locate('', '')", &[]);
+    run_ast(file, "locate('好世', '你好世界')", &[]);
     run_ast(file, "instr('foobarbar', 'bar')", &[]);
     run_ast(file, "instr('foobarbar', '')", &[]);
     run_ast(file, "instr('', '')", &[]);
@@ -516,6 +566,7 @@ fn test_locate(file: &mut impl Write) {
     run_ast(file, "position('' IN '')", &[]);
     run_ast(file, "position('foobarbar' IN 'bar')", &[]);
     run_ast(file, "locate('bar', 'foobarbar', 5)", &[]);
+    run_ast(file, "locate('好世', '你好世界', 1)", &[]);
 
     let table = [
         ("a", StringType::from_data(vec!["bar", "cc", "cc", "q"])),
@@ -570,6 +621,8 @@ fn test_ord(file: &mut impl Write) {
 fn test_repeat(file: &mut impl Write) {
     run_ast(file, "repeat('3', NULL)", &[]);
     run_ast(file, "repeat('3', 5)", &[]);
+    run_ast(file, "repeat('你好世界', 3)", &[]);
+    run_ast(file, "repeat('こんにちは', 2)", &[]);
     run_ast(file, "repeat('3', 1000001)", &[]);
     let table = [("a", StringType::from_data(vec!["a", "b", "c"]))];
     run_ast(file, "repeat(a, 3)", &table);
@@ -584,6 +637,7 @@ fn test_insert(file: &mut impl Write) {
     run_ast(file, "insert('Quadratic', 3, 100, NULL)", &[]);
     run_ast(file, "insert('Quadratic', 3, NULL, 'NULL')", &[]);
     run_ast(file, "insert('Quadratic', NULL, 100, 'NULL')", &[]);
+    run_ast(file, "insert('你好世界', 1, 2, 'こんにちは')", &[]);
     run_ast(file, "insert(NULL, 2, 100, 'NULL')", &[]);
 
     let table = [
@@ -653,6 +707,8 @@ fn test_substr(file: &mut impl Write) {
     run_ast(file, "substr('Sakila' from -4 for 2)", &[]);
     run_ast(file, "substr('sakila' FROM -4)", &[]);
     run_ast(file, "substr('abc',2)", &[]);
+    run_ast(file, "substr('你好世界', 3)", &[]);
+    run_ast(file, "substr('こんにちは', 2)", &[]);
     run_ast(file, "substr('abc', pos, len)", &[
         (
             "pos",
@@ -675,6 +731,8 @@ fn test_split(file: &mut impl Write) {
     run_ast(file, "split('Sakila', 'il')", &[]);
     run_ast(file, "split('sakila', 'a')", &[]);
     run_ast(file, "split('abc','b')", &[]);
+    run_ast(file, "split('你好世界', '好')", &[]);
+    run_ast(file, "split('こんにちは', 'に')", &[]);
     run_ast(file, "split(str, sep)", &[
         (
             "str",
diff --git a/src/query/functions/tests/it/scalars/testdata/arithmetic.txt b/src/query/functions/tests/it/scalars/testdata/arithmetic.txt
index fc61530bf044..2f506b6dfdab 100644
--- a/src/query/functions/tests/it/scalars/testdata/arithmetic.txt
+++ b/src/query/functions/tests/it/scalars/testdata/arithmetic.txt
@@ -69,8 +69,7 @@ evaluation (internal):
 
 ast            : c + 0.5
 raw expr       : plus(c::UInt32, 0.5)
-checked expr   : plus<Decimal(12, 1), Decimal(12, 1)>(to_decimal<UInt32>(12, 1)(c), to_decimal<Decimal(1, 1)>(12, 1)(0.5_d128(1,1)))
-optimized expr : plus<Decimal(12, 1), Decimal(12, 1)>(to_decimal<UInt32>(12, 1)(c), 0.5_d128(12,1))
+checked expr   : plus<UInt32, Decimal(1, 1)>(c, 0.5_d128(1,1))
 evaluation:
 +--------+-----------+----------------+
 |        | c         | Output         |
@@ -138,7 +137,7 @@ evaluation (internal):
 
 ast            : c + e
 raw expr       : plus(c::UInt32, e::Decimal(10, 1))
-checked expr   : plus<Decimal(12, 1), Decimal(12, 1)>(to_decimal<UInt32>(12, 1)(c), to_decimal<Decimal(10, 1)>(12, 1)(e))
+checked expr   : plus<UInt32, Decimal(10, 1)>(c, e)
 evaluation:
 +--------+-----------+----------------+----------------+
 |        | c         | e              | Output         |
@@ -184,7 +183,7 @@ evaluation (internal):
 
 ast            : d2 + e
 raw expr       : plus(d2::UInt8 NULL, e::Decimal(10, 1))
-checked expr   : plus<Decimal(11, 1) NULL, Decimal(11, 1) NULL>(CAST(d2 AS Decimal(11, 1) NULL), CAST(e AS Decimal(11, 1) NULL))
+checked expr   : plus<UInt8 NULL, Decimal(10, 1) NULL>(d2, CAST(e AS Decimal(10, 1) NULL))
 evaluation:
 +--------+------------------+----------------+------------------------+
 |        | d2               | e              | Output                 |
@@ -207,7 +206,7 @@ evaluation (internal):
 
 ast            : d2 + f
 raw expr       : plus(d2::UInt8 NULL, f::Decimal(76, 2))
-checked expr   : plus<Decimal(76, 2) NULL, Decimal(76, 2) NULL>(CAST(d2 AS Decimal(76, 2) NULL), CAST(f AS Decimal(76, 2) NULL))
+checked expr   : plus<UInt8 NULL, Decimal(76, 2) NULL>(d2, CAST(f AS Decimal(76, 2) NULL))
 evaluation:
 +--------+------------------+----------------+-------------------------+
 |        | d2               | f              | Output                  |
@@ -230,7 +229,7 @@ evaluation (internal):
 
 ast            : e + f
 raw expr       : plus(e::Decimal(10, 1), f::Decimal(76, 2))
-checked expr   : plus<Decimal(76, 2), Decimal(76, 2)>(to_decimal<Decimal(10, 1)>(76, 2)(e), f)
+checked expr   : plus<Decimal(10, 1), Decimal(76, 2)>(e, f)
 evaluation:
 +--------+----------------+----------------+-----------------+
 |        | e              | f              | Output          |
@@ -322,8 +321,7 @@ evaluation (internal):
 
 ast            : c - 0.5
 raw expr       : minus(c::UInt32, 0.5)
-checked expr   : minus<Decimal(12, 1), Decimal(12, 1)>(to_decimal<UInt32>(12, 1)(c), to_decimal<Decimal(1, 1)>(12, 1)(0.5_d128(1,1)))
-optimized expr : minus<Decimal(12, 1), Decimal(12, 1)>(to_decimal<UInt32>(12, 1)(c), 0.5_d128(12,1))
+checked expr   : minus<UInt32, Decimal(1, 1)>(c, 0.5_d128(1,1))
 evaluation:
 +--------+-----------+----------------+
 |        | c         | Output         |
@@ -391,7 +389,7 @@ evaluation (internal):
 
 ast            : c - e
 raw expr       : minus(c::UInt32, e::Decimal(10, 1))
-checked expr   : minus<Decimal(12, 1), Decimal(12, 1)>(to_decimal<UInt32>(12, 1)(c), to_decimal<Decimal(10, 1)>(12, 1)(e))
+checked expr   : minus<UInt32, Decimal(10, 1)>(c, e)
 evaluation:
 +--------+-----------+----------------+-----------------+
 |        | c         | e              | Output          |
@@ -437,7 +435,7 @@ evaluation (internal):
 
 ast            : d2 - e
 raw expr       : minus(d2::UInt8 NULL, e::Decimal(10, 1))
-checked expr   : minus<Decimal(11, 1) NULL, Decimal(11, 1) NULL>(CAST(d2 AS Decimal(11, 1) NULL), CAST(e AS Decimal(11, 1) NULL))
+checked expr   : minus<UInt8 NULL, Decimal(10, 1) NULL>(d2, CAST(e AS Decimal(10, 1) NULL))
 evaluation:
 +--------+------------------+----------------+--------------------------+
 |        | d2               | e              | Output                   |
@@ -460,7 +458,7 @@ evaluation (internal):
 
 ast            : d2 - f
 raw expr       : minus(d2::UInt8 NULL, f::Decimal(76, 2))
-checked expr   : minus<Decimal(76, 2) NULL, Decimal(76, 2) NULL>(CAST(d2 AS Decimal(76, 2) NULL), CAST(f AS Decimal(76, 2) NULL))
+checked expr   : minus<UInt8 NULL, Decimal(76, 2) NULL>(d2, CAST(f AS Decimal(76, 2) NULL))
 evaluation:
 +--------+------------------+----------------+--------------------------+
 |        | d2               | f              | Output                   |
@@ -483,7 +481,7 @@ evaluation (internal):
 
 ast            : e - f
 raw expr       : minus(e::Decimal(10, 1), f::Decimal(76, 2))
-checked expr   : minus<Decimal(76, 2), Decimal(76, 2)>(to_decimal<Decimal(10, 1)>(76, 2)(e), f)
+checked expr   : minus<Decimal(10, 1), Decimal(76, 2)>(e, f)
 evaluation:
 +--------+----------------+----------------+------------------+
 |        | e              | f              | Output           |
@@ -759,8 +757,7 @@ evaluation (internal):
 
 ast            : c * 0.5
 raw expr       : multiply(c::UInt32, 0.5)
-checked expr   : multiply<Decimal(11, 0), Decimal(11, 1)>(to_decimal<UInt32>(11, 0)(c), to_decimal<Decimal(1, 1)>(11, 1)(0.5_d128(1,1)))
-optimized expr : multiply<Decimal(11, 0), Decimal(11, 1)>(to_decimal<UInt32>(11, 0)(c), 0.5_d128(11,1))
+checked expr   : multiply<UInt32, Decimal(1, 1)>(c, 0.5_d128(1,1))
 evaluation:
 +--------+-----------+----------------+
 |        | c         | Output         |
@@ -828,7 +825,7 @@ evaluation (internal):
 
 ast            : c * e
 raw expr       : multiply(c::UInt32, e::Decimal(10, 1))
-checked expr   : multiply<Decimal(20, 0), Decimal(20, 1)>(to_decimal<UInt32>(20, 0)(c), to_decimal<Decimal(10, 1)>(20, 1)(e))
+checked expr   : multiply<UInt32, Decimal(10, 1)>(c, e)
 evaluation:
 +--------+-----------+----------------+-----------------+
 |        | c         | e              | Output          |
@@ -874,7 +871,7 @@ evaluation (internal):
 
 ast            : d2 * e
 raw expr       : multiply(d2::UInt8 NULL, e::Decimal(10, 1))
-checked expr   : multiply<Decimal(13, 0) NULL, Decimal(13, 1) NULL>(CAST(d2 AS Decimal(13, 0) NULL), CAST(e AS Decimal(13, 1) NULL))
+checked expr   : multiply<UInt8 NULL, Decimal(10, 1) NULL>(d2, CAST(e AS Decimal(10, 1) NULL))
 evaluation:
 +--------+------------------+----------------+------------------------+
 |        | d2               | e              | Output                 |
@@ -897,7 +894,7 @@ evaluation (internal):
 
 ast            : d2 * f
 raw expr       : multiply(d2::UInt8 NULL, f::Decimal(76, 2))
-checked expr   : multiply<Decimal(76, 0) NULL, Decimal(76, 2) NULL>(CAST(d2 AS Decimal(76, 0) NULL), CAST(f AS Decimal(76, 2) NULL))
+checked expr   : multiply<UInt8 NULL, Decimal(76, 2) NULL>(d2, CAST(f AS Decimal(76, 2) NULL))
 evaluation:
 +--------+------------------+----------------+-------------------------+
 |        | d2               | f              | Output                  |
@@ -920,7 +917,7 @@ evaluation (internal):
 
 ast            : e * f
 raw expr       : multiply(e::Decimal(10, 1), f::Decimal(76, 2))
-checked expr   : multiply<Decimal(76, 1), Decimal(76, 2)>(to_decimal<Decimal(10, 1)>(76, 1)(e), f)
+checked expr   : multiply<Decimal(10, 1), Decimal(76, 2)>(e, f)
 evaluation:
 +--------+----------------+----------------+--------------------+
 |        | e              | f              | Output             |
@@ -943,8 +940,7 @@ evaluation (internal):
 
 ast            : e * 0.5
 raw expr       : multiply(e::Decimal(10, 1), 0.5)
-checked expr   : multiply<Decimal(11, 1), Decimal(11, 1)>(to_decimal<Decimal(10, 1)>(11, 1)(e), to_decimal<Decimal(1, 1)>(11, 1)(0.5_d128(1,1)))
-optimized expr : multiply<Decimal(11, 1), Decimal(11, 1)>(to_decimal<Decimal(10, 1)>(11, 1)(e), 0.5_d128(11,1))
+checked expr   : multiply<Decimal(10, 1), Decimal(1, 1)>(e, 0.5_d128(1,1))
 evaluation:
 +--------+----------------+----------------+
 |        | e              | Output         |
@@ -1035,7 +1031,7 @@ evaluation (internal):
 
 ast            : c / 0.5
 raw expr       : divide(c::UInt32, 0.5)
-checked expr   : divide<Decimal(10, 0), Decimal(1, 1)>(to_decimal<UInt32>(10, 0)(c), 0.5_d128(1,1))
+checked expr   : divide<UInt32, Decimal(1, 1)>(c, 0.5_d128(1,1))
 evaluation:
 +--------+-----------+----------------+
 |        | c         | Output         |
@@ -1134,7 +1130,7 @@ error:
 
 ast            : c / e
 raw expr       : divide(c::UInt32, e::Decimal(10, 1))
-checked expr   : divide<Decimal(10, 0), Decimal(10, 1)>(to_decimal<UInt32>(10, 0)(c), e)
+checked expr   : divide<UInt32, Decimal(10, 1)>(c, e)
 evaluation:
 +--------+-----------+----------------+----------------+
 |        | c         | e              | Output         |
@@ -1180,7 +1176,7 @@ evaluation (internal):
 
 ast            : d2 / e
 raw expr       : divide(d2::UInt8 NULL, e::Decimal(10, 1))
-checked expr   : divide<Decimal(3, 0) NULL, Decimal(10, 1) NULL>(CAST(d2 AS Decimal(3, 0) NULL), CAST(e AS Decimal(10, 1) NULL))
+checked expr   : divide<UInt8 NULL, Decimal(10, 1) NULL>(d2, CAST(e AS Decimal(10, 1) NULL))
 evaluation:
 +--------+------------------+----------------+---------------------+
 |        | d2               | e              | Output              |
@@ -1203,7 +1199,7 @@ evaluation (internal):
 
 ast            : d2 / f
 raw expr       : divide(d2::UInt8 NULL, f::Decimal(76, 2))
-checked expr   : divide<Decimal(39, 0) NULL, Decimal(76, 2) NULL>(CAST(d2 AS Decimal(39, 0) NULL), CAST(f AS Decimal(76, 2) NULL))
+checked expr   : divide<UInt8 NULL, Decimal(76, 2) NULL>(d2, CAST(f AS Decimal(76, 2) NULL))
 evaluation:
 +--------+------------------+----------------+---------------------+
 |        | d2               | f              | Output              |
@@ -1226,7 +1222,7 @@ evaluation (internal):
 
 ast            : e / f
 raw expr       : divide(e::Decimal(10, 1), f::Decimal(76, 2))
-checked expr   : divide<Decimal(39, 1), Decimal(76, 2)>(to_decimal<Decimal(10, 1)>(39, 1)(e), f)
+checked expr   : divide<Decimal(10, 1), Decimal(76, 2)>(e, f)
 evaluation:
 +--------+----------------+----------------+----------------+
 |        | e              | f              | Output         |
diff --git a/src/query/functions/tests/it/scalars/testdata/array.txt b/src/query/functions/tests/it/scalars/testdata/array.txt
index b928a0187d23..41be493ae10d 100644
--- a/src/query/functions/tests/it/scalars/testdata/array.txt
+++ b/src/query/functions/tests/it/scalars/testdata/array.txt
@@ -439,6 +439,78 @@ evaluation (internal):
 +----------+-----------------------+
 
 
+ast            : contains(['5000', '6000', '7000'], string_col)
+raw expr       : contains(array('5000', '6000', '7000'), string_col::String)
+checked expr   : contains<Array(String), String>(array<T0=String><T0, T0, T0>("5000", "6000", "7000"), string_col)
+optimized expr : false
+evaluation:
++--------+-------------+---------+
+|        | string_col  | Output  |
++--------+-------------+---------+
+| Type   | String      | Boolean |
+| Domain | {"1"..="5"} | {FALSE} |
+| Row 0  | '1'         | false   |
+| Row 1  | '2'         | false   |
+| Row 2  | '5'         | false   |
+| Row 3  | '1234'      | false   |
++--------+-------------+---------+
+evaluation (internal):
++------------+-------------------------------------------------------------------+
+| Column     | Data                                                              |
++------------+-------------------------------------------------------------------+
+| string_col | StringColumn { data: 0x31323531323334, offsets: [0, 1, 2, 3, 7] } |
+| Output     | Boolean([0b____0000])                                             |
++------------+-------------------------------------------------------------------+
+
+
+ast            : contains(['1', '5'], string_col)
+raw expr       : contains(array('1', '5'), string_col::String)
+checked expr   : contains<Array(String), String>(array<T0=String><T0, T0>("1", "5"), string_col)
+optimized expr : contains<Array(String), String>(['1', '5'], string_col)
+evaluation:
++--------+-------------+---------------+
+|        | string_col  | Output        |
++--------+-------------+---------------+
+| Type   | String      | Boolean       |
+| Domain | {"1"..="5"} | {FALSE, TRUE} |
+| Row 0  | '1'         | true          |
+| Row 1  | '2'         | false         |
+| Row 2  | '5'         | true          |
+| Row 3  | '1234'      | false         |
++--------+-------------+---------------+
+evaluation (internal):
++------------+-------------------------------------------------------------------+
+| Column     | Data                                                              |
++------------+-------------------------------------------------------------------+
+| string_col | StringColumn { data: 0x31323531323334, offsets: [0, 1, 2, 3, 7] } |
+| Output     | Boolean([0b____0101])                                             |
++------------+-------------------------------------------------------------------+
+
+
+ast            : contains(['15000', '6000', '7000'], string_col)
+raw expr       : contains(array('15000', '6000', '7000'), string_col::String)
+checked expr   : contains<Array(String), String>(array<T0=String><T0, T0, T0>("15000", "6000", "7000"), string_col)
+optimized expr : contains<Array(String), String>(['15000', '6000', '7000'], string_col)
+evaluation:
++--------+-------------+---------------+
+|        | string_col  | Output        |
++--------+-------------+---------------+
+| Type   | String      | Boolean       |
+| Domain | {"1"..="5"} | {FALSE, TRUE} |
+| Row 0  | '1'         | false         |
+| Row 1  | '2'         | false         |
+| Row 2  | '5'         | false         |
+| Row 3  | '1234'      | false         |
++--------+-------------+---------------+
+evaluation (internal):
++------------+-------------------------------------------------------------------+
+| Column     | Data                                                              |
++------------+-------------------------------------------------------------------+
+| string_col | StringColumn { data: 0x31323531323334, offsets: [0, 1, 2, 3, 7] } |
+| Output     | Boolean([0b____0000])                                             |
++------------+-------------------------------------------------------------------+
+
+
 ast            : contains([1,2,null], nullable_col)
 raw expr       : contains(array(1, 2, NULL), nullable_col::Int64 NULL)
 checked expr   : contains<T0=Int64 NULL><Array(T0), T0>(CAST(array<T0=UInt8 NULL><T0, T0, T0>(CAST(1_u8 AS UInt8 NULL), CAST(2_u8 AS UInt8 NULL), CAST(NULL AS UInt8 NULL)) AS Array(Int64 NULL)), nullable_col)
diff --git a/src/query/functions/tests/it/scalars/testdata/comparison.txt b/src/query/functions/tests/it/scalars/testdata/comparison.txt
index 97194c73aef2..4949d145c888 100644
--- a/src/query/functions/tests/it/scalars/testdata/comparison.txt
+++ b/src/query/functions/tests/it/scalars/testdata/comparison.txt
@@ -1019,6 +1019,28 @@ evaluation (internal):
 +--------+-----------------------------------------------------------------------------------------------------------------------------------------------+
 
 
+ast            : col > 'efg'
+raw expr       : gt(col::String, 'efg')
+checked expr   : gt<String, String>(col, "efg")
+optimized expr : false
+evaluation:
++--------+-----------------+---------+
+|        | col             | Output  |
++--------+-----------------+---------+
+| Type   | String          | Boolean |
+| Domain | {"bcd"..="efg"} | {FALSE} |
+| Row 0  | 'bcd'           | false   |
+| Row 1  | 'efg'           | false   |
++--------+-----------------+---------+
+evaluation (internal):
++--------+-----------------------------------------------------------+
+| Column | Data                                                      |
++--------+-----------------------------------------------------------+
+| col    | StringColumn { data: 0x626364656667, offsets: [0, 3, 6] } |
+| Output | Boolean([0b______00])                                     |
++--------+-----------------------------------------------------------+
+
+
 ast            : '2'>='1'
 raw expr       : gte('2', '1')
 checked expr   : gte<String, String>("2", "1")
diff --git a/src/query/functions/tests/it/scalars/testdata/function_list.txt b/src/query/functions/tests/it/scalars/testdata/function_list.txt
index d6faa40fe3ff..66017085b9dc 100644
--- a/src/query/functions/tests/it/scalars/testdata/function_list.txt
+++ b/src/query/functions/tests/it/scalars/testdata/function_list.txt
@@ -7,7 +7,8 @@ array_slice -> slice
 bitmap_and_not -> bitmap_not
 bitmap_cardinality -> bitmap_count
 ceiling -> ceil
-character_length -> char_length
+char_length -> length
+character_length -> length
 current_timestamp -> now
 date_format -> to_string
 hex -> to_hex
@@ -16,21 +17,21 @@ ipv4_num_to_string -> inet_ntoa
 ipv4_string_to_num -> inet_aton
 json_to_string -> to_string
 lcase -> lower
-length_utf8 -> char_length
+length_utf8 -> length
 mid -> substr
 mod -> modulo
 neg -> minus
 negate -> minus
 object_keys -> json_object_keys
-octet_length -> length
 remove_nullable -> assume_not_null
 rlike -> regexp
 sha1 -> sha
 siphash -> siphash64
 str_to_date -> to_date
 str_to_timestamp -> to_timestamp
+substr_utf8 -> substr
 substring -> substr
-substring_utf8 -> substr_utf8
+substring_utf8 -> substr
 subtract -> minus
 to_datetime -> to_timestamp
 to_text -> to_string
@@ -785,8 +786,6 @@ Functions overloads:
 20 ceil(Float64 NULL) :: Float64 NULL
 0 char FACTORY
 1 char FACTORY
-0 char_length(String) :: UInt64
-1 char_length(String NULL) :: UInt64 NULL
 0 check_json(Variant) :: String NULL
 1 check_json(Variant NULL) :: String NULL
 2 check_json(String) :: String NULL
@@ -2774,6 +2773,8 @@ Functions overloads:
 0 now() :: Timestamp
 0 oct(Int64) :: String
 1 oct(Int64 NULL) :: String NULL
+0 octet_length(String) :: UInt64
+1 octet_length(String NULL) :: UInt64 NULL
 0 or(Boolean, Boolean) :: Boolean
 1 or(Boolean NULL, Boolean NULL) :: Boolean NULL
 0 ord(String) :: UInt64
@@ -3165,10 +3166,6 @@ Functions overloads:
 1 substr(String NULL, Int64 NULL) :: String NULL
 2 substr(String, Int64, UInt64) :: String
 3 substr(String NULL, Int64 NULL, UInt64 NULL) :: String NULL
-0 substr_utf8(String, Int64) :: String
-1 substr_utf8(String NULL, Int64 NULL) :: String NULL
-2 substr_utf8(String, Int64, UInt64) :: String
-3 substr_utf8(String NULL, Int64 NULL, UInt64 NULL) :: String NULL
 0 subtract_days(Date, Int64) :: Date
 1 subtract_days(Date NULL, Int64 NULL) :: Date NULL
 2 subtract_days(Timestamp, Int64) :: Timestamp
diff --git a/src/query/functions/tests/it/scalars/testdata/math.txt b/src/query/functions/tests/it/scalars/testdata/math.txt
index 1f46c3cf5bd0..0f664b7f9ee8 100644
--- a/src/query/functions/tests/it/scalars/testdata/math.txt
+++ b/src/query/functions/tests/it/scalars/testdata/math.txt
@@ -464,7 +464,7 @@ output         : 120
 
 ast            : truncate(0)(10.28*100, 0)
 raw expr       : truncate(0)(multiply(10.28, 100), 0)
-checked expr   : truncate<Decimal(7, 2), UInt8>(0)(multiply<Decimal(7, 2), Decimal(7, 0)>(to_decimal<Decimal(4, 2)>(7, 2)(10.28_d128(4,2)), to_decimal<UInt8>(7, 0)(100_u8)), 0_u8)
+checked expr   : truncate<Decimal(7, 2), UInt8>(0)(multiply<Decimal(4, 2), UInt8>(10.28_d128(4,2), 100_u8), 0_u8)
 optimized expr : 1028_d128(7,0)
 output type    : Decimal(7, 0)
 output domain  : {1028..=1028}
diff --git a/src/query/functions/tests/it/scalars/testdata/string.txt b/src/query/functions/tests/it/scalars/testdata/string.txt
index 38b4fb241f09..fc612631b73a 100644
--- a/src/query/functions/tests/it/scalars/testdata/string.txt
+++ b/src/query/functions/tests/it/scalars/testdata/string.txt
@@ -123,6 +123,42 @@ output domain  : {40..=40}
 output         : 40
 
 
+ast            : bit_length('CAFÉ')
+raw expr       : bit_length('CAFÉ')
+checked expr   : bit_length<String>("CAFÉ")
+optimized expr : 40_u64
+output type    : UInt64
+output domain  : {40..=40}
+output         : 40
+
+
+ast            : bit_length('数据库')
+raw expr       : bit_length('数据库')
+checked expr   : bit_length<String>("数据库")
+optimized expr : 72_u64
+output type    : UInt64
+output domain  : {72..=72}
+output         : 72
+
+
+ast            : bit_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+raw expr       : bit_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+checked expr   : bit_length<String>("НОЧЬ НА ОКРАИНЕ МОСКВЫ")
+optimized expr : 328_u64
+output type    : UInt64
+output domain  : {328..=328}
+output         : 328
+
+
+ast            : bit_length('قاعدة البيانات')
+raw expr       : bit_length('قاعدة البيانات')
+checked expr   : bit_length<String>("قاعدة البيانات")
+optimized expr : 216_u64
+output type    : UInt64
+output domain  : {216..=216}
+output         : 216
+
+
 ast            : bit_length(NULL)
 raw expr       : bit_length(NULL)
 checked expr   : bit_length<String NULL>(CAST(NULL AS String NULL))
@@ -156,25 +192,61 @@ evaluation (internal):
 
 ast            : octet_length('latin')
 raw expr       : octet_length('latin')
-checked expr   : length<String>("latin")
+checked expr   : octet_length<String>("latin")
+optimized expr : 5_u64
+output type    : UInt64
+output domain  : {5..=5}
+output         : 5
+
+
+ast            : octet_length('CAFÉ')
+raw expr       : octet_length('CAFÉ')
+checked expr   : octet_length<String>("CAFÉ")
 optimized expr : 5_u64
 output type    : UInt64
 output domain  : {5..=5}
 output         : 5
 
 
+ast            : octet_length('数据库')
+raw expr       : octet_length('数据库')
+checked expr   : octet_length<String>("数据库")
+optimized expr : 9_u64
+output type    : UInt64
+output domain  : {9..=9}
+output         : 9
+
+
+ast            : octet_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+raw expr       : octet_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+checked expr   : octet_length<String>("НОЧЬ НА ОКРАИНЕ МОСКВЫ")
+optimized expr : 41_u64
+output type    : UInt64
+output domain  : {41..=41}
+output         : 41
+
+
+ast            : octet_length('قاعدة البيانات')
+raw expr       : octet_length('قاعدة البيانات')
+checked expr   : octet_length<String>("قاعدة البيانات")
+optimized expr : 27_u64
+output type    : UInt64
+output domain  : {27..=27}
+output         : 27
+
+
 ast            : octet_length(NULL)
 raw expr       : octet_length(NULL)
-checked expr   : length<Variant NULL>(CAST(NULL AS Variant NULL))
+checked expr   : octet_length<String NULL>(CAST(NULL AS String NULL))
 optimized expr : NULL
-output type    : UInt32 NULL
+output type    : UInt64 NULL
 output domain  : {NULL}
 output         : NULL
 
 
-ast            : length(a)
-raw expr       : length(a::String)
-checked expr   : length<String>(a)
+ast            : octet_length(a)
+raw expr       : octet_length(a::String)
+checked expr   : octet_length<String>(a)
 evaluation:
 +--------+-----------------------------------+----------------------------+
 |        | a                                 | Output                     |
@@ -196,35 +268,71 @@ evaluation (internal):
 
 ast            : char_length('latin')
 raw expr       : char_length('latin')
-checked expr   : char_length<String>("latin")
+checked expr   : length<String>("latin")
 optimized expr : 5_u64
 output type    : UInt64
 output domain  : {5..=5}
 output         : 5
 
 
+ast            : char_length('CAFÉ')
+raw expr       : char_length('CAFÉ')
+checked expr   : length<String>("CAFÉ")
+optimized expr : 4_u64
+output type    : UInt64
+output domain  : {4..=4}
+output         : 4
+
+
+ast            : char_length('数据库')
+raw expr       : char_length('数据库')
+checked expr   : length<String>("数据库")
+optimized expr : 3_u64
+output type    : UInt64
+output domain  : {3..=3}
+output         : 3
+
+
+ast            : char_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+raw expr       : char_length('НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+checked expr   : length<String>("НОЧЬ НА ОКРАИНЕ МОСКВЫ")
+optimized expr : 22_u64
+output type    : UInt64
+output domain  : {22..=22}
+output         : 22
+
+
+ast            : char_length('قاعدة البيانات')
+raw expr       : char_length('قاعدة البيانات')
+checked expr   : length<String>("قاعدة البيانات")
+optimized expr : 14_u64
+output type    : UInt64
+output domain  : {14..=14}
+output         : 14
+
+
 ast            : char_length(NULL)
 raw expr       : char_length(NULL)
-checked expr   : char_length<String NULL>(CAST(NULL AS String NULL))
+checked expr   : length<Variant NULL>(CAST(NULL AS Variant NULL))
 optimized expr : NULL
-output type    : UInt64 NULL
+output type    : UInt32 NULL
 output domain  : {NULL}
 output         : NULL
 
 
 ast            : character_length(a)
 raw expr       : character_length(a::String)
-checked expr   : char_length<String>(a)
-evaluation:
-+--------+-----------------------------------+---------+
-|        | a                                 | Output  |
-+--------+-----------------------------------+---------+
-| Type   | String                            | UInt64  |
-| Domain | {"latin"..="кириллица and latin"} | Unknown |
-| Row 0  | 'latin'                           | 5       |
-| Row 1  | 'кириллица'                       | 9       |
-| Row 2  | 'кириллица and latin'             | 19      |
-+--------+-----------------------------------+---------+
+checked expr   : length<String>(a)
+evaluation:
++--------+-----------------------------------+----------------------------+
+|        | a                                 | Output                     |
++--------+-----------------------------------+----------------------------+
+| Type   | String                            | UInt64                     |
+| Domain | {"latin"..="кириллица and latin"} | {0..=18446744073709551615} |
+| Row 0  | 'latin'                           | 5                          |
+| Row 1  | 'кириллица'                       | 9                          |
+| Row 2  | 'кириллица and latin'             | 19                         |
++--------+-----------------------------------+----------------------------+
 evaluation (internal):
 +--------+----------------------------------------------------------------------------------------------------------------------------------------------------------+
 | Column | Data                                                                                                                                                     |
@@ -395,28 +503,28 @@ output         : ''
 ast            : reverse('你好')
 raw expr       : reverse('你好')
 checked expr   : reverse<String>("你好")
-optimized expr : BDA5E5A0BDE4
+optimized expr : "好你"
 output type    : String
-output domain  : {"��堽�"..="��堽�"}
-output         : BDA5E5A0BDE4
+output domain  : {"好你"..="好你"}
+output         : '好你'
 
 
 ast            : reverse('ß😀山')
 raw expr       : reverse('ß😀山')
 checked expr   : reverse<String>("ß😀山")
-optimized expr : B1B1E580989FF09FC3
+optimized expr : "山😀ß"
 output type    : String
-output domain  : {"��倘���"..="��倘���"}
-output         : B1B1E580989FF09FC3
+output domain  : {"山😀ß"..="山😀ß"}
+output         : '山😀ß'
 
 
 ast            : reverse('Dobrý den')
 raw expr       : reverse('Dobrý den')
 checked expr   : reverse<String>("Dobrý den")
-optimized expr : 6E656420BDC372626F44
+optimized expr : "ned ýrboD"
 output type    : String
-output domain  : {"ned ��rboD"..="ned ��rboD"}
-output         : 6E656420BDC372626F44
+output domain  : {"ned ýrboD"..="ned ýrboD"}
+output         : 'ned ýrboD'
 
 
 ast            : reverse(Null)
@@ -575,6 +683,33 @@ output domain  : {""..=""}
 output         : ''
 
 
+ast            : ltrim('  你  好  ')
+raw expr       : ltrim('  你  好  ')
+checked expr   : ltrim<String>("  你  好  ")
+optimized expr : "你  好  "
+output type    : String
+output domain  : {"你  好  "..="你  好  "}
+output         : '你  好  '
+
+
+ast            : ltrim('  분산 데이터베이스    ')
+raw expr       : ltrim('  분산 데이터베이스    ')
+checked expr   : ltrim<String>("  분산 데이터베이스    ")
+optimized expr : "분산 데이터베이스    "
+output type    : String
+output domain  : {"분산 데이터베이스    "..="분산 데이터베이스    "}
+output         : '분산 데이터베이스    '
+
+
+ast            : ltrim('   あなたのことが好きです   ')
+raw expr       : ltrim('   あなたのことが好きです   ')
+checked expr   : ltrim<String>("   あなたのことが好きです   ")
+optimized expr : "あなたのことが好きです   "
+output type    : String
+output domain  : {"あなたのことが好きです   "..="あなたのことが好きです   "}
+output         : 'あなたのことが好きです   '
+
+
 ast            : ltrim(NULL)
 raw expr       : ltrim(NULL)
 checked expr   : ltrim<String NULL>(CAST(NULL AS String NULL))
@@ -625,6 +760,33 @@ output domain  : {""..=""}
 output         : ''
 
 
+ast            : rtrim('  你  好  ')
+raw expr       : rtrim('  你  好  ')
+checked expr   : rtrim<String>("  你  好  ")
+optimized expr : "  你  好"
+output type    : String
+output domain  : {"  你  好"..="  你  好"}
+output         : '  你  好'
+
+
+ast            : rtrim('  분산 데이터베이스    ')
+raw expr       : rtrim('  분산 데이터베이스    ')
+checked expr   : rtrim<String>("  분산 데이터베이스    ")
+optimized expr : "  분산 데이터베이스"
+output type    : String
+output domain  : {"  분산 데이터베이스"..="  분산 데이터베이스"}
+output         : '  분산 데이터베이스'
+
+
+ast            : rtrim('   あなたのことが好きです   ')
+raw expr       : rtrim('   あなたのことが好きです   ')
+checked expr   : rtrim<String>("   あなたのことが好きです   ")
+optimized expr : "   あなたのことが好きです"
+output type    : String
+output domain  : {"   あなたのことが好きです"..="   あなたのことが好きです"}
+output         : '   あなたのことが好きです'
+
+
 ast            : rtrim(NULL)
 raw expr       : rtrim(NULL)
 checked expr   : rtrim<String NULL>(CAST(NULL AS String NULL))
@@ -720,6 +882,24 @@ output domain  : {"aaaaaaaa"..="aaaaaaaa"}
 output         : 'aaaaaaaa'
 
 
+ast            : trim_leading('분산 데이터베이스', '분산 ')
+raw expr       : trim_leading('분산 데이터베이스', '분산 ')
+checked expr   : trim_leading<String, String>("분산 데이터베이스", "분산 ")
+optimized expr : "데이터베이스"
+output type    : String
+output domain  : {"데이터베이스"..="데이터베이스"}
+output         : '데이터베이스'
+
+
+ast            : trim_leading('あなたのことが好きです', 'あなたの')
+raw expr       : trim_leading('あなたのことが好きです', 'あなたの')
+checked expr   : trim_leading<String, String>("あなたのことが好きです", "あなたの")
+optimized expr : "ことが好きです"
+output type    : String
+output domain  : {"ことが好きです"..="ことが好きです"}
+output         : 'ことが好きです'
+
+
 ast            : trim_leading(a, 'a')
 raw expr       : trim_leading(a::String, 'a')
 checked expr   : trim_leading<String, String>(a, "a")
@@ -853,6 +1033,24 @@ output domain  : {"aaaaaaaa"..="aaaaaaaa"}
 output         : 'aaaaaaaa'
 
 
+ast            : trim_trailing('분산 데이터베이스', '베이스')
+raw expr       : trim_trailing('분산 데이터베이스', '베이스')
+checked expr   : trim_trailing<String, String>("분산 데이터베이스", "베이스")
+optimized expr : "분산 데이터"
+output type    : String
+output domain  : {"분산 데이터"..="분산 데이터"}
+output         : '분산 데이터'
+
+
+ast            : trim_trailing('あなたのことが好きです', '好きです')
+raw expr       : trim_trailing('あなたのことが好きです', '好きです')
+checked expr   : trim_trailing<String, String>("あなたのことが好きです", "好きです")
+optimized expr : "あなたのことが"
+output type    : String
+output domain  : {"あなたのことが"..="あなたのことが"}
+output         : 'あなたのことが'
+
+
 ast            : trim_trailing(a, 'b')
 raw expr       : trim_trailing(a::String, 'b')
 checked expr   : trim_trailing<String, String>(a, "b")
@@ -986,6 +1184,33 @@ output domain  : {"aaaaaaaa"..="aaaaaaaa"}
 output         : 'aaaaaaaa'
 
 
+ast            : trim_both('  你  好  ', ' ')
+raw expr       : trim_both('  你  好  ', ' ')
+checked expr   : trim_both<String, String>("  你  好  ", " ")
+optimized expr : "你  好"
+output type    : String
+output domain  : {"你  好"..="你  好"}
+output         : '你  好'
+
+
+ast            : trim_both('  분산 데이터베이스    ', ' ')
+raw expr       : trim_both('  분산 데이터베이스    ', ' ')
+checked expr   : trim_both<String, String>("  분산 데이터베이스    ", " ")
+optimized expr : "분산 데이터베이스"
+output type    : String
+output domain  : {"분산 데이터베이스"..="분산 데이터베이스"}
+output         : '분산 데이터베이스'
+
+
+ast            : trim_both('   あなたのことが好きです   ', ' ')
+raw expr       : trim_both('   あなたのことが好きです   ', ' ')
+checked expr   : trim_both<String, String>("   あなたのことが好きです   ", " ")
+optimized expr : "あなたのことが好きです"
+output type    : String
+output domain  : {"あなたのことが好きです"..="あなたのことが好きです"}
+output         : 'あなたのことが好きです'
+
+
 ast            : trim_both(a, 'a')
 raw expr       : trim_both(a::String, 'a')
 checked expr   : trim_both<String, String>(a, "a")
@@ -1074,6 +1299,33 @@ output domain  : {""..=""}
 output         : ''
 
 
+ast            : trim('  你  好  ')
+raw expr       : trim('  你  好  ')
+checked expr   : trim<String>("  你  好  ")
+optimized expr : "你  好"
+output type    : String
+output domain  : {"你  好"..="你  好"}
+output         : '你  好'
+
+
+ast            : trim('  분산 데이터베이스    ')
+raw expr       : trim('  분산 데이터베이스    ')
+checked expr   : trim<String>("  분산 데이터베이스    ")
+optimized expr : "분산 데이터베이스"
+output type    : String
+output domain  : {"분산 데이터베이스"..="분산 데이터베이스"}
+output         : '분산 데이터베이스'
+
+
+ast            : trim('   あなたのことが好きです   ')
+raw expr       : trim('   あなたのことが好きです   ')
+checked expr   : trim<String>("   あなたのことが好きです   ")
+optimized expr : "あなたのことが好きです"
+output type    : String
+output domain  : {"あなたのことが好きです"..="あなたのことが好きです"}
+output         : 'あなたのことが好きです'
+
+
 ast            : trim(NULL)
 raw expr       : trim(NULL)
 checked expr   : trim<String NULL>(CAST(NULL AS String NULL))
@@ -1553,6 +1805,15 @@ output domain  : {NULL}
 output         : NULL
 
 
+ast            : concat('忠犬ハチ公', 'CAFÉ', '数据库', 'قاعدة البيانات', 'НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+raw expr       : concat('忠犬ハチ公', 'CAFÉ', '数据库', 'قاعدة البيانات', 'НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+checked expr   : concat<String, String, String, String, String>("忠犬ハチ公", "CAFÉ", "数据库", "قاعدة البيانات", "НОЧЬ НА ОКРАИНЕ МОСКВЫ")
+optimized expr : "忠犬ハチ公CAFÉ数据库قاعدة البياناتНОЧЬ НА ОКРАИНЕ МОСКВЫ"
+output type    : String
+output domain  : {"忠犬ハチ公CAFÉ数据库قاعدة البياناتНОЧЬ НА ОКРАИНЕ МОСКВЫ"..="忠犬ハチ公CAFÉ数据库قاعدة البياناتНОЧЬ НА ОКРАИНЕ МОСКВЫ"}
+output         : '忠犬ハチ公CAFÉ数据库قاعدة البياناتНОЧЬ НА ОКРАИНЕ МОСКВЫ'
+
+
 ast            : concat(a, '3', '4', '5')
 raw expr       : concat(a::String, '3', '4', '5')
 checked expr   : concat<String, String, String, String>(a, "3", "4", "5")
@@ -1618,6 +1879,15 @@ output domain  : {NULL}
 output         : NULL
 
 
+ast            : concat_ws(',', '忠犬ハチ公', 'CAFÉ', '数据库', 'قاعدة البيانات', 'НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+raw expr       : concat_ws(',', '忠犬ハチ公', 'CAFÉ', '数据库', 'قاعدة البيانات', 'НОЧЬ НА ОКРАИНЕ МОСКВЫ')
+checked expr   : concat_ws<String, String, String, String, String, String>(",", "忠犬ハチ公", "CAFÉ", "数据库", "قاعدة البيانات", "НОЧЬ НА ОКРАИНЕ МОСКВЫ")
+optimized expr : "忠犬ハチ公,CAFÉ,数据库,قاعدة البيانات,НОЧЬ НА ОКРАИНЕ МОСКВЫ"
+output type    : String
+output domain  : {"忠犬ハチ公,CAFÉ,数据库,قاعدة البيانات,НОЧЬ НА ОКРАИНЕ МОСКВЫ"..="忠犬ハチ公,CAFÉ,数据库,قاعدة البيانات,НОЧЬ НА ОКРАИНЕ МОСКВЫ"}
+output         : '忠犬ハチ公,CAFÉ,数据库,قاعدة البيانات,НОЧЬ НА ОКРАИНЕ МОСКВЫ'
+
+
 ast            : concat_ws(a, '3', '4', '5')
 raw expr       : concat_ws(a::String, '3', '4', '5')
 checked expr   : concat_ws<String, String, String, String>(a, "3", "4", "5")
@@ -2430,18 +2700,19 @@ evaluation:
 +--------+-----------------+------------------+--------------+
 | Type   | String          | String           | Int8         |
 | Domain | {"cc"..="test"} | {"ccb"..="test"} | {-128..=127} |
-| Row 0  | 'hi'            | 'i'              | 1            |
-| Row 1  | 'test'          | 'test'           | 0            |
-| Row 2  | 'cc'            | 'ccb'            | -1           |
+| Row 0  | 'i'             | 'hi'             | 1            |
+| Row 1  | 'h'             | 'hi'             | -1           |
+| Row 2  | 'test'          | 'test'           | 0            |
+| Row 3  | 'cc'            | 'ccb'            | -1           |
 +--------+-----------------+------------------+--------------+
 evaluation (internal):
-+--------+------------------------------------------------------------------+
-| Column | Data                                                             |
-+--------+------------------------------------------------------------------+
-| a      | StringColumn { data: 0x6869746573746363, offsets: [0, 2, 6, 8] } |
-| b      | StringColumn { data: 0x6974657374636362, offsets: [0, 1, 5, 8] } |
-| Output | Int8([1, 0, -1])                                                 |
-+--------+------------------------------------------------------------------+
++--------+----------------------------------------------------------------------------+
+| Column | Data                                                                       |
++--------+----------------------------------------------------------------------------+
+| a      | StringColumn { data: 0x6968746573746363, offsets: [0, 1, 2, 6, 8] }        |
+| b      | StringColumn { data: 0x6869686974657374636362, offsets: [0, 2, 4, 8, 11] } |
+| Output | Int8([1, -1, 0, -1])                                                       |
++--------+----------------------------------------------------------------------------+
 
 
 ast            : locate('bar', 'foobarbar')
@@ -2471,6 +2742,15 @@ output domain  : {1..=1}
 output         : 1
 
 
+ast            : locate('好世', '你好世界')
+raw expr       : locate('好世', '你好世界')
+checked expr   : locate<String, String>("好世", "你好世界")
+optimized expr : 2_u64
+output type    : UInt64
+output domain  : {2..=2}
+output         : 2
+
+
 ast            : instr('foobarbar', 'bar')
 raw expr       : instr('foobarbar', 'bar')
 checked expr   : instr<String, String>("foobarbar", "bar")
@@ -2543,6 +2823,15 @@ output domain  : {7..=7}
 output         : 7
 
 
+ast            : locate('好世', '你好世界', 1)
+raw expr       : locate('好世', '你好世界', 1)
+checked expr   : locate<String, String, UInt64>("好世", "你好世界", to_uint64<UInt8>(1_u8))
+optimized expr : 2_u64
+output type    : UInt64
+output domain  : {2..=2}
+output         : 2
+
+
 ast            : locate(a, b, c)
 raw expr       : locate(a::String, b::String, c::UInt8)
 checked expr   : locate<String, String, UInt64>(a, b, to_uint64<UInt8>(c))
@@ -2571,17 +2860,17 @@ evaluation (internal):
 ast            : char(65,66,67)
 raw expr       : char(65, 66, 67)
 checked expr   : char<UInt8, UInt8, UInt8>(65_u8, 66_u8, 67_u8)
-optimized expr : "ABC"
-output type    : String
-output domain  : {"ABC"..="ABC"}
-output         : 'ABC'
+optimized expr : 414243
+output type    : Binary
+output domain  : Undefined
+output         : 414243
 
 
 ast            : char(65, null)
 raw expr       : char(65, NULL)
 checked expr   : char<UInt8 NULL, UInt8 NULL>(CAST(65_u8 AS UInt8 NULL), CAST(NULL AS UInt8 NULL))
 optimized expr : NULL
-output type    : String NULL
+output type    : Binary NULL
 output domain  : {NULL}
 output         : NULL
 
@@ -2590,14 +2879,14 @@ ast            : char(a, b, c)
 raw expr       : char(a::UInt8, b::UInt8, c::UInt8)
 checked expr   : char<UInt8, UInt8, UInt8>(a, b, c)
 evaluation:
-+--------+-----------+-----------+-----------+--------+
-|        | a         | b         | c         | Output |
-+--------+-----------+-----------+-----------+--------+
-| Type   | UInt8     | UInt8     | UInt8     | String |
-| Domain | {66..=67} | {98..=99} | {68..=69} | {""..} |
-| Row 0  | 66        | 98        | 68        | 'BbD'  |
-| Row 1  | 67        | 99        | 69        | 'CcE'  |
-+--------+-----------+-----------+-----------+--------+
++--------+-----------+-----------+-----------+-----------+
+|        | a         | b         | c         | Output    |
++--------+-----------+-----------+-----------+-----------+
+| Type   | UInt8     | UInt8     | UInt8     | Binary    |
+| Domain | {66..=67} | {98..=99} | {68..=69} | Undefined |
+| Row 0  | 66        | 98        | 68        | 426244    |
+| Row 1  | 67        | 99        | 69        | 436345    |
++--------+-----------+-----------+-----------+-----------+
 evaluation (internal):
 +--------+-----------------------------------------------------------+
 | Column | Data                                                      |
@@ -2605,7 +2894,7 @@ evaluation (internal):
 | a      | UInt8([66, 67])                                           |
 | b      | UInt8([98, 99])                                           |
 | c      | UInt8([68, 69])                                           |
-| Output | StringColumn { data: 0x426244436345, offsets: [0, 3, 6] } |
+| Output | BinaryColumn { data: 0x426244436345, offsets: [0, 3, 6] } |
 +--------+-----------------------------------------------------------+
 
 
@@ -2613,14 +2902,14 @@ ast            : char(a2, b, c)
 raw expr       : char(a2::UInt8 NULL, b::UInt8, c::UInt8)
 checked expr   : char<UInt8 NULL, UInt8 NULL, UInt8 NULL>(a2, CAST(b AS UInt8 NULL), CAST(c AS UInt8 NULL))
 evaluation:
-+--------+-----------+-----------+--------------------+-----------------+
-|        | b         | c         | a2                 | Output          |
-+--------+-----------+-----------+--------------------+-----------------+
-| Type   | UInt8     | UInt8     | UInt8 NULL         | String NULL     |
-| Domain | {98..=99} | {68..=69} | {66..=67} ∪ {NULL} | {""..} ∪ {NULL} |
-| Row 0  | 98        | 68        | 66                 | 'BbD'           |
-| Row 1  | 99        | 69        | NULL               | NULL            |
-+--------+-----------+-----------+--------------------+-----------------+
++--------+-----------+-----------+--------------------+--------------------+
+|        | b         | c         | a2                 | Output             |
++--------+-----------+-----------+--------------------+--------------------+
+| Type   | UInt8     | UInt8     | UInt8 NULL         | Binary NULL        |
+| Domain | {98..=99} | {68..=69} | {66..=67} ∪ {NULL} | Undefined ∪ {NULL} |
+| Row 0  | 98        | 68        | 66                 | 426244             |
+| Row 1  | 99        | 69        | NULL               | NULL               |
++--------+-----------+-----------+--------------------+--------------------+
 evaluation (internal):
 +--------+--------------------------------------------------------------------------------------------------------------+
 | Column | Data                                                                                                         |
@@ -2628,7 +2917,7 @@ evaluation (internal):
 | b      | UInt8([98, 99])                                                                                              |
 | c      | UInt8([68, 69])                                                                                              |
 | a2     | NullableColumn { column: UInt8([66, 67]), validity: [0b______01] }                                           |
-| Output | NullableColumn { column: StringColumn { data: 0x426244436345, offsets: [0, 3, 6] }, validity: [0b______01] } |
+| Output | NullableColumn { column: BinaryColumn { data: 0x426244436345, offsets: [0, 3, 6] }, validity: [0b______01] } |
 +--------+--------------------------------------------------------------------------------------------------------------+
 
 
@@ -2639,8 +2928,8 @@ error:
   | ^^^^^^^^ no overload satisfies `char(UInt16)`
 
 has tried possible overloads:
-  char(UInt8) :: String            : unable to unify `UInt16` with `UInt8`
-  char(UInt8 NULL) :: String NULL  : unable to unify `UInt16` with `UInt8`
+  char(UInt8) :: Binary            : unable to unify `UInt16` with `UInt8`
+  char(UInt8 NULL) :: Binary NULL  : unable to unify `UInt16` with `UInt8`
 
 
 
@@ -2757,6 +3046,24 @@ output domain  : {"33333"..="33333"}
 output         : '33333'
 
 
+ast            : repeat('你好世界', 3)
+raw expr       : repeat('你好世界', 3)
+checked expr   : repeat<String, UInt64>("你好世界", to_uint64<UInt8>(3_u8))
+optimized expr : "你好世界你好世界你好世界"
+output type    : String
+output domain  : {"你好世界你好世界你好世界"..="你好世界你好世界你好世界"}
+output         : '你好世界你好世界你好世界'
+
+
+ast            : repeat('こんにちは', 2)
+raw expr       : repeat('こんにちは', 2)
+checked expr   : repeat<String, UInt64>("こんにちは", to_uint64<UInt8>(2_u8))
+optimized expr : "こんにちはこんにちは"
+output type    : String
+output domain  : {"こんにちはこんにちは"..="こんにちはこんにちは"}
+output         : 'こんにちはこんにちは'
+
+
 error: 
   --> SQL:1:1
   |
@@ -2858,6 +3165,15 @@ output domain  : {NULL}
 output         : NULL
 
 
+ast            : insert('你好世界', 1, 2, 'こんにちは')
+raw expr       : insert('你好世界', 1, 2, 'こんにちは')
+checked expr   : insert<String, Int64, Int64, String>("你好世界", to_int64<UInt8>(1_u8), to_int64<UInt8>(2_u8), "こんにちは")
+optimized expr : "こんにちは世界"
+output type    : String
+output domain  : {"こんにちは世界"..="こんにちは世界"}
+output         : 'こんにちは世界'
+
+
 ast            : insert(NULL, 2, 100, 'NULL')
 raw expr       : insert(NULL, 2, 100, 'NULL')
 checked expr   : insert<String NULL, Int64 NULL, Int64 NULL, String NULL>(CAST(NULL AS String NULL), CAST(2_u8 AS Int64 NULL), CAST(100_u8 AS Int64 NULL), CAST("NULL" AS String NULL))
@@ -3124,6 +3440,24 @@ output domain  : {"bc"..="bc"}
 output         : 'bc'
 
 
+ast            : substr('你好世界', 3)
+raw expr       : substr('你好世界', 3)
+checked expr   : substr<String, Int64>("你好世界", to_int64<UInt8>(3_u8))
+optimized expr : "世界"
+output type    : String
+output domain  : {"世界"..="世界"}
+output         : '世界'
+
+
+ast            : substr('こんにちは', 2)
+raw expr       : substr('こんにちは', 2)
+checked expr   : substr<String, Int64>("こんにちは", to_int64<UInt8>(2_u8))
+optimized expr : "んにちは"
+output type    : String
+output domain  : {"んにちは"..="んにちは"}
+output         : 'んにちは'
+
+
 ast            : substr('abc', pos, len)
 raw expr       : substr('abc', pos::Int8, len::UInt8)
 checked expr   : substr<String, Int64, UInt64>("abc", to_int64<Int8>(pos), to_uint64<UInt8>(len))
@@ -3216,6 +3550,24 @@ output domain  : [{"a"..="c"}]
 output         : ['a', 'c']
 
 
+ast            : split('你好世界', '好')
+raw expr       : split('你好世界', '好')
+checked expr   : split<String, String>("你好世界", "好")
+optimized expr : ['你', '世界']
+output type    : Array(String)
+output domain  : [{"世界"..="你"}]
+output         : ['你', '世界']
+
+
+ast            : split('こんにちは', 'に')
+raw expr       : split('こんにちは', 'に')
+checked expr   : split<String, String>("こんにちは", "に")
+optimized expr : ['こん', 'ちは']
+output type    : Array(String)
+output domain  : [{"こん"..="ちは"}]
+output         : ['こん', 'ちは']
+
+
 ast            : split(str, sep)
 raw expr       : split(str::String NULL, sep::String NULL)
 checked expr   : split<String NULL, String NULL>(str, sep)
diff --git a/src/query/pipeline/core/Cargo.toml b/src/query/pipeline/core/Cargo.toml
index 7e7c96c9a2f3..0984b255a25a 100644
--- a/src/query/pipeline/core/Cargo.toml
+++ b/src/query/pipeline/core/Cargo.toml
@@ -19,6 +19,7 @@ async-backtrace = { workspace = true }
 async-trait = { workspace = true }
 futures = { workspace = true }
 minitrace = { workspace = true }
+once_cell = { workspace = true }
 petgraph = "0.6.2"
 serde = { workspace = true }
 
diff --git a/src/query/pipeline/core/src/lib.rs b/src/query/pipeline/core/src/lib.rs
index 731667028d39..ff9ca5a777e3 100644
--- a/src/query/pipeline/core/src/lib.rs
+++ b/src/query/pipeline/core/src/lib.rs
@@ -12,6 +12,8 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+#![feature(once_cell_try)]
+#![feature(variant_count)]
 #![allow(clippy::arc_with_non_send_sync)]
 
 pub mod processors;
@@ -33,5 +35,7 @@ pub use pipe::SourcePipeBuilder;
 pub use pipe::TransformPipeBuilder;
 pub use pipeline::query_spill_prefix;
 pub use pipeline::Pipeline;
+pub use processors::get_statistics_desc;
 pub use processors::PlanScope;
 pub use processors::PlanScopeGuard;
+pub use processors::ProfileLabel;
diff --git a/src/query/pipeline/core/src/processors/mod.rs b/src/query/pipeline/core/src/processors/mod.rs
index 5c51f4ee2392..fc2a3d899d41 100644
--- a/src/query/pipeline/core/src/processors/mod.rs
+++ b/src/query/pipeline/core/src/processors/mod.rs
@@ -18,6 +18,7 @@ mod processor;
 mod duplicate_processor;
 mod port_trigger;
 pub mod profile;
+mod profiles;
 mod resize_processor;
 mod shuffle_processor;
 
@@ -34,6 +35,11 @@ pub use processor::Processor;
 pub use processor::ProcessorPtr;
 pub use profile::PlanScope;
 pub use profile::PlanScopeGuard;
+pub use profile::Profile;
+pub use profile::ProfileLabel;
+pub use profiles::get_statistics_desc;
+pub use profiles::ProfileDesc;
+pub use profiles::ProfileStatisticsName;
 pub use resize_processor::create_resize_item;
 pub use resize_processor::ResizeProcessor;
 pub use shuffle_processor::ShuffleProcessor;
diff --git a/src/query/pipeline/core/src/processors/port.rs b/src/query/pipeline/core/src/processors/port.rs
index d39e519775fb..2d7c521edf60 100644
--- a/src/query/pipeline/core/src/processors/port.rs
+++ b/src/query/pipeline/core/src/processors/port.rs
@@ -19,6 +19,8 @@ use std::sync::Arc;
 use databend_common_exception::Result;
 use databend_common_expression::DataBlock;
 
+use crate::processors::Profile;
+use crate::processors::ProfileStatisticsName;
 use crate::processors::UpdateTrigger;
 use crate::unsafe_cell_wrap::UnSafeCellWrap;
 
@@ -206,6 +208,7 @@ impl InputPort {
 }
 
 pub struct OutputPort {
+    record_profile: UnSafeCellWrap<bool>,
     shared: UnSafeCellWrap<Arc<SharedStatus>>,
     update_trigger: UnSafeCellWrap<*mut UpdateTrigger>,
 }
@@ -213,6 +216,7 @@ pub struct OutputPort {
 impl OutputPort {
     pub fn create() -> Arc<OutputPort> {
         Arc::new(OutputPort {
+            record_profile: UnSafeCellWrap::create(false),
             shared: UnSafeCellWrap::create(SharedStatus::create()),
             update_trigger: UnSafeCellWrap::create(std::ptr::null_mut()),
         })
@@ -223,6 +227,19 @@ impl OutputPort {
         unsafe {
             UpdateTrigger::update_output(&self.update_trigger);
 
+            if *self.record_profile {
+                if let Ok(data_block) = &data {
+                    Profile::record_usize_profile(
+                        ProfileStatisticsName::OutputRows,
+                        data_block.num_rows(),
+                    );
+                    Profile::record_usize_profile(
+                        ProfileStatisticsName::OutputBytes,
+                        data_block.memory_size(),
+                    );
+                }
+            }
+
             let data = Box::into_raw(Box::new(SharedData(data)));
             self.shared.swap(data, HAS_DATA, HAS_DATA);
         }
@@ -271,6 +288,13 @@ impl OutputPort {
     pub unsafe fn set_trigger(&self, update_trigger: *mut UpdateTrigger) {
         self.update_trigger.set_value(update_trigger)
     }
+
+    /// # Safety
+    ///
+    /// Method is thread unsafe and require thread safe call
+    pub unsafe fn record_profile(&self) {
+        self.record_profile.set_value(true);
+    }
 }
 
 /// Connect input and output ports.
diff --git a/src/query/pipeline/core/src/processors/processor.rs b/src/query/pipeline/core/src/processors/processor.rs
index e9e01d0af9c4..444d500da61a 100644
--- a/src/query/pipeline/core/src/processors/processor.rs
+++ b/src/query/pipeline/core/src/processors/processor.rs
@@ -25,8 +25,6 @@ use minitrace::prelude::*;
 use petgraph::graph::node_index;
 use petgraph::prelude::NodeIndex;
 
-use crate::processors::profile::Profile;
-
 #[derive(Debug)]
 pub enum Event {
     NeedData,
@@ -82,8 +80,6 @@ pub trait Processor: Send {
         Err(ErrorCode::Unimplemented("Unimplemented async_process."))
     }
 
-    fn record_profile(&self, _profile: &Profile) {}
-
     fn details_status(&self) -> Option<String> {
         None
     }
@@ -158,11 +154,6 @@ impl ProcessorPtr {
         (*self.inner.get()).un_reacted(cause, self.id().index())
     }
 
-    /// # Safety
-    pub unsafe fn record_profile(&self, profile: &Profile) {
-        (*self.inner.get()).record_profile(profile)
-    }
-
     /// # Safety
     pub unsafe fn interrupt(&self) {
         (*self.inner.get()).interrupt()
@@ -246,8 +237,4 @@ impl<T: Processor + ?Sized> Processor for Box<T> {
     fn details_status(&self) -> Option<String> {
         (**self).details_status()
     }
-
-    fn record_profile(&self, profile: &Profile) {
-        (**self).record_profile(profile)
-    }
 }
diff --git a/src/query/pipeline/core/src/processors/profile.rs b/src/query/pipeline/core/src/processors/profile.rs
index c68b447c2977..8bc36a33b0fc 100644
--- a/src/query/pipeline/core/src/processors/profile.rs
+++ b/src/query/pipeline/core/src/processors/profile.rs
@@ -12,53 +12,20 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::atomic::AtomicU64;
+use std::cell::RefCell;
 use std::sync::atomic::AtomicUsize;
 use std::sync::atomic::Ordering;
 use std::sync::Arc;
 
-#[derive(Default, Debug)]
-pub struct Profile {
-    /// The id of processor
-    pub pid: usize,
-    /// The name of processor
-    pub p_name: String,
-
-    pub plan_id: Option<u32>,
-    pub plan_name: Option<String>,
-    pub plan_parent_id: Option<u32>,
-
-    /// The time spent to process in nanoseconds
-    pub cpu_time: AtomicU64,
-    /// The time spent to wait in nanoseconds, usually used to
-    /// measure the time spent on waiting for I/O
-    pub wait_time: AtomicU64,
-
-    pub exchange_rows: AtomicUsize,
-    pub exchange_bytes: AtomicUsize,
-}
-
-impl Profile {
-    pub fn create(pid: usize, p_name: String, scope: Option<PlanScope>) -> Profile {
-        Profile {
-            pid,
-            p_name,
-            cpu_time: AtomicU64::new(0),
-            wait_time: AtomicU64::new(0),
-            exchange_rows: AtomicUsize::new(0),
-            exchange_bytes: AtomicUsize::new(0),
-            plan_id: scope.as_ref().map(|x| x.id),
-            plan_name: scope.as_ref().map(|x| x.name.clone()),
-            plan_parent_id: scope.as_ref().and_then(|x| x.parent_id),
-        }
-    }
-}
+use crate::processors::profiles::ProfileStatisticsName;
 
 #[derive(Clone, serde::Serialize, serde::Deserialize)]
 pub struct PlanProfile {
     pub id: Option<u32>,
     pub name: Option<String>,
     pub parent_id: Option<u32>,
+    pub title: Arc<String>,
+    pub labels: Arc<Vec<ProfileLabel>>,
 
     /// The time spent to process in nanoseconds
     pub cpu_time: usize,
@@ -68,6 +35,8 @@ pub struct PlanProfile {
 
     pub exchange_rows: usize,
     pub exchange_bytes: usize,
+
+    pub statistics: [usize; std::mem::variant_count::<ProfileStatisticsName>()],
 }
 
 impl PlanProfile {
@@ -76,21 +45,34 @@ impl PlanProfile {
             id: profile.plan_id,
             name: profile.plan_name.clone(),
             parent_id: profile.plan_parent_id,
-            cpu_time: profile.cpu_time.load(Ordering::SeqCst) as usize,
-            wait_time: profile.wait_time.load(Ordering::SeqCst) as usize,
-            exchange_rows: profile.exchange_rows.load(Ordering::SeqCst),
-            exchange_bytes: profile.exchange_bytes.load(Ordering::SeqCst),
+            title: profile.title.clone(),
+            labels: profile.labels.clone(),
+            cpu_time: profile.load_profile(ProfileStatisticsName::CpuTime),
+            wait_time: profile.load_profile(ProfileStatisticsName::WaitTime),
+            exchange_rows: profile.load_profile(ProfileStatisticsName::ExchangeRows),
+            exchange_bytes: profile.load_profile(ProfileStatisticsName::ExchangeBytes),
+            statistics: std::array::from_fn(|index| {
+                profile.statistics[index].load(Ordering::SeqCst)
+            }),
         }
     }
 
     pub fn accumulate(&mut self, profile: &Profile) {
-        self.cpu_time += profile.cpu_time.load(Ordering::SeqCst) as usize;
-        self.wait_time += profile.wait_time.load(Ordering::SeqCst) as usize;
-        self.exchange_rows += profile.exchange_rows.load(Ordering::SeqCst);
-        self.exchange_bytes += profile.exchange_bytes.load(Ordering::SeqCst);
+        for index in 0..std::mem::variant_count::<ProfileStatisticsName>() {
+            self.statistics[index] += profile.statistics[index].load(Ordering::SeqCst);
+        }
+
+        self.cpu_time += profile.load_profile(ProfileStatisticsName::CpuTime);
+        self.wait_time += profile.load_profile(ProfileStatisticsName::WaitTime);
+        self.exchange_rows += profile.load_profile(ProfileStatisticsName::ExchangeRows);
+        self.exchange_bytes += profile.load_profile(ProfileStatisticsName::ExchangeBytes);
     }
 
     pub fn merge(&mut self, profile: &PlanProfile) {
+        for index in 0..std::mem::variant_count::<ProfileStatisticsName>() {
+            self.statistics[index] += profile.statistics[index];
+        }
+
         self.cpu_time += profile.cpu_time;
         self.wait_time += profile.wait_time;
         self.exchange_rows += profile.exchange_rows;
@@ -124,14 +106,95 @@ pub struct PlanScope {
     pub id: u32,
     pub name: String,
     pub parent_id: Option<u32>,
+    pub title: Arc<String>,
+    pub labels: Arc<Vec<ProfileLabel>>,
 }
 
 impl PlanScope {
-    pub fn create(id: u32, name: String) -> PlanScope {
+    pub fn create(
+        id: u32,
+        name: String,
+        title: Arc<String>,
+        labels: Arc<Vec<ProfileLabel>>,
+    ) -> PlanScope {
         PlanScope {
             id,
+            labels,
+            title,
             parent_id: None,
             name,
         }
     }
 }
+
+#[derive(Clone, Debug, serde::Serialize, serde::Deserialize)]
+pub struct ProfileLabel {
+    pub name: String,
+    pub value: Vec<String>,
+}
+
+impl ProfileLabel {
+    pub fn create(name: String, value: Vec<String>) -> ProfileLabel {
+        ProfileLabel { name, value }
+    }
+}
+
+pub struct Profile {
+    /// The id of processor
+    pub pid: usize,
+    /// The name of processor
+    pub p_name: String,
+
+    pub plan_id: Option<u32>,
+    pub plan_name: Option<String>,
+    pub plan_parent_id: Option<u32>,
+    pub labels: Arc<Vec<ProfileLabel>>,
+    pub title: Arc<String>,
+
+    pub statistics: [AtomicUsize; std::mem::variant_count::<ProfileStatisticsName>()],
+}
+
+thread_local! {
+    static CURRENT_PROFILE: RefCell<Option<Arc<Profile>>> = const { RefCell::new(None) };
+}
+
+impl Profile {
+    fn create_items() -> [AtomicUsize; std::mem::variant_count::<ProfileStatisticsName>()] {
+        std::array::from_fn(|_| AtomicUsize::new(0))
+    }
+    pub fn create(pid: usize, p_name: String, scope: Option<PlanScope>) -> Profile {
+        Profile {
+            pid,
+            p_name,
+            plan_id: scope.as_ref().map(|x| x.id),
+            plan_name: scope.as_ref().map(|x| x.name.clone()),
+            plan_parent_id: scope.as_ref().and_then(|x| x.parent_id),
+            statistics: Self::create_items(),
+            labels: scope
+                .as_ref()
+                .map(|x| x.labels.clone())
+                .unwrap_or(Arc::new(vec![])),
+            title: scope
+                .as_ref()
+                .map(|x| x.title.clone())
+                .unwrap_or(Arc::new(String::new())),
+        }
+    }
+
+    pub fn track_profile(profile: &Arc<Profile>) {
+        CURRENT_PROFILE.set(Some(profile.clone()))
+    }
+
+    pub fn record_usize_profile(name: ProfileStatisticsName, value: usize) {
+        CURRENT_PROFILE.with(|x| match x.borrow().as_ref() {
+            None => {}
+            Some(profile) => {
+                profile.statistics[name as usize].fetch_add(value, Ordering::SeqCst);
+            }
+        });
+    }
+
+    pub fn load_profile(&self, name: ProfileStatisticsName) -> usize {
+        self.statistics[name as usize].load(Ordering::SeqCst)
+    }
+}
diff --git a/src/query/pipeline/core/src/processors/profiles.rs b/src/query/pipeline/core/src/processors/profiles.rs
new file mode 100644
index 000000000000..b9119343417c
--- /dev/null
+++ b/src/query/pipeline/core/src/processors/profiles.rs
@@ -0,0 +1,175 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::collections::HashMap;
+use std::fmt::Display;
+use std::fmt::Formatter;
+use std::sync::Arc;
+
+use once_cell::sync::OnceCell;
+
+#[derive(Clone, Hash, Eq, PartialEq, serde::Serialize, serde::Deserialize, Debug)]
+pub enum ProfileStatisticsName {
+    /// The time spent to process in nanoseconds
+    CpuTime,
+    /// The time spent to wait in nanoseconds, usually used to
+    /// measure the time spent on waiting for I/O
+    WaitTime,
+    ExchangeRows,
+    ExchangeBytes,
+    OutputRows,
+    OutputBytes,
+    ScanBytes,
+    ScanCacheBytes,
+    ScanPartitions,
+    SpillWriteCount,
+    SpillWriteBytes,
+    SpillWriteTime,
+    SpillReadCount,
+    SpillReadBytes,
+    SpillReadTime,
+}
+
+impl Display for ProfileStatisticsName {
+    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
+        write!(f, "{:?}", self)
+    }
+}
+
+impl From<usize> for ProfileStatisticsName {
+    fn from(value: usize) -> Self {
+        let statistics_index = get_statistics_name_index();
+
+        if value > statistics_index.len() {
+            panic!("logical error");
+        }
+
+        match &statistics_index[value] {
+            None => panic!("logical error {}", value),
+            Some(statistics_name) => statistics_name.clone(),
+        }
+    }
+}
+
+#[derive(Clone, serde::Serialize, serde::Deserialize)]
+pub struct ProfileDesc {
+    desc: &'static str,
+    display_name: &'static str,
+    index: usize,
+}
+
+pub static PROFILES_DESC: OnceCell<Arc<HashMap<ProfileStatisticsName, ProfileDesc>>> =
+    OnceCell::new();
+
+pub static PROFILES_INDEX: OnceCell<
+    Arc<[Option<ProfileStatisticsName>; std::mem::variant_count::<ProfileStatisticsName>()]>,
+> = OnceCell::new();
+
+fn get_statistics_name_index()
+-> Arc<[Option<ProfileStatisticsName>; std::mem::variant_count::<ProfileStatisticsName>()]> {
+    PROFILES_INDEX
+        .get_or_init(|| {
+            let statistics_desc = get_statistics_desc();
+            let mut statistics_index = std::array::from_fn(|_v| None);
+
+            for (k, v) in statistics_desc.iter() {
+                statistics_index[v.index] = Some(k.clone());
+            }
+
+            Arc::new(statistics_index)
+        })
+        .clone()
+}
+
+pub fn get_statistics_desc() -> Arc<HashMap<ProfileStatisticsName, ProfileDesc>> {
+    PROFILES_DESC.get_or_init(|| {
+        Arc::new(HashMap::from([
+            (ProfileStatisticsName::CpuTime, ProfileDesc {
+                display_name: "cpu time",
+                desc: "The time spent to process in nanoseconds",
+                index: ProfileStatisticsName::CpuTime as usize,
+            }),
+            (ProfileStatisticsName::WaitTime, ProfileDesc {
+                display_name: "wait time",
+                desc: "The time spent to wait in nanoseconds, usually used to measure the time spent on waiting for I/O",
+                index: ProfileStatisticsName::WaitTime as usize,
+            }),
+            (ProfileStatisticsName::ExchangeRows, ProfileDesc {
+                display_name: "exchange rows",
+                desc: "The number of data rows exchange between nodes in cluster mode",
+                index: ProfileStatisticsName::ExchangeRows as usize,
+            }),
+            (ProfileStatisticsName::ExchangeBytes, ProfileDesc {
+                display_name: "exchange bytes",
+                desc: "The number of data bytes exchange between nodes in cluster mode",
+                index: ProfileStatisticsName::ExchangeBytes as usize,
+            }),
+            (ProfileStatisticsName::OutputRows, ProfileDesc {
+                display_name: "output rows",
+                desc: "The number of rows from the physical plan output to the next physical plan",
+                index: ProfileStatisticsName::OutputRows as usize,
+            }),
+            (ProfileStatisticsName::OutputBytes, ProfileDesc {
+                display_name: "output bytes",
+                desc: "The number of bytes from the physical plan output to the next physical plan",
+                index: ProfileStatisticsName::OutputBytes as usize,
+            }),
+            (ProfileStatisticsName::ScanBytes, ProfileDesc {
+                display_name: "bytes scanned",
+                desc: "The bytes scanned of query",
+                index: ProfileStatisticsName::ScanBytes as usize,
+            }),
+            (ProfileStatisticsName::ScanCacheBytes, ProfileDesc {
+                display_name: "bytes scanned from cache",
+                desc: "The bytes scanned from cache of query",
+                index: ProfileStatisticsName::ScanCacheBytes as usize,
+            }),
+            (ProfileStatisticsName::ScanPartitions, ProfileDesc {
+                display_name: "partitions scanned",
+                desc: "The partitions scanned of query",
+                index: ProfileStatisticsName::ScanPartitions as usize,
+            }),
+            (ProfileStatisticsName::SpillWriteCount, ProfileDesc {
+                display_name: "numbers spilled by write",
+                desc: "The number of spilled by write",
+                index: ProfileStatisticsName::SpillWriteCount as usize,
+            }),
+            (ProfileStatisticsName::SpillWriteBytes, ProfileDesc {
+                display_name: "bytes spilled by write",
+                desc: "The bytes spilled by write",
+                index: ProfileStatisticsName::SpillWriteBytes as usize,
+            }),
+            (ProfileStatisticsName::SpillWriteTime, ProfileDesc {
+                display_name: "spilled time by write",
+                desc: "The time spent to write spill in millisecond",
+                index: ProfileStatisticsName::SpillWriteTime as usize,
+            }),
+            (ProfileStatisticsName::SpillReadCount, ProfileDesc {
+                display_name: "numbers spilled by read",
+                desc: "The number of spilled by read",
+                index: ProfileStatisticsName::SpillReadCount as usize,
+            }),
+            (ProfileStatisticsName::SpillReadBytes, ProfileDesc {
+                display_name: "bytes spilled by read",
+                desc: "The bytes spilled by read",
+                index: ProfileStatisticsName::SpillReadBytes as usize,
+            }),
+            (ProfileStatisticsName::SpillReadTime, ProfileDesc {
+                display_name: "spilled time by read",
+                desc: "The time spent to read spill in millisecond",
+                index: ProfileStatisticsName::SpillReadTime as usize,
+            }),
+        ]))
+    }).clone()
+}
diff --git a/src/query/pipeline/sinks/src/async_sink.rs b/src/query/pipeline/sinks/src/async_sink.rs
index be7ed120acea..47c170cb9ffb 100644
--- a/src/query/pipeline/sinks/src/async_sink.rs
+++ b/src/query/pipeline/sinks/src/async_sink.rs
@@ -22,7 +22,6 @@ use databend_common_base::runtime::TrySpawn;
 use databend_common_catalog::table_context::TableContext;
 use databend_common_exception::Result;
 use databend_common_expression::DataBlock;
-use databend_common_pipeline_core::processors::profile::Profile;
 use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::Processor;
@@ -47,8 +46,6 @@ pub trait AsyncSink: Send {
     fn details_status(&self) -> Option<String> {
         None
     }
-
-    fn record_profile(&self, _profile: &Profile) {}
 }
 
 pub struct AsyncSinker<T: AsyncSink + 'static> {
@@ -173,8 +170,4 @@ impl<T: AsyncSink + 'static> Processor for AsyncSinker<T> {
     fn details_status(&self) -> Option<String> {
         self.inner.as_ref().and_then(|x| x.details_status())
     }
-
-    fn record_profile(&self, profile: &Profile) {
-        self.inner.as_ref().unwrap().record_profile(profile);
-    }
 }
diff --git a/src/query/pipeline/sources/src/async_source.rs b/src/query/pipeline/sources/src/async_source.rs
index ca1b651a0c2e..f3477f96219c 100644
--- a/src/query/pipeline/sources/src/async_source.rs
+++ b/src/query/pipeline/sources/src/async_source.rs
@@ -25,6 +25,8 @@ use databend_common_pipeline_core::processors::EventCause;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 
 #[async_trait::async_trait]
 pub trait AsyncSource: Send {
@@ -120,6 +122,10 @@ impl<T: 'static + AsyncSource> Processor for AsyncSourcer<T> {
                         bytes: data_block.memory_size(),
                     };
                     self.scan_progress.incr(&progress_values);
+                    Profile::record_usize_profile(
+                        ProfileStatisticsName::ScanBytes,
+                        data_block.memory_size(),
+                    );
                 }
 
                 if !T::SKIP_EMPTY_DATA_BLOCK || !data_block.is_empty() {
diff --git a/src/query/pipeline/sources/src/input_formats/source_aligner.rs b/src/query/pipeline/sources/src/input_formats/source_aligner.rs
index a9c3fe92461b..6c23694c08df 100644
--- a/src/query/pipeline/sources/src/input_formats/source_aligner.rs
+++ b/src/query/pipeline/sources/src/input_formats/source_aligner.rs
@@ -26,6 +26,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use log::debug;
 
 use crate::input_formats::input_pipeline::AligningStateTrait;
@@ -131,6 +133,10 @@ impl<I: InputFormatPipe> Processor for Aligner<I> {
                     self.state = None;
                     self.batch_rx = None;
                 }
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::ScanBytes,
+                    process_values.bytes,
+                );
                 self.ctx.scan_progress.incr(&process_values);
                 Ok(())
             }
diff --git a/src/query/pipeline/sources/src/sync_source.rs b/src/query/pipeline/sources/src/sync_source.rs
index c35907efff3d..3f3bcdc2348c 100644
--- a/src/query/pipeline/sources/src/sync_source.rs
+++ b/src/query/pipeline/sources/src/sync_source.rs
@@ -24,6 +24,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 
 /// Synchronized source. such as:
 ///     - Memory storage engine.
@@ -106,6 +108,10 @@ impl<T: 'static + SyncSource> Processor for SyncSourcer<T> {
                     bytes: data_block.memory_size(),
                 };
                 self.scan_progress.incr(&progress_values);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::ScanBytes,
+                    data_block.memory_size(),
+                );
                 self.generated_data = Some(data_block)
             }
         };
diff --git a/src/query/pipeline/transforms/Cargo.toml b/src/query/pipeline/transforms/Cargo.toml
index 3fff363caa10..9637be9611d8 100644
--- a/src/query/pipeline/transforms/Cargo.toml
+++ b/src/query/pipeline/transforms/Cargo.toml
@@ -13,7 +13,6 @@ databend-common-base = { path = "../../../common/base" }
 databend-common-exception = { path = "../../../common/exception" }
 databend-common-expression = { path = "../../expression" }
 databend-common-pipeline-core = { path = "../core" }
-databend-common-profile = { path = "../../profile" }
 
 async-backtrace = { workspace = true }
 async-trait = { workspace = true }
diff --git a/src/query/pipeline/transforms/src/processors/mod.rs b/src/query/pipeline/transforms/src/processors/mod.rs
index 0ffd8d9a2580..a9c76cf3f0a3 100644
--- a/src/query/pipeline/transforms/src/processors/mod.rs
+++ b/src/query/pipeline/transforms/src/processors/mod.rs
@@ -12,8 +12,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-mod profile_wrapper;
 mod transforms;
 
-pub use profile_wrapper::*;
 pub use transforms::*;
diff --git a/src/query/pipeline/transforms/src/processors/profile_wrapper.rs b/src/query/pipeline/transforms/src/processors/profile_wrapper.rs
deleted file mode 100644
index 51cc4388ab27..000000000000
--- a/src/query/pipeline/transforms/src/processors/profile_wrapper.rs
+++ /dev/null
@@ -1,280 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::sync::Arc;
-use std::time::Instant;
-
-use databend_common_exception::Result;
-use databend_common_expression::DataBlock;
-use databend_common_pipeline_core::processors::Event;
-use databend_common_pipeline_core::processors::InputPort;
-use databend_common_pipeline_core::processors::OutputPort;
-use databend_common_pipeline_core::processors::Processor;
-use databend_common_profile::ProcessorProfile;
-use databend_common_profile::SharedProcessorProfiles;
-
-use crate::processors::transforms::Transform;
-use crate::processors::transforms::Transformer;
-
-/// A profile wrapper for `Processor` trait.
-/// This wrapper will record the time cost of each processor.
-/// But because of the limitation of `Processor` trait,
-/// we can't get the number of rows processed by the processor.
-pub struct ProcessorProfileWrapper<T> {
-    inner: T,
-    prof_id: u32,
-    proc_profs: SharedProcessorProfiles,
-
-    prof: ProcessorProfile,
-}
-
-impl<T> ProcessorProfileWrapper<T>
-where T: Processor + 'static
-{
-    pub fn create(
-        inner: T,
-        prof_id: u32,
-        proc_profs: SharedProcessorProfiles,
-    ) -> Box<dyn Processor> {
-        Box::new(Self {
-            inner,
-            prof_id,
-            proc_profs,
-            prof: ProcessorProfile::default(),
-        })
-    }
-}
-
-#[async_trait::async_trait]
-impl<T> Processor for ProcessorProfileWrapper<T>
-where T: Processor + 'static
-{
-    fn name(&self) -> String {
-        self.inner.name()
-    }
-
-    fn as_any(&mut self) -> &mut dyn std::any::Any {
-        self
-    }
-
-    fn event(&mut self) -> Result<Event> {
-        match self.inner.event()? {
-            Event::Finished => {
-                self.proc_profs
-                    .lock()
-                    .unwrap()
-                    .update(self.prof_id, self.prof);
-                Ok(Event::Finished)
-            }
-            v => Ok(v),
-        }
-    }
-
-    fn process(&mut self) -> Result<()> {
-        let instant = Instant::now();
-        self.inner.process()?;
-        let elapsed = instant.elapsed();
-        self.prof = self.prof
-            + ProcessorProfile {
-                cpu_time: elapsed,
-                ..Default::default()
-            };
-        Ok(())
-    }
-
-    #[async_backtrace::framed]
-    async fn async_process(&mut self) -> Result<()> {
-        let instant = Instant::now();
-        self.inner.async_process().await?;
-        let elapsed = instant.elapsed();
-        self.prof = self.prof
-            + ProcessorProfile {
-                wait_time: elapsed,
-                ..Default::default()
-            };
-        Ok(())
-    }
-}
-
-/// A profile wrapper for `Transform` trait.
-/// This wrapper will record the time cost and the information
-/// about the number of rows processed by the processor.
-pub struct TransformProfileWrapper<T> {
-    inner: T,
-    prof_id: u32,
-    proc_profs: SharedProcessorProfiles,
-
-    prof: ProcessorProfile,
-}
-
-impl<T> TransformProfileWrapper<T>
-where T: Transform + 'static
-{
-    pub fn create(
-        inner: T,
-        input_port: Arc<InputPort>,
-        output_port: Arc<OutputPort>,
-        prof_id: u32,
-        proc_profs: SharedProcessorProfiles,
-    ) -> Box<dyn Processor> {
-        Box::new(Transformer::create(input_port, output_port, Self {
-            inner,
-            prof_id,
-            proc_profs,
-            prof: ProcessorProfile::default(),
-        }))
-    }
-}
-
-impl<T> Transform for TransformProfileWrapper<T>
-where T: Transform + 'static
-{
-    const NAME: &'static str = "TransformProfileWrapper";
-
-    fn transform(&mut self, data: DataBlock) -> Result<DataBlock> {
-        let input_rows = data.num_rows();
-        let input_bytes = data.memory_size();
-
-        let instant = Instant::now();
-        let res = self.inner.transform(data)?;
-        let elapsed = instant.elapsed();
-        self.prof = self.prof
-            + ProcessorProfile {
-                cpu_time: elapsed,
-                wait_time: Default::default(),
-                input_rows,
-                input_bytes,
-                output_rows: res.num_rows(),
-                output_bytes: res.memory_size(),
-            };
-        Ok(res)
-    }
-
-    fn on_finish(&mut self) -> Result<()> {
-        self.proc_profs
-            .lock()
-            .unwrap()
-            .update(self.prof_id, self.prof);
-        Ok(())
-    }
-}
-
-/// A stub transform for collecting profile information
-/// at some point of the pipeline.
-/// For example, we can profiling the output data of a
-/// processor by adding a [`ProfileStub`] after it.
-/// [`ProfileStub`] will pass through all the data without
-/// any modification.
-#[allow(clippy::type_complexity)]
-pub struct ProfileStub {
-    prof_id: u32,
-    proc_profs: SharedProcessorProfiles,
-    prof: ProcessorProfile,
-
-    /// Callback function for processing the start event.
-    on_start: Box<dyn Fn(&ProcessorProfile) -> ProcessorProfile + Send + Sync + 'static>,
-    /// Callback function for processing the input data.
-    on_process:
-        Box<dyn Fn(&DataBlock, &ProcessorProfile) -> ProcessorProfile + Send + Sync + 'static>,
-    /// Callback function for processing the finish event.
-    on_finish: Box<dyn Fn(&ProcessorProfile) -> ProcessorProfile + Send + Sync + 'static>,
-}
-
-impl ProfileStub {
-    pub fn new(prof_id: u32, proc_profs: SharedProcessorProfiles) -> Self {
-        Self {
-            prof_id,
-            proc_profs,
-            prof: Default::default(),
-            on_start: Box::new(|_| Default::default()),
-            on_process: Box::new(|_, _| Default::default()),
-            on_finish: Box::new(|_| Default::default()),
-        }
-    }
-
-    /// Create a new [`ProfileStub`] with `on_start` callback.
-    /// The previous callback will be called before the new one.
-    pub fn on_start(
-        self,
-        f: impl Fn(&ProcessorProfile) -> ProcessorProfile + Sync + Send + 'static,
-    ) -> Self {
-        Self {
-            on_start: Box::new(move |prof| f(&(self.on_start)(prof))),
-            ..self
-        }
-    }
-
-    /// Create a new [`ProfileStub`] with `on_process` callback.
-    /// The previous callback will be called before the new one.
-    pub fn on_process(
-        self,
-        f: impl Fn(&DataBlock, &ProcessorProfile) -> ProcessorProfile + Sync + Send + 'static,
-    ) -> Self {
-        Self {
-            on_process: Box::new(move |data, prof| f(data, &(self.on_process)(data, prof))),
-            ..self
-        }
-    }
-
-    /// Create a new [`ProfileStub`] with `on_finish` callback.
-    /// The previous callback will be called before the new one.
-    pub fn on_finish(
-        self,
-        f: impl Fn(&ProcessorProfile) -> ProcessorProfile + Sync + Send + 'static,
-    ) -> Self {
-        Self {
-            on_finish: Box::new(move |prof| f(&(self.on_finish)(prof))),
-            ..self
-        }
-    }
-
-    /// Accumulate the number of output rows.
-    pub fn accumulate_output_rows(self) -> Self {
-        self.on_process(|data, prof| ProcessorProfile {
-            output_rows: prof.output_rows + data.num_rows(),
-            ..*prof
-        })
-    }
-
-    /// Accumulate the number of output bytes.
-    pub fn accumulate_output_bytes(self) -> Self {
-        self.on_process(|data, prof| ProcessorProfile {
-            output_bytes: prof.output_bytes + data.memory_size(),
-            ..*prof
-        })
-    }
-}
-
-impl Transform for ProfileStub {
-    const NAME: &'static str = "ProfileStub";
-
-    fn transform(&mut self, data: DataBlock) -> Result<DataBlock> {
-        self.prof = self.prof + (self.on_process)(&data, &self.prof);
-        Ok(data)
-    }
-
-    fn on_start(&mut self) -> Result<()> {
-        self.prof = self.prof + (self.on_start)(&self.prof);
-        Ok(())
-    }
-
-    fn on_finish(&mut self) -> Result<()> {
-        self.prof = self.prof + (self.on_finish)(&self.prof);
-        self.proc_profs
-            .lock()
-            .unwrap()
-            .update(self.prof_id, self.prof);
-        Ok(())
-    }
-}
diff --git a/src/query/pipeline/transforms/src/processors/transforms/transform.rs b/src/query/pipeline/transforms/src/processors/transforms/transform.rs
index 91c8cde21e00..ae6ec7a0b785 100644
--- a/src/query/pipeline/transforms/src/processors/transforms/transform.rs
+++ b/src/query/pipeline/transforms/src/processors/transforms/transform.rs
@@ -21,7 +21,6 @@ use databend_common_exception::Result;
 use databend_common_expression::BlockMetaInfo;
 use databend_common_expression::BlockMetaInfoDowncast;
 use databend_common_expression::DataBlock;
-use databend_common_pipeline_core::processors::profile::Profile;
 use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
@@ -45,8 +44,6 @@ pub trait Transform: Send {
     fn on_finish(&mut self) -> Result<()> {
         Ok(())
     }
-
-    fn record_profile(&self, _: &Profile) {}
 }
 
 pub struct Transformer<T: Transform + 'static> {
@@ -125,10 +122,6 @@ impl<T: Transform + 'static> Processor for Transformer<T> {
 
         Ok(())
     }
-
-    fn record_profile(&self, profile: &Profile) {
-        self.transform.record_profile(profile)
-    }
 }
 
 impl<T: Transform> Transformer<T> {
diff --git a/src/query/pipeline/transforms/src/processors/transforms/transform_multi_sort_merge.rs b/src/query/pipeline/transforms/src/processors/transforms/transform_multi_sort_merge.rs
index 51f995edca21..8b227df8c583 100644
--- a/src/query/pipeline/transforms/src/processors/transforms/transform_multi_sort_merge.rs
+++ b/src/query/pipeline/transforms/src/processors/transforms/transform_multi_sort_merge.rs
@@ -38,14 +38,12 @@ use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_core::Pipe;
 use databend_common_pipeline_core::PipeItem;
 use databend_common_pipeline_core::Pipeline;
-use databend_common_profile::SharedProcessorProfiles;
 
 use super::sort::HeapMerger;
 use super::sort::Rows;
 use super::sort::SimpleRows;
 use super::sort::SortedStream;
 use crate::processors::sort::utils::ORDER_COL_NAME;
-use crate::processors::ProcessorProfileWrapper;
 
 pub fn try_add_multi_sort_merge(
     pipeline: &mut Pipeline,
@@ -53,7 +51,6 @@ pub fn try_add_multi_sort_merge(
     block_size: usize,
     limit: Option<usize>,
     sort_columns_descriptions: Arc<Vec<SortColumnDescription>>,
-    prof_info: Option<(u32, SharedProcessorProfiles)>,
     remove_order_col: bool,
 ) -> Result<()> {
     debug_assert!(if !remove_order_col {
@@ -75,7 +72,8 @@ pub fn try_add_multi_sort_merge(
                 inputs_port.push(InputPort::create());
             }
             let output_port = OutputPort::create();
-            let processor = create_processor(
+
+            let processor = ProcessorPtr::create(create_processor(
                 inputs_port.clone(),
                 output_port.clone(),
                 schema,
@@ -83,17 +81,7 @@ pub fn try_add_multi_sort_merge(
                 limit,
                 sort_columns_descriptions,
                 remove_order_col,
-            )?;
-
-            let processor = if let Some((plan_id, prof)) = &prof_info {
-                ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    processor,
-                    *plan_id,
-                    prof.clone(),
-                ))
-            } else {
-                ProcessorPtr::create(processor)
-            };
+            )?);
 
             pipeline.add_pipe(Pipe::create(inputs_port.len(), 1, vec![PipeItem::create(
                 processor,
diff --git a/src/query/profile/Cargo.toml b/src/query/profile/Cargo.toml
deleted file mode 100644
index ed81eeee1078..000000000000
--- a/src/query/profile/Cargo.toml
+++ /dev/null
@@ -1,16 +0,0 @@
-[package]
-name = "databend-common-profile"
-version = { workspace = true }
-authors = { workspace = true }
-license = { workspace = true }
-publish = { workspace = true }
-edition = { workspace = true }
-
-[lib]
-doctest = false
-test = false
-
-[dependencies]
-databend-common-base = { path = "../../common/base" }
-
-dashmap = { workspace = true }
diff --git a/src/query/profile/src/mgr.rs b/src/query/profile/src/mgr.rs
deleted file mode 100644
index 9f3e29f02af7..000000000000
--- a/src/query/profile/src/mgr.rs
+++ /dev/null
@@ -1,134 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::collections::VecDeque;
-use std::sync::Arc;
-use std::sync::Mutex;
-
-use dashmap::mapref::entry::Entry;
-use dashmap::DashMap;
-use databend_common_base::base::GlobalInstance;
-
-use crate::prof::QueryProfile;
-
-/// Default capacity of the LRU cache of query profiles.
-const DEFAULT_QUERY_PROFILE_LIMIT: usize = 20;
-
-/// Manager of query profiling.
-/// This is a singleton in every databend-query process.
-pub struct QueryProfileManager {
-    /// The LRU cache of query profiles.
-    profiles: Lru,
-}
-
-impl QueryProfileManager {
-    fn new(capacity: usize) -> Self {
-        QueryProfileManager {
-            profiles: Lru::new(capacity),
-        }
-    }
-
-    pub fn init() {
-        GlobalInstance::set(Arc::new(Self::new(DEFAULT_QUERY_PROFILE_LIMIT)));
-    }
-
-    pub fn instance() -> Arc<Self> {
-        GlobalInstance::get()
-    }
-
-    /// Try to get the query profile by query ID.
-    pub fn get(&self, query_id: &str) -> Option<Arc<QueryProfile>> {
-        self.profiles.get(query_id)
-    }
-
-    /// Inserts a query profile.
-    pub fn insert(&self, query_profile: Arc<QueryProfile>) {
-        self.profiles
-            .insert(query_profile.query_id.clone(), query_profile);
-    }
-
-    /// Lists all query profiles.
-    pub fn list_all(&self) -> Vec<Arc<QueryProfile>> {
-        self.profiles.list_all()
-    }
-}
-
-impl Default for QueryProfileManager {
-    fn default() -> Self {
-        QueryProfileManager::new(DEFAULT_QUERY_PROFILE_LIMIT)
-    }
-}
-
-/// An LRU cache of query profiles. The expired query profiles
-/// will be removed.
-struct Lru {
-    /// The maximum number of query profiles to keep in memory.
-    /// If the number of query profiles exceeds this number,
-    /// the oldest one will be removed.
-    capacity: usize,
-
-    /// The query profiles.
-    /// The key is the query ID.
-    /// The value is the query profile.
-    profiles: DashMap<String, Arc<QueryProfile>>,
-
-    /// An LRU list of query IDs.
-    lru: Mutex<VecDeque<String>>,
-}
-
-impl Lru {
-    /// Creates a new LRU cache.
-    pub fn new(capacity: usize) -> Self {
-        Lru {
-            capacity,
-            profiles: DashMap::with_capacity(capacity),
-            lru: Mutex::new(VecDeque::with_capacity(capacity)),
-        }
-    }
-
-    /// Gets the query profile by the query ID.
-    /// Notice that this method required to acquire the shared lock of the LRU list.
-    /// So don't call this method when the lock is already acquired.
-    pub fn get(&self, query_id: &str) -> Option<Arc<QueryProfile>> {
-        self.profiles.get(query_id).map(|v| v.value().clone())
-    }
-
-    /// Inserts a query profile.
-    /// This operation is thread-safe.
-    pub fn insert(&self, query_id: String, query_profile: Arc<QueryProfile>) {
-        // Lock the LRU list to ensure the consistency between the LRU list and the query profiles.
-        let mut lru = self.lru.lock().unwrap();
-
-        if let Entry::Occupied(mut prof) = self.profiles.entry(query_id.clone()) {
-            prof.insert(query_profile);
-            return;
-        }
-
-        if self.profiles.len() >= self.capacity {
-            if let Some(query_id) = lru.pop_front() {
-                self.profiles.remove(&query_id);
-            }
-        }
-
-        self.profiles.insert(query_id.clone(), query_profile);
-        lru.push_back(query_id);
-    }
-
-    /// Lists all query profiles.
-    /// Notice that this method required to acquire the shared lock of the LRU list.
-    /// So don't call this method when the lock is already acquired.
-    pub fn list_all(&self) -> Vec<Arc<QueryProfile>> {
-        self.profiles.iter().map(|v| v.value().clone()).collect()
-    }
-}
diff --git a/src/query/profile/src/proc.rs b/src/query/profile/src/proc.rs
deleted file mode 100644
index bd91dc88fabf..000000000000
--- a/src/query/profile/src/proc.rs
+++ /dev/null
@@ -1,85 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::collections::HashMap;
-use std::fmt::Debug;
-use std::sync::Arc;
-use std::sync::Mutex;
-use std::time::Duration;
-
-pub type SharedProcessorProfiles = Arc<Mutex<ProcessorProfiles<u32>>>;
-
-/// Execution profile information of a `Processor`.
-/// Can be merged with other `ProcessorProfile` using
-/// `add` or `+` operator.
-///
-/// # Example
-/// ```
-/// let profile1 = ProcessorProfile::default();
-/// let profile2 = ProcessorProfile::default();
-/// let profile = profile1 + profile2;
-/// ```
-#[derive(Default, Clone, Copy, Debug)]
-pub struct ProcessorProfile {
-    /// The time spent to process in nanoseconds
-    pub cpu_time: Duration,
-    /// The time spent to wait in nanoseconds, usually used to
-    /// measure the time spent on waiting for I/O
-    pub wait_time: Duration,
-    /// Row count of the input data
-    pub input_rows: usize,
-    /// Byte size of the input data
-    pub input_bytes: usize,
-    /// Row count of the output data
-    pub output_rows: usize,
-    /// Byte size of the output data
-    pub output_bytes: usize,
-}
-
-impl std::ops::Add for ProcessorProfile {
-    type Output = Self;
-
-    fn add(self, rhs: Self) -> Self::Output {
-        Self {
-            cpu_time: self.cpu_time + rhs.cpu_time,
-            wait_time: self.wait_time + rhs.wait_time,
-            input_rows: self.input_rows + rhs.input_rows,
-            input_bytes: self.input_bytes + rhs.input_bytes,
-            output_rows: self.output_rows + rhs.output_rows,
-            output_bytes: self.output_bytes + rhs.output_bytes,
-        }
-    }
-}
-
-#[derive(Default)]
-pub struct ProcessorProfiles<K = u32> {
-    spans: HashMap<K, ProcessorProfile>,
-}
-
-impl<K> ProcessorProfiles<K>
-where K: std::hash::Hash + Eq + PartialEq + Clone + Debug
-{
-    pub fn update(&mut self, key: K, span: ProcessorProfile) {
-        let entry = self.spans.entry(key).or_default();
-        *entry = *entry + span;
-    }
-
-    pub fn iter(&self) -> impl Iterator<Item = (&K, &ProcessorProfile)> {
-        self.spans.iter()
-    }
-
-    pub fn get(&self, k: &K) -> Option<&ProcessorProfile> {
-        self.spans.get(k)
-    }
-}
diff --git a/src/query/profile/src/prof.rs b/src/query/profile/src/prof.rs
deleted file mode 100644
index b42e7852076f..000000000000
--- a/src/query/profile/src/prof.rs
+++ /dev/null
@@ -1,217 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::fmt::Display;
-use std::fmt::Formatter;
-use std::time::Duration;
-
-use crate::ProcessorProfile;
-
-#[derive(Debug, Clone)]
-pub struct QueryProfile {
-    /// Query ID of the query profile
-    pub query_id: String,
-
-    /// Flattened plan node profiles
-    pub operator_profiles: Vec<OperatorProfile>,
-}
-
-impl QueryProfile {
-    pub fn new(query_id: String, operator_profiles: Vec<OperatorProfile>) -> Self {
-        QueryProfile {
-            query_id,
-            operator_profiles,
-        }
-    }
-}
-
-#[derive(Debug, Clone)]
-pub struct OperatorProfile {
-    /// ID of the plan node
-    pub id: u32,
-
-    /// Type of the plan operator, e.g. `HashJoin`
-    pub operator_type: OperatorType,
-
-    /// IDs of the children plan nodes
-    pub children: Vec<u32>,
-
-    /// The execution information of the plan operator
-    pub execution_info: OperatorExecutionInfo,
-
-    /// Attribute of the plan operator
-    pub attribute: OperatorAttribute,
-}
-
-#[derive(Debug, Clone)]
-pub enum OperatorType {
-    Join,
-    Aggregate,
-    AggregateExpand,
-    Filter,
-    ProjectSet,
-    EvalScalar,
-    Limit,
-    TableScan,
-    CteScan,
-    Sort,
-    UnionAll,
-    Project,
-    Window,
-    RowFetch,
-    Exchange,
-    Insert,
-    ConstantTableScan,
-    Udf,
-}
-
-impl Display for OperatorType {
-    fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result {
-        match self {
-            OperatorType::Join => write!(f, "Join"),
-            OperatorType::Aggregate => write!(f, "Aggregate"),
-            OperatorType::AggregateExpand => write!(f, "AggregateExpand"),
-            OperatorType::Filter => write!(f, "Filter"),
-            OperatorType::ProjectSet => write!(f, "ProjectSet"),
-            OperatorType::EvalScalar => write!(f, "EvalScalar"),
-            OperatorType::Limit => write!(f, "Limit"),
-            OperatorType::TableScan => write!(f, "TableScan"),
-            OperatorType::Sort => write!(f, "Sort"),
-            OperatorType::UnionAll => write!(f, "UnionAll"),
-            OperatorType::Project => write!(f, "Project"),
-            OperatorType::Window => write!(f, "Window"),
-            OperatorType::RowFetch => write!(f, "RowFetch"),
-            OperatorType::Exchange => write!(f, "Exchange"),
-            OperatorType::Insert => write!(f, "Insert"),
-            OperatorType::CteScan => write!(f, "CteScan"),
-            OperatorType::ConstantTableScan => write!(f, "ConstantTableScan"),
-            OperatorType::Udf => write!(f, "Udf"),
-        }
-    }
-}
-
-#[derive(Debug, Clone, Default)]
-pub struct OperatorExecutionInfo {
-    pub process_time: Duration,
-    pub input_rows: usize,
-    pub input_bytes: usize,
-    pub output_rows: usize,
-    pub output_bytes: usize,
-}
-
-impl From<ProcessorProfile> for OperatorExecutionInfo {
-    fn from(value: ProcessorProfile) -> Self {
-        (&value).into()
-    }
-}
-
-impl From<&ProcessorProfile> for OperatorExecutionInfo {
-    fn from(value: &ProcessorProfile) -> Self {
-        OperatorExecutionInfo {
-            process_time: value.cpu_time,
-            input_rows: value.input_rows,
-            input_bytes: value.input_bytes,
-            output_rows: value.output_rows,
-            output_bytes: value.output_bytes,
-        }
-    }
-}
-
-#[derive(Debug, Clone)]
-pub enum OperatorAttribute {
-    Join(JoinAttribute),
-    Aggregate(AggregateAttribute),
-    AggregateExpand(AggregateExpandAttribute),
-    Filter(FilterAttribute),
-    EvalScalar(EvalScalarAttribute),
-    ProjectSet(ProjectSetAttribute),
-    Limit(LimitAttribute),
-    TableScan(TableScanAttribute),
-    Sort(SortAttribute),
-    Window(WindowAttribute),
-    Exchange(ExchangeAttribute),
-    CteScan(CteScanAttribute),
-    Udf(UdfAttribute),
-    Empty,
-}
-
-#[derive(Debug, Clone)]
-pub struct JoinAttribute {
-    pub join_type: String,
-    pub equi_conditions: String,
-    pub non_equi_conditions: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct AggregateAttribute {
-    pub group_keys: String,
-    pub functions: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct AggregateExpandAttribute {
-    pub group_keys: String,
-    pub aggr_exprs: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct EvalScalarAttribute {
-    pub scalars: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct ProjectSetAttribute {
-    pub functions: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct FilterAttribute {
-    pub predicate: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct LimitAttribute {
-    pub limit: usize,
-    pub offset: usize,
-}
-
-#[derive(Debug, Clone)]
-pub struct SortAttribute {
-    pub sort_keys: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct TableScanAttribute {
-    pub qualified_name: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct CteScanAttribute {
-    pub cte_idx: usize,
-}
-
-#[derive(Debug, Clone)]
-pub struct WindowAttribute {
-    pub functions: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct ExchangeAttribute {
-    pub exchange_mode: String,
-}
-
-#[derive(Debug, Clone)]
-pub struct UdfAttribute {
-    pub scalars: String,
-}
diff --git a/src/query/service/Cargo.toml b/src/query/service/Cargo.toml
index f2d59c499aa2..8b4e0cadd97e 100644
--- a/src/query/service/Cargo.toml
+++ b/src/query/service/Cargo.toml
@@ -61,7 +61,6 @@ databend-common-pipeline-core = { path = "../pipeline/core" }
 databend-common-pipeline-sinks = { path = "../pipeline/sinks" }
 databend-common-pipeline-sources = { path = "../pipeline/sources" }
 databend-common-pipeline-transforms = { path = "../pipeline/transforms" }
-databend-common-profile = { path = "../profile" }
 databend-common-settings = { path = "../settings" }
 databend-common-sharing = { path = "../sharing" }
 databend-common-sql = { path = "../sql" }
diff --git a/src/query/service/src/api/rpc/exchange/exchange_manager.rs b/src/query/service/src/api/rpc/exchange/exchange_manager.rs
index 2a9512bd26ce..f6e7ca1e2a3e 100644
--- a/src/query/service/src/api/rpc/exchange/exchange_manager.rs
+++ b/src/query/service/src/api/rpc/exchange/exchange_manager.rs
@@ -31,7 +31,6 @@ use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_grpc::ConnectionFactory;
 use databend_common_pipeline_core::processors::profile::Profile;
-use databend_common_profile::SharedProcessorProfiles;
 use databend_common_sql::executor::PhysicalPlan;
 use minitrace::prelude::*;
 use parking_lot::Mutex;
@@ -231,9 +230,7 @@ impl DataExchangeManager {
                 "Query {} not found in cluster.",
                 packet.query_id
             ))),
-            Some(query_coordinator) => {
-                query_coordinator.prepare_pipeline(ctx, packet.enable_profiling, packet)
-            }
+            Some(query_coordinator) => query_coordinator.prepare_pipeline(ctx, packet),
         }
     }
 
@@ -300,7 +297,6 @@ impl DataExchangeManager {
     pub async fn commit_actions(
         &self,
         ctx: Arc<QueryContext>,
-        enable_profiling: bool,
         actions: QueryFragmentsActions,
     ) -> Result<PipelineBuildResult> {
         let settings = ctx.get_settings();
@@ -327,7 +323,7 @@ impl DataExchangeManager {
         self.init_query_fragments_plan(&ctx, &local_query_fragments_plan_packet)?;
 
         // Get local pipeline of local task
-        let build_res = self.get_root_pipeline(ctx, enable_profiling, root_actions)?;
+        let build_res = self.get_root_pipeline(ctx, root_actions)?;
 
         actions
             .get_execute_partial_query_packets()?
@@ -339,7 +335,6 @@ impl DataExchangeManager {
     fn get_root_pipeline(
         &self,
         ctx: Arc<QueryContext>,
-        enable_profiling: bool,
         root_actions: &QueryFragmentActions,
     ) -> Result<PipelineBuildResult> {
         let query_id = ctx.get_id();
@@ -353,12 +348,8 @@ impl DataExchangeManager {
             Some(query_coordinator) => {
                 assert!(query_coordinator.fragment_exchanges.is_empty());
                 let injector = DefaultExchangeInjector::create();
-                let mut build_res = query_coordinator.subscribe_fragment(
-                    &ctx,
-                    enable_profiling,
-                    fragment_id,
-                    injector,
-                )?;
+                let mut build_res =
+                    query_coordinator.subscribe_fragment(&ctx, fragment_id, injector)?;
 
                 let exchanges = std::mem::take(&mut query_coordinator.statistics_exchanges);
                 let statistics_receiver = StatisticsReceiver::spawn_receiver(&ctx, exchanges)?;
@@ -415,7 +406,6 @@ impl DataExchangeManager {
         &self,
         query_id: &str,
         fragment_id: usize,
-        enable_profiling: bool,
         injector: Arc<dyn ExchangeInjector>,
     ) -> Result<PipelineBuildResult> {
         let queries_coordinator_guard = self.queries_coordinator.lock();
@@ -431,12 +421,7 @@ impl DataExchangeManager {
                     .query_ctx
                     .clone();
 
-                query_coordinator.subscribe_fragment(
-                    &query_ctx,
-                    enable_profiling,
-                    fragment_id,
-                    injector,
-                )
+                query_coordinator.subscribe_fragment(&query_ctx, fragment_id, injector)
             }
         }
     }
@@ -599,7 +584,6 @@ impl QueryCoordinator {
     pub fn prepare_pipeline(
         &mut self,
         ctx: &Arc<QueryContext>,
-        enable_profiling: bool,
         packet: &QueryFragmentsPlanPacket,
     ) -> Result<()> {
         self.info = Some(QueryInfo {
@@ -619,7 +603,7 @@ impl QueryCoordinator {
         for fragment in &packet.fragments {
             let fragment_id = fragment.fragment_id;
             if let Some(coordinator) = self.fragments_coordinator.get_mut(&fragment_id) {
-                coordinator.prepare_pipeline(ctx.clone(), enable_profiling)?;
+                coordinator.prepare_pipeline(ctx.clone())?;
             }
         }
 
@@ -629,14 +613,13 @@ impl QueryCoordinator {
     pub fn subscribe_fragment(
         &mut self,
         ctx: &Arc<QueryContext>,
-        enable_profiling: bool,
         fragment_id: usize,
         injector: Arc<dyn ExchangeInjector>,
     ) -> Result<PipelineBuildResult> {
         // Merge pipelines if exist locally pipeline
         if let Some(mut fragment_coordinator) = self.fragments_coordinator.remove(&fragment_id) {
             let info = self.info.as_ref().expect("QueryInfo is none");
-            fragment_coordinator.prepare_pipeline(ctx.clone(), enable_profiling)?;
+            fragment_coordinator.prepare_pipeline(ctx.clone())?;
 
             if fragment_coordinator.pipeline_build_res.is_none() {
                 return Err(ErrorCode::Internal(
@@ -843,11 +826,7 @@ impl FragmentCoordinator {
         Err(ErrorCode::Internal("Cannot find data exchange."))
     }
 
-    pub fn prepare_pipeline(
-        &mut self,
-        ctx: Arc<QueryContext>,
-        enable_profiling: bool,
-    ) -> Result<()> {
+    pub fn prepare_pipeline(&mut self, ctx: Arc<QueryContext>) -> Result<()> {
         if !self.initialized {
             self.initialized = true;
 
@@ -857,8 +836,6 @@ impl FragmentCoordinator {
                 pipeline_ctx.get_function_context()?,
                 pipeline_ctx.get_settings(),
                 pipeline_ctx,
-                enable_profiling,
-                SharedProcessorProfiles::default(),
                 vec![],
             );
 
diff --git a/src/query/service/src/api/rpc/exchange/exchange_sink_writer.rs b/src/query/service/src/api/rpc/exchange/exchange_sink_writer.rs
index 4de8c9941cc6..90c8ed4093ed 100644
--- a/src/query/service/src/api/rpc/exchange/exchange_sink_writer.rs
+++ b/src/query/service/src/api/rpc/exchange/exchange_sink_writer.rs
@@ -12,8 +12,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::atomic::AtomicUsize;
-use std::sync::atomic::Ordering;
 use std::sync::Arc;
 
 use databend_common_catalog::table_context::TableContext;
@@ -26,6 +24,7 @@ use databend_common_pipeline_core::processors::profile::Profile;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_core::PipeItem;
 use databend_common_pipeline_sinks::AsyncSink;
 use databend_common_pipeline_sinks::AsyncSinker;
@@ -41,7 +40,6 @@ pub struct ExchangeWriterSink {
     source: String,
     destination: String,
     fragment: usize,
-    exchange_bytes: AtomicUsize,
 }
 
 impl ExchangeWriterSink {
@@ -58,7 +56,6 @@ impl ExchangeWriterSink {
             source: source_id.to_string(),
             destination: destination_id.to_string(),
             fragment: fragment_id,
-            exchange_bytes: AtomicUsize::new(0),
         })
     }
 }
@@ -101,19 +98,12 @@ impl AsyncSink for ExchangeWriterSink {
         {
             metrics_inc_exchange_write_count(count);
             metrics_inc_exchange_write_bytes(bytes);
-            self.exchange_bytes.fetch_add(bytes, Ordering::Relaxed);
+            Profile::record_usize_profile(ProfileStatisticsName::ExchangeBytes, bytes);
         }
 
         Ok(false)
     }
 
-    fn record_profile(&self, profile: &Profile) {
-        profile.exchange_bytes.fetch_add(
-            self.exchange_bytes.swap(0, Ordering::Relaxed),
-            Ordering::Relaxed,
-        );
-    }
-
     fn details_status(&self) -> Option<String> {
         #[derive(Debug)]
         #[allow(dead_code)]
diff --git a/src/query/service/src/api/rpc/exchange/serde/exchange_serializer.rs b/src/query/service/src/api/rpc/exchange/serde/exchange_serializer.rs
index 9599311c0fbf..00f645232d78 100644
--- a/src/query/service/src/api/rpc/exchange/serde/exchange_serializer.rs
+++ b/src/query/service/src/api/rpc/exchange/serde/exchange_serializer.rs
@@ -14,8 +14,6 @@
 
 use std::fmt::Debug;
 use std::fmt::Formatter;
-use std::sync::atomic::AtomicUsize;
-use std::sync::atomic::Ordering;
 use std::sync::Arc;
 
 use databend_common_arrow::arrow::chunk::Chunk;
@@ -32,10 +30,11 @@ use databend_common_expression::BlockMetaInfoPtr;
 use databend_common_expression::DataBlock;
 use databend_common_io::prelude::bincode_serialize_into_buf;
 use databend_common_io::prelude::BinaryWrite;
-use databend_common_pipeline_core::processors::profile::Profile;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_transforms::processors::BlockMetaTransform;
 use databend_common_pipeline_transforms::processors::BlockMetaTransformer;
 use databend_common_pipeline_transforms::processors::Transform;
@@ -99,7 +98,6 @@ impl BlockMetaInfo for ExchangeSerializeMeta {
 pub struct TransformExchangeSerializer {
     options: WriteOptions,
     ipc_fields: Vec<IpcField>,
-    exchange_rows: AtomicUsize,
 }
 
 impl TransformExchangeSerializer {
@@ -125,7 +123,6 @@ impl TransformExchangeSerializer {
             TransformExchangeSerializer {
                 ipc_fields,
                 options: WriteOptions { compression },
-                exchange_rows: AtomicUsize::new(0),
             },
         )))
     }
@@ -135,17 +132,9 @@ impl Transform for TransformExchangeSerializer {
     const NAME: &'static str = "ExchangeSerializerTransform";
 
     fn transform(&mut self, data_block: DataBlock) -> Result<DataBlock> {
-        self.exchange_rows
-            .fetch_add(data_block.num_rows(), Ordering::Relaxed);
+        Profile::record_usize_profile(ProfileStatisticsName::ExchangeRows, data_block.num_rows());
         serialize_block(0, data_block, &self.ipc_fields, &self.options)
     }
-
-    fn record_profile(&self, profile: &Profile) {
-        profile.exchange_rows.fetch_add(
-            self.exchange_rows.swap(0, Ordering::Relaxed),
-            Ordering::Relaxed,
-        );
-    }
 }
 
 pub struct TransformScatterExchangeSerializer {
diff --git a/src/query/service/src/api/rpc/packets/packet_executor.rs b/src/query/service/src/api/rpc/packets/packet_executor.rs
index 77272cee6646..430a6630ecfd 100644
--- a/src/query/service/src/api/rpc/packets/packet_executor.rs
+++ b/src/query/service/src/api/rpc/packets/packet_executor.rs
@@ -38,8 +38,6 @@ pub struct QueryFragmentsPlanPacket {
     pub changed_settings: Arc<Settings>,
     // We send nodes info for each node. This is a bad choice
     pub executors_info: HashMap<String, Arc<NodeInfo>>,
-    /// Enable profiling for this query
-    pub enable_profiling: bool,
 }
 
 impl QueryFragmentsPlanPacket {
@@ -52,7 +50,6 @@ impl QueryFragmentsPlanPacket {
         executors_info: HashMap<String, Arc<NodeInfo>>,
         changed_settings: Arc<Settings>,
         request_executor: String,
-        enable_profiling: bool,
     ) -> QueryFragmentsPlanPacket {
         QueryFragmentsPlanPacket {
             query_id,
@@ -62,7 +59,6 @@ impl QueryFragmentsPlanPacket {
             executors_info,
             changed_settings,
             request_executor,
-            enable_profiling,
         }
     }
 }
diff --git a/src/query/service/src/databases/system/system_database.rs b/src/query/service/src/databases/system/system_database.rs
index 328b118121b1..04b686a5d6db 100644
--- a/src/query/service/src/databases/system/system_database.rs
+++ b/src/query/service/src/databases/system/system_database.rs
@@ -46,8 +46,6 @@ use databend_common_storages_system::ProcessesTable;
 use databend_common_storages_system::ProcessorProfileTable;
 use databend_common_storages_system::QueryCacheTable;
 use databend_common_storages_system::QueryLogTable;
-use databend_common_storages_system::QueryProfileTable;
-use databend_common_storages_system::QuerySummaryTable;
 use databend_common_storages_system::RolesTable;
 use databend_common_storages_system::SettingsTable;
 use databend_common_storages_system::StagesTable;
@@ -117,12 +115,10 @@ impl SystemDatabase {
             TableFunctionsTable::create(sys_db_meta.next_table_id()),
             CachesTable::create(sys_db_meta.next_table_id()),
             IndexesTable::create(sys_db_meta.next_table_id()),
-            QueryProfileTable::create(sys_db_meta.next_table_id()),
             BackgroundTaskTable::create(sys_db_meta.next_table_id()),
             BackgroundJobTable::create(sys_db_meta.next_table_id()),
             BacktraceTable::create(sys_db_meta.next_table_id()),
             TempFilesTable::create(sys_db_meta.next_table_id()),
-            QuerySummaryTable::create(sys_db_meta.next_table_id()),
             TasksTable::create(sys_db_meta.next_table_id()),
             TaskHistoryTable::create(sys_db_meta.next_table_id()),
             ProcessorProfileTable::create(sys_db_meta.next_table_id()),
diff --git a/src/query/service/src/global_services.rs b/src/query/service/src/global_services.rs
index 72061499d66d..3dbe19e6284d 100644
--- a/src/query/service/src/global_services.rs
+++ b/src/query/service/src/global_services.rs
@@ -25,7 +25,6 @@ use databend_common_config::GlobalConfig;
 use databend_common_config::InnerConfig;
 use databend_common_exception::Result;
 use databend_common_meta_app::schema::CatalogType;
-use databend_common_profile::QueryProfileManager;
 use databend_common_sharing::ShareEndpointManager;
 use databend_common_storage::DataOperator;
 use databend_common_storage::ShareTableConfig;
@@ -110,7 +109,6 @@ impl GlobalServices {
         .await?;
         RoleCacheManager::init()?;
         ShareEndpointManager::init()?;
-        QueryProfileManager::init();
 
         DataOperator::init(&config.storage).await?;
         ShareTableConfig::init(
diff --git a/src/query/service/src/interpreters/interpreter.rs b/src/query/service/src/interpreters/interpreter.rs
index 6e8b2f981fa9..dbdd515fd560 100644
--- a/src/query/service/src/interpreters/interpreter.rs
+++ b/src/query/service/src/interpreters/interpreter.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::HashMap;
 use std::sync::Arc;
 use std::time::SystemTime;
 
@@ -19,7 +20,10 @@ use databend_common_catalog::table_context::TableContext;
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_expression::SendableDataBlockStream;
+use databend_common_pipeline_core::get_statistics_desc;
 use databend_common_pipeline_core::processors::profile::PlanProfile;
+use databend_common_pipeline_core::processors::ProfileDesc;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_core::SourcePipeBuilder;
 use log::error;
 use log::info;
@@ -91,12 +95,18 @@ pub trait Interpreter: Sync + Send {
                     struct QueryProfiles {
                         query_id: String,
                         profiles: Vec<PlanProfile>,
+                        statistics_desc: Arc<HashMap<ProfileStatisticsName, ProfileDesc>>,
                     }
 
-                    info!(target: "databend::log::profile", "{}", serde_json::to_string(&QueryProfiles {
-                        query_id: query_ctx.get_id(),
-                        profiles: query_profiles,
-                    })?);
+                    info!(
+                        target: "databend::log::profile",
+                        "{}",
+                        serde_json::to_string(&QueryProfiles {
+                            query_id: query_ctx.get_id(),
+                            profiles: query_profiles,
+                            statistics_desc: get_statistics_desc(),
+                        })?
+                    );
                 }
             }
 
diff --git a/src/query/service/src/interpreters/interpreter_catalog_show_create.rs b/src/query/service/src/interpreters/interpreter_catalog_show_create.rs
index df5a8c81066a..e16b571ac897 100644
--- a/src/query/service/src/interpreters/interpreter_catalog_show_create.rs
+++ b/src/query/service/src/interpreters/interpreter_catalog_show_create.rs
@@ -72,18 +72,12 @@ impl Interpreter for ShowCreateCatalogInterpreter {
 
         let block = DataBlock::new(
             vec![
+                BlockEntry::new(DataType::String, Value::Scalar(Scalar::String(name))),
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(name.into_bytes())),
-                ),
-                BlockEntry::new(
-                    DataType::String,
-                    Value::Scalar(Scalar::String(catalog_type.into_bytes())),
-                ),
-                BlockEntry::new(
-                    DataType::String,
-                    Value::Scalar(Scalar::String(option.into_bytes())),
+                    Value::Scalar(Scalar::String(catalog_type)),
                 ),
+                BlockEntry::new(DataType::String, Value::Scalar(Scalar::String(option))),
             ],
             1,
         );
diff --git a/src/query/service/src/interpreters/interpreter_connection_desc.rs b/src/query/service/src/interpreters/interpreter_connection_desc.rs
index 8f15a2f35cb4..19cf3ed569ce 100644
--- a/src/query/service/src/interpreters/interpreter_connection_desc.rs
+++ b/src/query/service/src/interpreters/interpreter_connection_desc.rs
@@ -54,11 +54,11 @@ impl Interpreter for DescConnectionInterpreter {
             .get_connection(&tenant, self.plan.name.as_str())
             .await?;
 
-        let names = vec![connection.name.as_bytes().to_vec()];
-        let types = vec![connection.storage_type.as_bytes().to_vec()];
+        let names = vec![connection.name.clone()];
+        let types = vec![connection.storage_type.clone()];
         let conn = Connection::new(connection.storage_params).mask();
         connection.storage_params = conn.conns;
-        let params = vec![connection.storage_params_display().as_bytes().to_vec()];
+        let params = vec![connection.storage_params_display().clone()];
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
             StringType::from_data(names),
diff --git a/src/query/service/src/interpreters/interpreter_connection_show.rs b/src/query/service/src/interpreters/interpreter_connection_show.rs
index 4f5683daf83d..42b65606c62a 100644
--- a/src/query/service/src/interpreters/interpreter_connection_show.rs
+++ b/src/query/service/src/interpreters/interpreter_connection_show.rs
@@ -55,14 +55,11 @@ impl Interpreter for ShowConnectionsInterpreter {
 
         formats.sort_by(|a, b| a.name.cmp(&b.name));
 
-        let names = formats
-            .iter()
-            .map(|x| x.name.as_bytes().to_vec())
-            .collect::<Vec<_>>();
+        let names = formats.iter().map(|x| x.name.clone()).collect::<Vec<_>>();
 
         let types = formats
             .iter()
-            .map(|x| x.storage_type.as_bytes().to_vec())
+            .map(|x| x.storage_type.clone())
             .collect::<Vec<_>>();
 
         let options = formats
@@ -70,7 +67,7 @@ impl Interpreter for ShowConnectionsInterpreter {
             .map(|x| {
                 let conn = Connection::new(x.storage_params.clone()).mask();
                 x.storage_params = conn.conns;
-                x.storage_params_display().as_bytes().to_vec()
+                x.storage_params_display().clone()
             })
             .collect::<Vec<_>>();
 
diff --git a/src/query/service/src/interpreters/interpreter_copy_into_table.rs b/src/query/service/src/interpreters/interpreter_copy_into_table.rs
index 5c2042f4da49..1fba025d4cb1 100644
--- a/src/query/service/src/interpreters/interpreter_copy_into_table.rs
+++ b/src/query/service/src/interpreters/interpreter_copy_into_table.rs
@@ -206,13 +206,13 @@ impl CopyIntoTableInterpreter {
         for entry in results {
             let status = entry.value();
             if let Some(err) = &status.error {
-                files.push(entry.key().as_bytes().to_vec());
+                files.push(entry.key().clone());
                 rows_loaded.push(status.num_rows_loaded as i32);
                 errors_seen.push(err.num_errors as i32);
-                first_error.push(Some(err.first_error.error.to_string().as_bytes().to_vec()));
+                first_error.push(Some(err.first_error.error.to_string().clone()));
                 first_error_line.push(Some(err.first_error.line as i32 + 1));
             } else if return_all {
-                files.push(entry.key().as_bytes().to_vec());
+                files.push(entry.key().clone());
                 rows_loaded.push(status.num_rows_loaded as i32);
                 errors_seen.push(0);
                 first_error.push(None);
@@ -304,8 +304,7 @@ impl Interpreter for CopyIntoTableInterpreter {
         let (physical_plan, files, update_stream_meta) =
             self.build_physical_plan(&self.plan).await?;
         let mut build_res =
-            build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                .await?;
+            build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
 
         // Build commit insertion pipeline.
         {
diff --git a/src/query/service/src/interpreters/interpreter_data_mask_desc.rs b/src/query/service/src/interpreters/interpreter_data_mask_desc.rs
index d5a3c6ba259b..491bed72cd66 100644
--- a/src/query/service/src/interpreters/interpreter_data_mask_desc.rs
+++ b/src/query/service/src/interpreters/interpreter_data_mask_desc.rs
@@ -71,8 +71,8 @@ impl Interpreter for DescDataMaskInterpreter {
             }
         };
 
-        let name: Vec<Vec<u8>> = vec![self.plan.name.as_bytes().to_vec()];
-        let create_on: Vec<Vec<u8>> = vec![policy.create_on.to_string().as_bytes().to_vec()];
+        let name: Vec<String> = vec![self.plan.name.clone()];
+        let create_on: Vec<String> = vec![policy.create_on.to_string().clone()];
         let args = format!(
             "({})",
             policy
@@ -83,12 +83,12 @@ impl Interpreter for DescDataMaskInterpreter {
                 .collect::<Vec<_>>()
                 .join(",")
         );
-        let signature: Vec<Vec<u8>> = vec![args.as_bytes().to_vec()];
-        let return_type = vec![policy.return_type.as_bytes().to_vec()];
-        let body = vec![policy.body.as_bytes().to_vec()];
+        let signature: Vec<String> = vec![args.clone()];
+        let return_type = vec![policy.return_type.clone()];
+        let body = vec![policy.body.clone()];
         let comment = vec![match policy.comment {
-            Some(comment) => comment.as_bytes().to_vec(),
-            None => "".to_string().as_bytes().to_vec(),
+            Some(comment) => comment.clone(),
+            None => "".to_string().clone(),
         }];
 
         let blocks = vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_database_show_create.rs b/src/query/service/src/interpreters/interpreter_database_show_create.rs
index 7211abb0438a..7a5df3743619 100644
--- a/src/query/service/src/interpreters/interpreter_database_show_create.rs
+++ b/src/query/service/src/interpreters/interpreter_database_show_create.rs
@@ -74,11 +74,11 @@ impl Interpreter for ShowCreateDatabaseInterpreter {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(name.as_bytes().to_vec())),
+                    Value::Scalar(Scalar::String(name.to_string())),
                 ),
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(info.as_bytes().to_vec())),
+                    Value::Scalar(Scalar::String(info.clone())),
                 ),
             ],
             1,
diff --git a/src/query/service/src/interpreters/interpreter_delete.rs b/src/query/service/src/interpreters/interpreter_delete.rs
index 15d6ca7ae3dc..ac9168212296 100644
--- a/src/query/service/src/interpreters/interpreter_delete.rs
+++ b/src/query/service/src/interpreters/interpreter_delete.rs
@@ -230,8 +230,7 @@ impl Interpreter for DeleteInterpreter {
             )?;
 
             build_res =
-                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                    .await?;
+                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
         }
 
         build_res.main_pipeline.add_lock_guard(lock_guard);
diff --git a/src/query/service/src/interpreters/interpreter_explain.rs b/src/query/service/src/interpreters/interpreter_explain.rs
index 1e2662543f21..dd4149fdfc2d 100644
--- a/src/query/service/src/interpreters/interpreter_explain.rs
+++ b/src/query/service/src/interpreters/interpreter_explain.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::HashMap;
 use std::sync::Arc;
 
 use databend_common_ast::ast::ExplainKind;
@@ -22,9 +23,7 @@ use databend_common_exception::Result;
 use databend_common_expression::types::StringType;
 use databend_common_expression::DataBlock;
 use databend_common_expression::FromData;
-use databend_common_profile::QueryProfileManager;
-use databend_common_profile::SharedProcessorProfiles;
-use databend_common_sql::executor::ProfileHelper;
+use databend_common_pipeline_core::processors::profile::PlanProfile;
 use databend_common_sql::optimizer::ColumnSet;
 use databend_common_sql::plans::UpdatePlan;
 use databend_common_sql::BindContext;
@@ -260,7 +259,7 @@ impl ExplainInterpreter {
         }
 
         let result = plan
-            .format(metadata.clone(), SharedProcessorProfiles::default())?
+            .format(metadata.clone(), Default::default())?
             .format_pretty()?;
         let line_split_result: Vec<&str> = result.lines().collect();
         let formatted_plan = StringType::from_data(line_split_result);
@@ -283,7 +282,7 @@ impl ExplainInterpreter {
         // Format root pipeline
         let line_split_result = format!("{}", build_res.main_pipeline.display_indent())
             .lines()
-            .map(|s| s.as_bytes().to_vec())
+            .map(|l| l.to_string())
             .collect::<Vec<_>>();
         let column = StringType::from_data(line_split_result);
         blocks.push(DataBlock::new_from_columns(vec![column]));
@@ -291,7 +290,7 @@ impl ExplainInterpreter {
         for pipeline in build_res.sources_pipelines.iter() {
             let line_split_result = format!("\n{}", pipeline.display_indent())
                 .lines()
-                .map(|s| s.as_bytes().to_vec())
+                .map(|l| l.to_string())
                 .collect::<Vec<_>>();
             let column = StringType::from_data(line_split_result);
             blocks.push(DataBlock::new_from_columns(vec![column]));
@@ -313,14 +312,11 @@ impl ExplainInterpreter {
 
         let root_fragment = Fragmenter::try_create(ctx.clone())?.build_fragment(&plan)?;
 
-        let mut fragments_actions = QueryFragmentsActions::create(ctx.clone(), false);
+        let mut fragments_actions = QueryFragmentsActions::create(ctx.clone());
         root_fragment.get_actions(ctx, &mut fragments_actions)?;
 
         let display_string = fragments_actions.display_indent(&metadata).to_string();
-        let line_split_result = display_string
-            .lines()
-            .map(|s| s.as_bytes().to_vec())
-            .collect::<Vec<_>>();
+        let line_split_result = display_string.lines().collect::<Vec<_>>();
         let formatted_plan = StringType::from_data(line_split_result);
         Ok(vec![DataBlock::new_from_columns(vec![formatted_plan])])
     }
@@ -331,7 +327,7 @@ impl ExplainInterpreter {
         let display_string = if let Some(plan) = interpreter.get_physical_plan().await? {
             let root_fragment = Fragmenter::try_create(self.ctx.clone())?.build_fragment(&plan)?;
 
-            let mut fragments_actions = QueryFragmentsActions::create(self.ctx.clone(), false);
+            let mut fragments_actions = QueryFragmentsActions::create(self.ctx.clone());
             root_fragment.get_actions(self.ctx.clone(), &mut fragments_actions)?;
 
             let ident = fragments_actions.display_indent(&update.metadata);
@@ -339,10 +335,7 @@ impl ExplainInterpreter {
         } else {
             "Nothing to update".to_string()
         };
-        let line_split_result = display_string
-            .lines()
-            .map(|s| s.as_bytes().to_vec())
-            .collect::<Vec<_>>();
+        let line_split_result = display_string.lines().collect::<Vec<_>>();
         let formatted_plan = StringType::from_data(line_split_result);
         Ok(vec![DataBlock::new_from_columns(vec![formatted_plan])])
     }
@@ -357,46 +350,68 @@ impl ExplainInterpreter {
     ) -> Result<Vec<DataBlock>> {
         let mut builder = PhysicalPlanBuilder::new(metadata.clone(), self.ctx.clone(), true);
         let plan = builder.build(s_expr, required).await?;
-        let mut build_res = build_query_pipeline(&self.ctx, &[], &plan, ignore_result).await?;
+        let build_res = build_query_pipeline(&self.ctx, &[], &plan, ignore_result).await?;
 
-        let prof_span_set = build_res.prof_span_set.clone();
+        // Drain the data
+        let query_profiles = self.execute_and_get_profiles(build_res)?;
 
+        let result = plan
+            .format(metadata.clone(), query_profiles)?
+            .format_pretty()?;
+        let line_split_result: Vec<&str> = result.lines().collect();
+        let formatted_plan = StringType::from_data(line_split_result);
+        Ok(vec![DataBlock::new_from_columns(vec![formatted_plan])])
+    }
+
+    fn execute_and_get_profiles(
+        &self,
+        mut build_res: PipelineBuildResult,
+    ) -> Result<HashMap<u32, PlanProfile>> {
         let settings = self.ctx.get_settings();
         let query_id = self.ctx.get_id();
         build_res.set_max_threads(settings.get_max_threads()? as usize);
         let settings = ExecutorSettings::try_create(&settings, query_id.clone())?;
 
-        // Drain the data
-        if build_res.main_pipeline.is_complete_pipeline()? {
-            let mut pipelines = build_res.sources_pipelines;
-            pipelines.push(build_res.main_pipeline);
-
-            let complete_executor = PipelineCompleteExecutor::from_pipelines(pipelines, settings)?;
-            complete_executor.execute()?;
-        } else {
-            let mut pulling_executor =
-                PipelinePullingExecutor::from_pipelines(build_res, settings)?;
-            pulling_executor.start();
-            while (pulling_executor.pull_data()?).is_some() {}
+        match build_res.main_pipeline.is_complete_pipeline()? {
+            true => {
+                let mut pipelines = build_res.sources_pipelines;
+                pipelines.push(build_res.main_pipeline);
+
+                let executor = PipelineCompleteExecutor::from_pipelines(pipelines, settings)?;
+                executor.execute()?;
+                self.ctx.add_query_profiles(
+                    &executor
+                        .get_inner()
+                        .get_profiles()
+                        .iter()
+                        .filter(|x| x.plan_id.is_some())
+                        .map(|x| PlanProfile::create(x))
+                        .collect::<Vec<_>>(),
+                );
+            }
+            false => {
+                let mut executor = PipelinePullingExecutor::from_pipelines(build_res, settings)?;
+                executor.start();
+                while (executor.pull_data()?).is_some() {}
+                self.ctx.add_query_profiles(
+                    &executor
+                        .get_inner()
+                        .get_profiles()
+                        .iter()
+                        .filter(|x| x.plan_id.is_some())
+                        .map(|x| PlanProfile::create(x))
+                        .collect::<Vec<_>>(),
+                );
+            }
         }
 
-        let profile = ProfileHelper::build_query_profile(
-            &query_id,
-            metadata,
-            &plan,
-            &prof_span_set.lock().unwrap(),
-        )?;
-
-        // Record the query profile
-        let prof_mgr = QueryProfileManager::instance();
-        prof_mgr.insert(Arc::new(profile));
-
-        let result = plan
-            .format(metadata.clone(), prof_span_set)?
-            .format_pretty()?;
-        let line_split_result: Vec<&str> = result.lines().collect();
-        let formatted_plan = StringType::from_data(line_split_result);
-        Ok(vec![DataBlock::new_from_columns(vec![formatted_plan])])
+        Ok(self
+            .ctx
+            .get_query_profiles()
+            .into_iter()
+            .filter(|x| x.id.is_some())
+            .map(|x| (x.id.unwrap(), x))
+            .collect::<HashMap<_, _>>())
     }
 
     async fn explain_query(
diff --git a/src/query/service/src/interpreters/interpreter_file_format_show.rs b/src/query/service/src/interpreters/interpreter_file_format_show.rs
index dace4ee3430b..93edb23c74d7 100644
--- a/src/query/service/src/interpreters/interpreter_file_format_show.rs
+++ b/src/query/service/src/interpreters/interpreter_file_format_show.rs
@@ -54,14 +54,11 @@ impl Interpreter for ShowFileFormatsInterpreter {
 
         formats.sort_by(|a, b| a.name.cmp(&b.name));
 
-        let names = formats
-            .iter()
-            .map(|x| x.name.as_bytes().to_vec())
-            .collect::<Vec<_>>();
+        let names = formats.iter().map(|x| x.name.clone()).collect::<Vec<_>>();
 
         let options = formats
             .iter()
-            .map(|x| x.file_format_params.to_string().as_bytes().to_vec())
+            .map(|x| x.file_format_params.to_string())
             .collect::<Vec<_>>();
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_index_refresh.rs b/src/query/service/src/interpreters/interpreter_index_refresh.rs
index 4058364f67e4..3edbef5be492 100644
--- a/src/query/service/src/interpreters/interpreter_index_refresh.rs
+++ b/src/query/service/src/interpreters/interpreter_index_refresh.rs
@@ -291,7 +291,7 @@ impl Interpreter for RefreshIndexInterpreter {
         query_plan = replace_read_source.replace(&query_plan)?;
 
         let mut build_res =
-            build_query_pipeline_without_render_result_set(&self.ctx, &query_plan, false).await?;
+            build_query_pipeline_without_render_result_set(&self.ctx, &query_plan).await?;
 
         let input_schema = query_plan.output_schema()?;
 
diff --git a/src/query/service/src/interpreters/interpreter_insert.rs b/src/query/service/src/interpreters/interpreter_insert.rs
index 9bd6b8dd4e14..ad60dc2b4404 100644
--- a/src/query/service/src/interpreters/interpreter_insert.rs
+++ b/src/query/service/src/interpreters/interpreter_insert.rs
@@ -230,12 +230,9 @@ impl Interpreter for InsertInterpreter {
                     }
                 };
 
-                let mut build_res = build_query_pipeline_without_render_result_set(
-                    &self.ctx,
-                    &insert_select_plan,
-                    false,
-                )
-                .await?;
+                let mut build_res =
+                    build_query_pipeline_without_render_result_set(&self.ctx, &insert_select_plan)
+                        .await?;
 
                 table.commit_insertion(
                     self.ctx.clone(),
diff --git a/src/query/service/src/interpreters/interpreter_merge_into.rs b/src/query/service/src/interpreters/interpreter_merge_into.rs
index 930a24c5cdc5..88ac30ca6222 100644
--- a/src/query/service/src/interpreters/interpreter_merge_into.rs
+++ b/src/query/service/src/interpreters/interpreter_merge_into.rs
@@ -92,8 +92,7 @@ impl Interpreter for MergeIntoInterpreter {
     async fn execute2(&self) -> Result<PipelineBuildResult> {
         let (physical_plan, _) = self.build_physical_plan().await?;
         let mut build_res =
-            build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                .await?;
+            build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
 
         // Add table lock before execution.
         // todo!(@zhyass) :But for now the lock maybe exist problem, let's open this after fix it.
diff --git a/src/query/service/src/interpreters/interpreter_network_policies_show.rs b/src/query/service/src/interpreters/interpreter_network_policies_show.rs
index d92d0d1be1c9..f38be2b417f4 100644
--- a/src/query/service/src/interpreters/interpreter_network_policies_show.rs
+++ b/src/query/service/src/interpreters/interpreter_network_policies_show.rs
@@ -53,10 +53,10 @@ impl Interpreter for ShowNetworkPoliciesInterpreter {
         let mut blocked_ip_lists = Vec::with_capacity(network_policies.len());
         let mut comments = Vec::with_capacity(network_policies.len());
         for network_policy in network_policies {
-            names.push(network_policy.name.as_bytes().to_vec());
-            allowed_ip_lists.push(network_policy.allowed_ip_list.join(",").as_bytes().to_vec());
-            blocked_ip_lists.push(network_policy.blocked_ip_list.join(",").as_bytes().to_vec());
-            comments.push(network_policy.comment.as_bytes().to_vec());
+            names.push(network_policy.name.clone());
+            allowed_ip_lists.push(network_policy.allowed_ip_list.join(",").clone());
+            blocked_ip_lists.push(network_policy.blocked_ip_list.join(",").clone());
+            comments.push(network_policy.comment.clone());
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_network_policy_desc.rs b/src/query/service/src/interpreters/interpreter_network_policy_desc.rs
index 94d2fb1a0300..080b9e20a2ea 100644
--- a/src/query/service/src/interpreters/interpreter_network_policy_desc.rs
+++ b/src/query/service/src/interpreters/interpreter_network_policy_desc.rs
@@ -53,10 +53,10 @@ impl Interpreter for DescNetworkPolicyInterpreter {
             .get_network_policy(&tenant, self.plan.name.as_str())
             .await?;
 
-        let names = vec![network_policy.name.as_bytes().to_vec()];
-        let allowed_ip_lists = vec![network_policy.allowed_ip_list.join(",").as_bytes().to_vec()];
-        let blocked_ip_lists = vec![network_policy.blocked_ip_list.join(",").as_bytes().to_vec()];
-        let comments = vec![network_policy.comment.as_bytes().to_vec()];
+        let names = vec![network_policy.name.clone()];
+        let allowed_ip_lists = vec![network_policy.allowed_ip_list.join(",").clone()];
+        let blocked_ip_lists = vec![network_policy.blocked_ip_list.join(",").clone()];
+        let comments = vec![network_policy.comment.clone()];
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
             StringType::from_data(names),
diff --git a/src/query/service/src/interpreters/interpreter_password_policy_desc.rs b/src/query/service/src/interpreters/interpreter_password_policy_desc.rs
index 8aafd73e6aeb..cdc9759a3793 100644
--- a/src/query/service/src/interpreters/interpreter_password_policy_desc.rs
+++ b/src/query/service/src/interpreters/interpreter_password_policy_desc.rs
@@ -64,19 +64,19 @@ impl Interpreter for DescPasswordPolicyInterpreter {
             .await?;
 
         let properties = vec![
-            "NAME".as_bytes().to_vec(),
-            "COMMENT".as_bytes().to_vec(),
-            "PASSWORD_MIN_LENGTH".as_bytes().to_vec(),
-            "PASSWORD_MAX_LENGTH".as_bytes().to_vec(),
-            "PASSWORD_MIN_UPPER_CASE_CHARS".as_bytes().to_vec(),
-            "PASSWORD_MIN_LOWER_CASE_CHARS".as_bytes().to_vec(),
-            "PASSWORD_MIN_NUMERIC_CHARS".as_bytes().to_vec(),
-            "PASSWORD_MIN_SPECIAL_CHARS".as_bytes().to_vec(),
-            "PASSWORD_MIN_AGE_DAYS".as_bytes().to_vec(),
-            "PASSWORD_MAX_AGE_DAYS".as_bytes().to_vec(),
-            "PASSWORD_MAX_RETRIES".as_bytes().to_vec(),
-            "PASSWORD_LOCKOUT_TIME_MINS".as_bytes().to_vec(),
-            "PASSWORD_HISTORY".as_bytes().to_vec(),
+            "NAME".to_string(),
+            "COMMENT".to_string(),
+            "PASSWORD_MIN_LENGTH".to_string(),
+            "PASSWORD_MAX_LENGTH".to_string(),
+            "PASSWORD_MIN_UPPER_CASE_CHARS".to_string(),
+            "PASSWORD_MIN_LOWER_CASE_CHARS".to_string(),
+            "PASSWORD_MIN_NUMERIC_CHARS".to_string(),
+            "PASSWORD_MIN_SPECIAL_CHARS".to_string(),
+            "PASSWORD_MIN_AGE_DAYS".to_string(),
+            "PASSWORD_MAX_AGE_DAYS".to_string(),
+            "PASSWORD_MAX_RETRIES".to_string(),
+            "PASSWORD_LOCKOUT_TIME_MINS".to_string(),
+            "PASSWORD_HISTORY".to_string(),
         ];
 
         let min_length = format!("{}", password_policy.min_length);
@@ -92,19 +92,19 @@ impl Interpreter for DescPasswordPolicyInterpreter {
         let history = format!("{}", password_policy.history);
 
         let values = vec![
-            password_policy.name.as_bytes().to_vec(),
-            password_policy.comment.as_bytes().to_vec(),
-            min_length.as_bytes().to_vec(),
-            max_length.as_bytes().to_vec(),
-            min_upper_case_chars.as_bytes().to_vec(),
-            min_lower_case_chars.as_bytes().to_vec(),
-            min_numeric_chars.as_bytes().to_vec(),
-            min_special_chars.as_bytes().to_vec(),
-            min_age_days.as_bytes().to_vec(),
-            max_age_days.as_bytes().to_vec(),
-            max_retries.as_bytes().to_vec(),
-            lockout_time_mins.as_bytes().to_vec(),
-            history.as_bytes().to_vec(),
+            password_policy.name.clone(),
+            password_policy.comment.clone(),
+            min_length.clone(),
+            max_length.clone(),
+            min_upper_case_chars.clone(),
+            min_lower_case_chars.clone(),
+            min_numeric_chars.clone(),
+            min_special_chars.clone(),
+            min_age_days.clone(),
+            max_age_days.clone(),
+            max_retries.clone(),
+            lockout_time_mins.clone(),
+            history.clone(),
         ];
 
         let defaults = vec![
@@ -124,19 +124,19 @@ impl Interpreter for DescPasswordPolicyInterpreter {
         ];
 
         let descriptions = vec![
-            "Name of password policy.".as_bytes().to_vec(),
-            "Comment of password policy.".as_bytes().to_vec(),
-            "Minimum length of new password.".as_bytes().to_vec(),
-            "Maximum length of new password.".as_bytes().to_vec(),
-            "Minimum number of uppercase characters in new password.".as_bytes().to_vec(),
-            "Minimum number of lowercase characters in new password.".as_bytes().to_vec(),
-            "Minimum number of numeric characters in new password.".as_bytes().to_vec(),
-            "Minimum number of special characters in new password.".as_bytes().to_vec(),
-            "Period after a password is changed during which a password cannot be changed again, in days.".as_bytes().to_vec(),
-            "Period after which password must be changed, in days.".as_bytes().to_vec(),
-            "Number of attempts users have to enter the correct password before their account is locked.".as_bytes().to_vec(),
-            "Period of time for which users will be locked after entering their password incorrectly many times (specified by MAX_RETRIES), in minutes.".as_bytes().to_vec(),
-            "Number of most recent passwords that may not be repeated by the user.".as_bytes().to_vec(),
+            "Name of password policy.".to_string(),
+            "Comment of password policy.".to_string(),
+            "Minimum length of new password.".to_string(),
+            "Maximum length of new password.".to_string(),
+            "Minimum number of uppercase characters in new password.".to_string(),
+            "Minimum number of lowercase characters in new password.".to_string(),
+            "Minimum number of numeric characters in new password.".to_string(),
+            "Minimum number of special characters in new password.".to_string(),
+            "Period after a password is changed during which a password cannot be changed again, in days.".to_string(),
+            "Period after which password must be changed, in days.".to_string(),
+            "Number of attempts users have to enter the correct password before their account is locked.".to_string(),
+            "Period of time for which users will be locked after entering their password incorrectly many times (specified by MAX_RETRIES), in minutes.".to_string(),
+            "Number of most recent passwords that may not be repeated by the user.".to_string(),
         ];
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_presign.rs b/src/query/service/src/interpreters/interpreter_presign.rs
index 82cacd65c501..8ef714a32118 100644
--- a/src/query/service/src/interpreters/interpreter_presign.rs
+++ b/src/query/service/src/interpreters/interpreter_presign.rs
@@ -95,9 +95,7 @@ impl Interpreter for PresignInterpreter {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(
-                        presigned_req.method().as_str().as_bytes().to_vec(),
-                    )),
+                    Value::Scalar(Scalar::String(presigned_req.method().as_str().to_string())),
                 ),
                 BlockEntry::new(
                     DataType::Variant,
@@ -105,9 +103,7 @@ impl Interpreter for PresignInterpreter {
                 ),
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(
-                        presigned_req.uri().to_string().as_bytes().to_vec(),
-                    )),
+                    Value::Scalar(Scalar::String(presigned_req.uri().to_string())),
                 ),
             ],
             1,
diff --git a/src/query/service/src/interpreters/interpreter_replace.rs b/src/query/service/src/interpreters/interpreter_replace.rs
index 6834a4e1d0df..71f740e9fb20 100644
--- a/src/query/service/src/interpreters/interpreter_replace.rs
+++ b/src/query/service/src/interpreters/interpreter_replace.rs
@@ -86,8 +86,7 @@ impl Interpreter for ReplaceInterpreter {
         // replace
         let (physical_plan, purge_info) = self.build_physical_plan().await?;
         let mut pipeline =
-            build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                .await?;
+            build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
 
         // purge
         if let Some((files, stage_info)) = purge_info {
diff --git a/src/query/service/src/interpreters/interpreter_role_show.rs b/src/query/service/src/interpreters/interpreter_role_show.rs
index e39322148dbe..c6023caf851d 100644
--- a/src/query/service/src/interpreters/interpreter_role_show.rs
+++ b/src/query/service/src/interpreters/interpreter_role_show.rs
@@ -64,10 +64,7 @@ impl Interpreter for ShowRolesInterpreter {
             .cloned()
             .unwrap_or_default();
 
-        let names = roles
-            .iter()
-            .map(|x| x.name.as_bytes().to_vec())
-            .collect::<Vec<_>>();
+        let names = roles.iter().map(|x| x.name.clone()).collect::<Vec<_>>();
         let inherited_roles: Vec<u64> = roles
             .iter()
             .map(|x| x.grants.roles().len() as u64)
diff --git a/src/query/service/src/interpreters/interpreter_select.rs b/src/query/service/src/interpreters/interpreter_select.rs
index 4c1493ea61f5..ff0d3a1250fc 100644
--- a/src/query/service/src/interpreters/interpreter_select.rs
+++ b/src/query/service/src/interpreters/interpreter_select.rs
@@ -26,7 +26,6 @@ use databend_common_pipeline_core::Pipe;
 use databend_common_pipeline_core::PipeItem;
 use databend_common_pipeline_core::Pipeline;
 use databend_common_pipeline_transforms::processors::TransformDummy;
-use databend_common_profile::SharedProcessorProfiles;
 use databend_common_sql::executor::physical_plans::FragmentKind;
 use databend_common_sql::executor::PhysicalPlan;
 use databend_common_sql::parse_result_scan_args;
@@ -233,7 +232,7 @@ impl Interpreter for SelectInterpreter {
         // 0. Need to build physical plan first to get the partitions.
         let physical_plan = self.build_physical_plan().await?;
         let query_plan = physical_plan
-            .format(self.metadata.clone(), SharedProcessorProfiles::default())?
+            .format(self.metadata.clone(), Default::default())?
             .format_pretty()?;
         info!(
             "Query id: {}, query plan: \n{}",
diff --git a/src/query/service/src/interpreters/interpreter_share_desc.rs b/src/query/service/src/interpreters/interpreter_share_desc.rs
index 815410d9660e..32ea39406a5f 100644
--- a/src/query/service/src/interpreters/interpreter_share_desc.rs
+++ b/src/query/service/src/interpreters/interpreter_share_desc.rs
@@ -61,21 +61,21 @@ impl Interpreter for DescShareInterpreter {
             return Ok(PipelineBuildResult::create());
         }
 
-        let mut names: Vec<Vec<u8>> = vec![];
-        let mut kinds: Vec<Vec<u8>> = vec![];
-        let mut shared_owns: Vec<Vec<u8>> = vec![];
+        let mut names: Vec<String> = vec![];
+        let mut kinds: Vec<String> = vec![];
+        let mut shared_owns: Vec<String> = vec![];
         for entry in resp.objects.iter() {
             match &entry.object {
                 ShareGrantObjectName::Database(db) => {
-                    kinds.push("DATABASE".to_string().as_bytes().to_vec());
-                    names.push(db.clone().as_bytes().to_vec());
+                    kinds.push("DATABASE".to_string());
+                    names.push(db.clone());
                 }
                 ShareGrantObjectName::Table(db, table_name) => {
-                    kinds.push("TABLE".to_string().as_bytes().to_vec());
-                    names.push(format!("{}.{}", db, table_name).as_bytes().to_vec());
+                    kinds.push("TABLE".to_string());
+                    names.push(format!("{}.{}", db, table_name));
                 }
             }
-            shared_owns.push(entry.grant_on.to_string().as_bytes().to_vec());
+            shared_owns.push(entry.grant_on.to_string());
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_share_endpoint_show.rs b/src/query/service/src/interpreters/interpreter_share_endpoint_show.rs
index 1664feb9de28..25bbbcb18f46 100644
--- a/src/query/service/src/interpreters/interpreter_share_endpoint_show.rs
+++ b/src/query/service/src/interpreters/interpreter_share_endpoint_show.rs
@@ -49,19 +49,19 @@ impl Interpreter for ShowShareEndpointInterpreter {
             .get_share_endpoint(self.plan.clone().into())
             .await?;
 
-        let mut endpoints: Vec<Vec<u8>> = vec![];
-        let mut urls: Vec<Vec<u8>> = vec![];
-        let mut to_tenants: Vec<Vec<u8>> = vec![];
-        let mut args: Vec<Vec<u8>> = vec![];
-        let mut comments: Vec<Vec<u8>> = vec![];
-        let mut created_on_vec: Vec<Vec<u8>> = vec![];
+        let mut endpoints: Vec<String> = vec![];
+        let mut urls: Vec<String> = vec![];
+        let mut to_tenants: Vec<String> = vec![];
+        let mut args: Vec<String> = vec![];
+        let mut comments: Vec<String> = vec![];
+        let mut created_on_vec: Vec<String> = vec![];
         for (endpoint, meta) in resp.share_endpoint_meta_vec {
-            endpoints.push(endpoint.endpoint.clone().as_bytes().to_vec());
-            urls.push(meta.url.clone().as_bytes().to_vec());
-            to_tenants.push(meta.tenant.clone().as_bytes().to_vec());
-            args.push(format!("{:?}", meta.args).as_bytes().to_vec());
-            comments.push(meta.comment.unwrap_or_default().as_bytes().to_vec());
-            created_on_vec.push(meta.create_on.to_string().as_bytes().to_vec());
+            endpoints.push(endpoint.endpoint.clone());
+            urls.push(meta.url.clone());
+            to_tenants.push(meta.tenant.clone());
+            args.push(format!("{:?}", meta.args));
+            comments.push(meta.comment.unwrap_or_default());
+            created_on_vec.push(meta.create_on.to_string());
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_share_show.rs b/src/query/service/src/interpreters/interpreter_share_show.rs
index 18f18c59a039..91bea064e650 100644
--- a/src/query/service/src/interpreters/interpreter_share_show.rs
+++ b/src/query/service/src/interpreters/interpreter_share_show.rs
@@ -48,40 +48,26 @@ impl Interpreter for ShowSharesInterpreter {
     async fn execute2(&self) -> Result<PipelineBuildResult> {
         let meta_api = UserApiProvider::instance().get_meta_store_client();
         let tenant = self.ctx.get_tenant();
-        let mut names: Vec<Vec<u8>> = vec![];
-        let mut kinds: Vec<Vec<u8>> = vec![];
-        let mut created_owns: Vec<Vec<u8>> = vec![];
-        let mut database_names: Vec<Vec<u8>> = vec![];
-        let mut from: Vec<Vec<u8>> = vec![];
-        let mut to: Vec<Vec<u8>> = vec![];
-        let mut comments: Vec<Vec<u8>> = vec![];
+        let mut names: Vec<String> = vec![];
+        let mut kinds: Vec<String> = vec![];
+        let mut created_owns: Vec<String> = vec![];
+        let mut database_names: Vec<String> = vec![];
+        let mut from: Vec<String> = vec![];
+        let mut to: Vec<String> = vec![];
+        let mut comments: Vec<String> = vec![];
 
         // query all share endpoint for other tenant inbound shares
         let share_specs = ShareEndpointManager::instance()
             .get_inbound_shares(&tenant, None, None)
             .await?;
         for (from_tenant, share_spec) in share_specs {
-            names.push(share_spec.name.clone().as_bytes().to_vec());
-            kinds.push("INBOUND".to_string().as_bytes().to_vec());
-            created_owns.push(
-                share_spec
-                    .share_on
-                    .unwrap_or_default()
-                    .to_string()
-                    .as_bytes()
-                    .to_vec(),
-            );
-            database_names.push(
-                share_spec
-                    .database
-                    .unwrap_or_default()
-                    .name
-                    .as_bytes()
-                    .to_vec(),
-            );
-            from.push(from_tenant.as_bytes().to_vec());
-            to.push(tenant.clone().as_bytes().to_vec());
-            comments.push(share_spec.comment.unwrap_or_default().as_bytes().to_vec());
+            names.push(share_spec.name.clone());
+            kinds.push("INBOUND".to_string());
+            created_owns.push(share_spec.share_on.unwrap_or_default().to_string());
+            database_names.push(share_spec.database.unwrap_or_default().name);
+            from.push(from_tenant);
+            to.push(tenant.clone());
+            comments.push(share_spec.comment.unwrap_or_default());
         }
 
         let req = ShowSharesReq {
@@ -90,19 +76,17 @@ impl Interpreter for ShowSharesInterpreter {
         let resp = meta_api.show_shares(req).await?;
 
         for entry in resp.outbound_accounts {
-            names.push(entry.share_name.share_name.clone().as_bytes().to_vec());
-            kinds.push("OUTBOUND".to_string().as_bytes().to_vec());
-            created_owns.push(entry.create_on.to_string().as_bytes().to_vec());
-            database_names.push(entry.database_name.unwrap_or_default().as_bytes().to_vec());
-            from.push(entry.share_name.tenant.clone().as_bytes().to_vec());
+            names.push(entry.share_name.share_name.clone());
+            kinds.push("OUTBOUND".to_string());
+            created_owns.push(entry.create_on.to_string());
+            database_names.push(entry.database_name.unwrap_or_default());
+            from.push(entry.share_name.tenant.clone());
             to.push(
                 entry
                     .accounts
-                    .map_or("".to_string().as_bytes().to_vec(), |accounts| {
-                        accounts.join(",").as_bytes().to_vec()
-                    }),
+                    .map_or("".to_string(), |accounts| accounts.join(",")),
             );
-            comments.push(entry.comment.unwrap_or_default().as_bytes().to_vec());
+            comments.push(entry.comment.unwrap_or_default());
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_share_show_grant_tenants.rs b/src/query/service/src/interpreters/interpreter_share_show_grant_tenants.rs
index dc585d3cb3dc..d3c5b979ee3f 100644
--- a/src/query/service/src/interpreters/interpreter_share_show_grant_tenants.rs
+++ b/src/query/service/src/interpreters/interpreter_share_show_grant_tenants.rs
@@ -61,12 +61,12 @@ impl Interpreter for ShowGrantTenantsOfShareInterpreter {
             return Ok(PipelineBuildResult::create());
         }
 
-        let mut granted_owns: Vec<Vec<u8>> = vec![];
-        let mut accounts: Vec<Vec<u8>> = vec![];
+        let mut granted_owns: Vec<String> = vec![];
+        let mut accounts: Vec<String> = vec![];
 
         for account in resp.accounts {
-            granted_owns.push(account.grant_on.to_string().as_bytes().to_vec());
-            accounts.push(account.account.clone().as_bytes().to_vec());
+            granted_owns.push(account.grant_on.to_string());
+            accounts.push(account.account.clone());
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_show_grants.rs b/src/query/service/src/interpreters/interpreter_show_grants.rs
index 5f05122010e6..f3800577f432 100644
--- a/src/query/service/src/interpreters/interpreter_show_grants.rs
+++ b/src/query/service/src/interpreters/interpreter_show_grants.rs
@@ -81,7 +81,7 @@ impl Interpreter for ShowGrantsInterpreter {
             .fold(grant_set, |a, b| a | b)
             .entries();
 
-        let mut grant_list: Vec<Vec<u8>> = Vec::new();
+        let mut grant_list: Vec<String> = Vec::new();
         for grant_entry in grant_entries {
             let object = grant_entry.object();
             match object {
@@ -95,13 +95,10 @@ impl Interpreter for ShowGrantsInterpreter {
                     let catalog = self.ctx.get_catalog(catalog_name).await?;
                     let db_name = catalog.get_db_name_by_id(*db_id).await?;
                     let table_name = catalog.get_table_name_by_id(*table_id).await?;
-                    grant_list.push(
-                        format!(
-                            "GRANT {} ON '{}'.'{}'.'{}' TO {}",
-                            &privileges_str, catalog_name, db_name, table_name, identity
-                        )
-                        .into(),
-                    );
+                    grant_list.push(format!(
+                        "GRANT {} ON '{}'.'{}'.'{}' TO {}",
+                        &privileges_str, catalog_name, db_name, table_name, identity
+                    ));
                 }
                 GrantObject::DatabaseById(catalog_name, db_id) => {
                     let privileges_str = if grant_entry.has_all_available_privileges() {
@@ -112,21 +109,13 @@ impl Interpreter for ShowGrantsInterpreter {
                     };
                     let catalog = self.ctx.get_catalog(catalog_name).await?;
                     let db_name = catalog.get_db_name_by_id(*db_id).await?;
-                    grant_list.push(
-                        format!(
-                            "GRANT {} ON '{}'.'{}'.* TO {}",
-                            &privileges_str, catalog_name, db_name, identity
-                        )
-                        .as_bytes()
-                        .to_vec(),
-                    );
+                    grant_list.push(format!(
+                        "GRANT {} ON '{}'.'{}'.* TO {}",
+                        &privileges_str, catalog_name, db_name, identity
+                    ));
                 }
                 _ => {
-                    grant_list.push(
-                        format!("{} TO {}", grant_entry, identity)
-                            .as_bytes()
-                            .to_vec(),
-                    );
+                    grant_list.push(format!("{} TO {}", grant_entry, identity));
                 }
             }
         }
diff --git a/src/query/service/src/interpreters/interpreter_show_object_grant_privileges.rs b/src/query/service/src/interpreters/interpreter_show_object_grant_privileges.rs
index 403a6a503be5..4fb3e6215a28 100644
--- a/src/query/service/src/interpreters/interpreter_show_object_grant_privileges.rs
+++ b/src/query/service/src/interpreters/interpreter_show_object_grant_privileges.rs
@@ -56,14 +56,14 @@ impl Interpreter for ShowObjectGrantPrivilegesInterpreter {
         if resp.privileges.is_empty() {
             return Ok(PipelineBuildResult::create());
         }
-        let mut share_names: Vec<Vec<u8>> = vec![];
-        let mut privileges: Vec<Vec<u8>> = vec![];
-        let mut created_owns: Vec<Vec<u8>> = vec![];
+        let mut share_names: Vec<String> = vec![];
+        let mut privileges: Vec<String> = vec![];
+        let mut created_owns: Vec<String> = vec![];
 
         for privilege in resp.privileges {
-            share_names.push(privilege.share_name.as_bytes().to_vec());
-            privileges.push(privilege.privileges.to_string().as_bytes().to_vec());
-            created_owns.push(privilege.grant_on.to_string().as_bytes().to_vec());
+            share_names.push(privilege.share_name);
+            privileges.push(privilege.privileges.to_string());
+            created_owns.push(privilege.grant_on.to_string());
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_table_describe.rs b/src/query/service/src/interpreters/interpreter_table_describe.rs
index 1efd4285b148..71a01056f3be 100644
--- a/src/query/service/src/interpreters/interpreter_table_describe.rs
+++ b/src/query/service/src/interpreters/interpreter_table_describe.rs
@@ -71,30 +71,30 @@ impl Interpreter for DescribeTableInterpreter {
             Ok(table.schema())
         }?;
 
-        let mut names: Vec<Vec<u8>> = vec![];
-        let mut types: Vec<Vec<u8>> = vec![];
-        let mut nulls: Vec<Vec<u8>> = vec![];
-        let mut default_exprs: Vec<Vec<u8>> = vec![];
-        let mut extras: Vec<Vec<u8>> = vec![];
+        let mut names: Vec<String> = vec![];
+        let mut types: Vec<String> = vec![];
+        let mut nulls: Vec<String> = vec![];
+        let mut default_exprs: Vec<String> = vec![];
+        let mut extras: Vec<String> = vec![];
 
         for field in schema.fields().iter() {
-            names.push(field.name().to_string().as_bytes().to_vec());
+            names.push(field.name().to_string());
 
             let non_null_type = field.data_type().remove_recursive_nullable();
-            types.push(non_null_type.sql_name().as_bytes().to_vec());
+            types.push(non_null_type.sql_name());
             nulls.push(if field.is_nullable() {
-                "YES".to_string().as_bytes().to_vec()
+                "YES".to_string()
             } else {
-                "NO".to_string().as_bytes().to_vec()
+                "NO".to_string()
             });
             match field.default_expr() {
                 Some(expr) => {
-                    default_exprs.push(expr.as_bytes().to_vec());
+                    default_exprs.push(expr.clone());
                 }
 
                 None => {
                     let value = Scalar::default_value(&field.data_type().into());
-                    default_exprs.push(value.to_string().as_bytes().to_vec());
+                    default_exprs.push(value.to_string());
                 }
             }
             let extra = match field.computed_expr() {
@@ -102,7 +102,7 @@ impl Interpreter for DescribeTableInterpreter {
                 Some(ComputedExpr::Stored(expr)) => format!("STORED COMPUTED COLUMN `{}`", expr),
                 _ => "".to_string(),
             };
-            extras.push(extra.as_bytes().to_vec());
+            extras.push(extra);
         }
 
         PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/interpreters/interpreter_table_modify_column.rs b/src/query/service/src/interpreters/interpreter_table_modify_column.rs
index e9ee907e213b..06fd45622bfc 100644
--- a/src/query/service/src/interpreters/interpreter_table_modify_column.rs
+++ b/src/query/service/src/interpreters/interpreter_table_modify_column.rs
@@ -428,7 +428,7 @@ impl ModifyTableColumnInterpreter {
                 cast_needed: true,
             }));
         let mut build_res =
-            build_query_pipeline_without_render_result_set(&self.ctx, &insert_plan, false).await?;
+            build_query_pipeline_without_render_result_set(&self.ctx, &insert_plan).await?;
 
         // 6. commit new meta schema and snapshots
         new_table.commit_insertion(
diff --git a/src/query/service/src/interpreters/interpreter_table_optimize.rs b/src/query/service/src/interpreters/interpreter_table_optimize.rs
index 687c951b7299..1748769ffc16 100644
--- a/src/query/service/src/interpreters/interpreter_table_optimize.rs
+++ b/src/query/service/src/interpreters/interpreter_table_optimize.rs
@@ -188,8 +188,7 @@ impl OptimizeTableInterpreter {
             )?;
 
             let build_res =
-                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                    .await?;
+                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
             build_res.main_pipeline
         } else {
             Pipeline::create()
@@ -237,12 +236,9 @@ impl OptimizeTableInterpreter {
                         mutator.removed_segment_summary,
                     )?;
 
-                    build_res = build_query_pipeline_without_render_result_set(
-                        &self.ctx,
-                        &physical_plan,
-                        false,
-                    )
-                    .await?;
+                    build_res =
+                        build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan)
+                            .await?;
 
                     let ctx = self.ctx.clone();
                     let plan = self.plan.clone();
diff --git a/src/query/service/src/interpreters/interpreter_table_recluster.rs b/src/query/service/src/interpreters/interpreter_table_recluster.rs
index c7dc851a710e..accbc8171839 100644
--- a/src/query/service/src/interpreters/interpreter_table_recluster.rs
+++ b/src/query/service/src/interpreters/interpreter_table_recluster.rs
@@ -166,8 +166,7 @@ impl Interpreter for ReclusterTableInterpreter {
             )?;
 
             let mut build_res =
-                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                    .await?;
+                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
             assert!(build_res.main_pipeline.is_complete_pipeline()?);
             build_res.set_max_threads(max_threads);
 
diff --git a/src/query/service/src/interpreters/interpreter_table_show_create.rs b/src/query/service/src/interpreters/interpreter_table_show_create.rs
index 908d08282132..fde21dd3cf26 100644
--- a/src/query/service/src/interpreters/interpreter_table_show_create.rs
+++ b/src/query/service/src/interpreters/interpreter_table_show_create.rs
@@ -179,11 +179,11 @@ impl ShowCreateTableInterpreter {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(name.as_bytes().to_vec())),
+                    Value::Scalar(Scalar::String(name.to_string())),
                 ),
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(table_create_sql.into_bytes())),
+                    Value::Scalar(Scalar::String(table_create_sql)),
                 ),
             ],
             1,
@@ -204,11 +204,11 @@ impl ShowCreateTableInterpreter {
                 vec![
                     BlockEntry::new(
                         DataType::String,
-                        Value::Scalar(Scalar::String(name.as_bytes().to_vec())),
+                        Value::Scalar(Scalar::String(name.to_string())),
                     ),
                     BlockEntry::new(
                         DataType::String,
-                        Value::Scalar(Scalar::String(view_create_sql.into_bytes())),
+                        Value::Scalar(Scalar::String(view_create_sql)),
                     ),
                 ],
                 1,
@@ -240,12 +240,9 @@ impl ShowCreateTableInterpreter {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(stream_table.name().as_bytes().to_vec())),
-                ),
-                BlockEntry::new(
-                    DataType::String,
-                    Value::Scalar(Scalar::String(create_sql.into_bytes())),
+                    Value::Scalar(Scalar::String(stream_table.name().to_string())),
                 ),
+                BlockEntry::new(DataType::String, Value::Scalar(Scalar::String(create_sql))),
             ],
             1,
         );
@@ -277,12 +274,9 @@ impl ShowCreateTableInterpreter {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(name.as_bytes().to_vec())),
-                ),
-                BlockEntry::new(
-                    DataType::String,
-                    Value::Scalar(Scalar::String(ddl.into_bytes())),
+                    Value::Scalar(Scalar::String(name.to_string())),
                 ),
+                BlockEntry::new(DataType::String, Value::Scalar(Scalar::String(ddl))),
             ],
             1,
         );
diff --git a/src/query/service/src/interpreters/interpreter_table_vacuum.rs b/src/query/service/src/interpreters/interpreter_table_vacuum.rs
index 243739fb08b6..b409586dd1f8 100644
--- a/src/query/service/src/interpreters/interpreter_table_vacuum.rs
+++ b/src/query/service/src/interpreters/interpreter_table_vacuum.rs
@@ -86,16 +86,9 @@ impl Interpreter for VacuumTableInterpreter {
 
         match purge_files_opt {
             None => return Ok(PipelineBuildResult::create()),
-            Some(purge_files) => {
-                let mut files: Vec<Vec<u8>> = Vec::with_capacity(purge_files.len());
-                for file in purge_files.into_iter() {
-                    files.push(file.as_bytes().to_vec());
-                }
-
-                PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
-                    StringType::from_data(files),
-                ])])
-            }
+            Some(purge_files) => PipelineBuildResult::from_blocks(vec![
+                DataBlock::new_from_columns(vec![StringType::from_data(purge_files)]),
+            ]),
         }
     }
 }
diff --git a/src/query/service/src/interpreters/interpreter_task_alter.rs b/src/query/service/src/interpreters/interpreter_task_alter.rs
index ccdaf4be222f..8104c9661230 100644
--- a/src/query/service/src/interpreters/interpreter_task_alter.rs
+++ b/src/query/service/src/interpreters/interpreter_task_alter.rs
@@ -68,6 +68,8 @@ impl AlterTaskInterpreter {
             when_condition: None,
             add_after: vec![],
             remove_after: vec![],
+            set_session_parameters: false,
+            session_parameters: Default::default(),
         };
         match plan.alter_options {
             AlterTaskOptions::Resume => {
@@ -81,6 +83,7 @@ impl AlterTaskInterpreter {
                 comments,
                 warehouse,
                 suspend_task_after_num_failures,
+                session_parameters,
             } => {
                 req.alter_task_type = AlterTaskType::Set as i32;
                 req.schedule_options = schedule.map(make_schedule_options);
@@ -91,6 +94,10 @@ impl AlterTaskInterpreter {
                 });
                 req.suspend_task_after_num_failures =
                     suspend_task_after_num_failures.map(|i| i as i32);
+                if let Some(session_parameters) = session_parameters {
+                    req.set_session_parameters = true;
+                    req.session_parameters = session_parameters;
+                }
             }
             AlterTaskOptions::Unset { .. } => {
                 todo!()
diff --git a/src/query/service/src/interpreters/interpreter_task_create.rs b/src/query/service/src/interpreters/interpreter_task_create.rs
index 93ec5a276430..603274378485 100644
--- a/src/query/service/src/interpreters/interpreter_task_create.rs
+++ b/src/query/service/src/interpreters/interpreter_task_create.rs
@@ -63,6 +63,7 @@ impl CreateTaskInterpreter {
             if_not_exist: plan.if_not_exists,
             after: plan.after,
             when_condition: plan.when_condition,
+            session_parameters: plan.session_parameters,
         }
     }
 }
diff --git a/src/query/service/src/interpreters/interpreter_update.rs b/src/query/service/src/interpreters/interpreter_update.rs
index a8ef7f8f1682..11cdacd158f2 100644
--- a/src/query/service/src/interpreters/interpreter_update.rs
+++ b/src/query/service/src/interpreters/interpreter_update.rs
@@ -108,8 +108,7 @@ impl Interpreter for UpdateInterpreter {
         let mut build_res = PipelineBuildResult::create();
         if let Some(physical_plan) = physical_plan {
             build_res =
-                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan, false)
-                    .await?;
+                build_query_pipeline_without_render_result_set(&self.ctx, &physical_plan).await?;
             {
                 let hook_operator = HookOperator::create(
                     self.ctx.clone(),
diff --git a/src/query/service/src/interpreters/interpreter_vacuum_drop_tables.rs b/src/query/service/src/interpreters/interpreter_vacuum_drop_tables.rs
index 91e192f1945c..526a0fad20db 100644
--- a/src/query/service/src/interpreters/interpreter_vacuum_drop_tables.rs
+++ b/src/query/service/src/interpreters/interpreter_vacuum_drop_tables.rs
@@ -138,12 +138,12 @@ impl Interpreter for VacuumDropTablesInterpreter {
                 if let Some(limit) = self.plan.option.limit {
                     len = min(len, limit);
                 }
-                let mut tables: Vec<Vec<u8>> = Vec::with_capacity(len);
-                let mut files: Vec<Vec<u8>> = Vec::with_capacity(len);
+                let mut tables: Vec<String> = Vec::with_capacity(len);
+                let mut files: Vec<String> = Vec::with_capacity(len);
                 let purge_files = &purge_files[0..len];
                 for file in purge_files.iter() {
-                    tables.push(file.0.to_string().as_bytes().to_vec());
-                    files.push(file.1.to_string().as_bytes().to_vec());
+                    tables.push(file.0.to_string());
+                    files.push(file.1.to_string());
                 }
 
                 PipelineBuildResult::from_blocks(vec![DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/lib.rs b/src/query/service/src/lib.rs
index b9152a509b7b..64fdf4828f08 100644
--- a/src/query/service/src/lib.rs
+++ b/src/query/service/src/lib.rs
@@ -35,6 +35,7 @@
 #![feature(let_chains)]
 #![feature(try_blocks)]
 #![feature(lazy_cell)]
+#![feature(variant_count)]
 #![allow(clippy::diverging_sub_expression)]
 #![allow(clippy::arc_with_non_send_sync)]
 
diff --git a/src/query/service/src/local/executor.rs b/src/query/service/src/local/executor.rs
index fa3d986bc0d1..cc38ac6e5375 100644
--- a/src/query/service/src/local/executor.rs
+++ b/src/query/service/src/local/executor.rs
@@ -117,7 +117,7 @@ impl SessionExecutor {
                             .convert_to_full_column(&DataType::String, num_rows);
                         let value = StringType::try_downcast_column(&col).unwrap();
                         for r in value.iter() {
-                            keywords.push(unsafe { String::from_utf8_unchecked(r.to_vec()) });
+                            keywords.push(r.to_string());
                         }
                     }
                 }
diff --git a/src/query/service/src/pipelines/builders/builder_aggregate.rs b/src/query/service/src/pipelines/builders/builder_aggregate.rs
index d67fc5fff89e..d79e905ea637 100644
--- a/src/query/service/src/pipelines/builders/builder_aggregate.rs
+++ b/src/query/service/src/pipelines/builders/builder_aggregate.rs
@@ -25,9 +25,6 @@ use databend_common_expression::HashMethodKind;
 use databend_common_functions::aggregates::AggregateFunctionFactory;
 use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_core::query_spill_prefix;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
-use databend_common_pipeline_transforms::processors::ProfileStub;
-use databend_common_pipeline_transforms::processors::Transformer;
 use databend_common_sql::executor::physical_plans::AggregateExpand;
 use databend_common_sql::executor::physical_plans::AggregateFinal;
 use databend_common_sql::executor::physical_plans::AggregateFunctionDesc;
@@ -110,17 +107,9 @@ impl PipelineBuilder {
 
         if params.group_columns.is_empty() {
             return self.main_pipeline.add_transform(|input, output| {
-                let transform = PartialSingleStateAggregator::try_create(input, output, &params)?;
-
-                if self.enable_profiling {
-                    Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                        transform,
-                        aggregate.plan_id,
-                        self.proc_profs.clone(),
-                    )))
-                } else {
-                    Ok(ProcessorPtr::create(transform))
-                }
+                Ok(ProcessorPtr::create(
+                    PartialSingleStateAggregator::try_create(input, output, &params)?,
+                ))
             });
         }
 
@@ -132,76 +121,60 @@ impl PipelineBuilder {
         let method = DataBlock::choose_hash_method(&sample_block, group_cols, efficiently_memory)?;
 
         self.main_pipeline.add_transform(|input, output| {
-            let transform = match params.aggregate_functions.is_empty() {
-                true => with_mappedhash_method!(|T| match method.clone() {
-                    HashMethodKind::T(method) => TransformPartialGroupBy::try_create(
-                        self.ctx.clone(),
-                        method,
-                        input,
-                        output,
-                        params.clone()
-                    ),
-                }),
-                false => with_mappedhash_method!(|T| match method.clone() {
-                    HashMethodKind::T(method) => TransformPartialAggregate::try_create(
-                        self.ctx.clone(),
-                        method,
-                        input,
-                        output,
-                        params.clone()
-                    ),
-                }),
-            }?;
-
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    aggregate.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
-        })?;
-
-        // If cluster mode, spill write will be completed in exchange serialize, because we need scatter the block data first
-        if self.ctx.get_cluster().is_empty() {
-            let operator = DataOperator::instance().operator();
-            let location_prefix = query_spill_prefix(&self.ctx.get_tenant());
-            self.main_pipeline.add_transform(|input, output| {
-                let transform = match params.aggregate_functions.is_empty() {
+            Ok(ProcessorPtr::create(
+                match params.aggregate_functions.is_empty() {
                     true => with_mappedhash_method!(|T| match method.clone() {
-                        HashMethodKind::T(method) => TransformGroupBySpillWriter::create(
+                        HashMethodKind::T(method) => TransformPartialGroupBy::try_create(
                             self.ctx.clone(),
+                            method,
                             input,
                             output,
-                            method,
-                            operator.clone(),
-                            location_prefix.clone()
+                            params.clone()
                         ),
                     }),
                     false => with_mappedhash_method!(|T| match method.clone() {
-                        HashMethodKind::T(method) => TransformAggregateSpillWriter::create(
+                        HashMethodKind::T(method) => TransformPartialAggregate::try_create(
                             self.ctx.clone(),
+                            method,
                             input,
                             output,
-                            method,
-                            operator.clone(),
-                            params.clone(),
-                            location_prefix.clone()
+                            params.clone()
                         ),
                     }),
-                };
+                }?,
+            ))
+        })?;
 
-                if self.enable_profiling {
-                    Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                        transform,
-                        aggregate.plan_id,
-                        self.proc_profs.clone(),
-                    )))
-                } else {
-                    Ok(ProcessorPtr::create(transform))
-                }
+        // If cluster mode, spill write will be completed in exchange serialize, because we need scatter the block data first
+        if self.ctx.get_cluster().is_empty() {
+            let operator = DataOperator::instance().operator();
+            let location_prefix = query_spill_prefix(&self.ctx.get_tenant());
+            self.main_pipeline.add_transform(|input, output| {
+                Ok(ProcessorPtr::create(
+                    match params.aggregate_functions.is_empty() {
+                        true => with_mappedhash_method!(|T| match method.clone() {
+                            HashMethodKind::T(method) => TransformGroupBySpillWriter::create(
+                                self.ctx.clone(),
+                                input,
+                                output,
+                                method,
+                                operator.clone(),
+                                location_prefix.clone()
+                            ),
+                        }),
+                        false => with_mappedhash_method!(|T| match method.clone() {
+                            HashMethodKind::T(method) => TransformAggregateSpillWriter::create(
+                                self.ctx.clone(),
+                                input,
+                                output,
+                                method,
+                                operator.clone(),
+                                params.clone(),
+                                location_prefix.clone()
+                            ),
+                        }),
+                    },
+                ))
             })?;
         }
 
@@ -231,32 +204,11 @@ impl PipelineBuilder {
             self.build_pipeline(&aggregate.input)?;
             self.main_pipeline.try_resize(1)?;
             self.main_pipeline.add_transform(|input, output| {
-                let transform = FinalSingleStateAggregator::try_create(input, output, &params)?;
-
-                if self.enable_profiling {
-                    Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                        transform,
-                        aggregate.plan_id,
-                        self.proc_profs.clone(),
-                    )))
-                } else {
-                    Ok(ProcessorPtr::create(transform))
-                }
+                Ok(ProcessorPtr::create(
+                    FinalSingleStateAggregator::try_create(input, output, &params)?,
+                ))
             })?;
 
-            // Append a profile stub to record the output rows and bytes
-            if self.enable_profiling {
-                self.main_pipeline.add_transform(|input, output| {
-                    Ok(ProcessorPtr::create(Transformer::create(
-                        input,
-                        output,
-                        ProfileStub::new(aggregate.plan_id, self.proc_profs.clone())
-                            .accumulate_output_rows()
-                            .accumulate_output_bytes(),
-                    )))
-                })?;
-            }
-
             return Ok(());
         }
 
@@ -283,14 +235,7 @@ impl PipelineBuilder {
 
                     self.build_pipeline(&aggregate.input)?;
                     self.exchange_injector = old_inject;
-                    build_partition_bucket::<_, ()>(
-                        v,
-                        &mut self.main_pipeline,
-                        params.clone(),
-                        self.enable_profiling,
-                        aggregate.plan_id,
-                        self.proc_profs.clone(),
-                    )
+                    build_partition_bucket::<_, ()>(v, &mut self.main_pipeline, params.clone())
                 }
             }),
             false => with_hash_method!(|T| match method {
@@ -305,14 +250,7 @@ impl PipelineBuilder {
                     }
                     self.build_pipeline(&aggregate.input)?;
                     self.exchange_injector = old_inject;
-                    build_partition_bucket::<_, usize>(
-                        v,
-                        &mut self.main_pipeline,
-                        params.clone(),
-                        self.enable_profiling,
-                        aggregate.plan_id,
-                        self.proc_profs.clone(),
-                    )
+                    build_partition_bucket::<_, usize>(v, &mut self.main_pipeline, params.clone())
                 }
             }),
         }
diff --git a/src/query/service/src/pipelines/builders/builder_exchange.rs b/src/query/service/src/pipelines/builders/builder_exchange.rs
index 66d13b358d6a..25b93c469383 100644
--- a/src/query/service/src/pipelines/builders/builder_exchange.rs
+++ b/src/query/service/src/pipelines/builders/builder_exchange.rs
@@ -24,7 +24,6 @@ impl PipelineBuilder {
         let build_res = exchange_manager.get_fragment_source(
             &exchange_source.query_id,
             exchange_source.source_fragment_id,
-            self.enable_profiling,
             self.exchange_injector.clone(),
         )?;
         // add sharing data
diff --git a/src/query/service/src/pipelines/builders/builder_filter.rs b/src/query/service/src/pipelines/builders/builder_filter.rs
index 3768544d3c1b..e70fc1deaa39 100644
--- a/src/query/service/src/pipelines/builders/builder_filter.rs
+++ b/src/query/service/src/pipelines/builders/builder_filter.rs
@@ -19,7 +19,6 @@ use databend_common_expression::type_check::check_function;
 use databend_common_expression::types::DataType;
 use databend_common_functions::BUILTIN_FUNCTIONS;
 use databend_common_pipeline_core::processors::ProcessorPtr;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
 use databend_common_sql::executor::physical_plans::Filter;
 
 use crate::pipelines::processors::transforms::TransformFilter;
@@ -47,7 +46,7 @@ impl PipelineBuilder {
         let max_block_size = self.settings.get_max_block_size()? as usize;
         let (select_expr, has_or) = build_select_expr(&predicate).into();
         self.main_pipeline.add_transform(|input, output| {
-            let transform = TransformFilter::create(
+            Ok(ProcessorPtr::create(TransformFilter::create(
                 input,
                 output,
                 select_expr.clone(),
@@ -55,17 +54,7 @@ impl PipelineBuilder {
                 filter.projections.clone(),
                 self.func_ctx.clone(),
                 max_block_size,
-            );
-
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    filter.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            )))
         })?;
 
         Ok(())
diff --git a/src/query/service/src/pipelines/builders/builder_join.rs b/src/query/service/src/pipelines/builders/builder_join.rs
index 3db9373560dc..77ce87887b63 100644
--- a/src/query/service/src/pipelines/builders/builder_join.rs
+++ b/src/query/service/src/pipelines/builders/builder_join.rs
@@ -18,9 +18,6 @@ use databend_common_base::base::tokio::sync::Barrier;
 use databend_common_exception::Result;
 use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_sinks::Sinker;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
-use databend_common_pipeline_transforms::processors::ProfileStub;
-use databend_common_pipeline_transforms::processors::Transformer;
 use databend_common_sql::executor::physical_plans::HashJoin;
 use databend_common_sql::executor::physical_plans::MaterializedCte;
 use databend_common_sql::executor::physical_plans::RangeJoin;
@@ -50,17 +47,6 @@ impl PipelineBuilder {
         let state = Arc::new(RangeJoinState::new(self.ctx.clone(), range_join));
         self.expand_right_side_pipeline(range_join, state.clone())?;
         self.build_left_side(range_join, state)?;
-        if self.enable_profiling {
-            self.main_pipeline.add_transform(|input, output| {
-                Ok(ProcessorPtr::create(Transformer::create(
-                    input,
-                    output,
-                    ProfileStub::new(range_join.plan_id, self.proc_profs.clone())
-                        .accumulate_output_rows()
-                        .accumulate_output_bytes(),
-                )))
-            })?;
-        }
         Ok(())
     }
 
@@ -73,16 +59,11 @@ impl PipelineBuilder {
         let max_threads = self.settings.get_max_threads()? as usize;
         self.main_pipeline.try_resize(max_threads)?;
         self.main_pipeline.add_transform(|input, output| {
-            let transform = TransformRangeJoinLeft::create(input, output, state.clone());
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    range_join.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            Ok(ProcessorPtr::create(TransformRangeJoinLeft::create(
+                input,
+                output,
+                state.clone(),
+            )))
         })?;
         Ok(())
     }
@@ -97,26 +78,17 @@ impl PipelineBuilder {
             self.func_ctx.clone(),
             self.settings.clone(),
             right_side_context,
-            self.enable_profiling,
-            self.proc_profs.clone(),
             self.main_pipeline.get_scopes(),
         );
         right_side_builder.cte_state = self.cte_state.clone();
         let mut right_res = right_side_builder.finalize(&range_join.right)?;
         right_res.main_pipeline.add_sink(|input| {
-            let transform = Sinker::<TransformRangeJoinRight>::create(
-                input,
-                TransformRangeJoinRight::create(state.clone()),
-            );
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    range_join.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            Ok(ProcessorPtr::create(
+                Sinker::<TransformRangeJoinRight>::create(
+                    input,
+                    TransformRangeJoinRight::create(state.clone()),
+                ),
+            ))
         })?;
         self.pipelines.push(right_res.main_pipeline.finalize());
         self.pipelines.extend(right_res.sources_pipelines);
@@ -124,14 +96,12 @@ impl PipelineBuilder {
     }
 
     pub(crate) fn build_join(&mut self, join: &HashJoin) -> Result<()> {
-        let id = join.probe.get_table_index();
         // for merge into target table as build side.
         let (merge_into_build_table_index, merge_into_is_distributed) =
             self.merge_into_get_optimization_flag(join);
 
         let state = self.build_join_state(
             join,
-            id,
             merge_into_build_table_index,
             merge_into_is_distributed,
         )?;
@@ -142,7 +112,6 @@ impl PipelineBuilder {
     fn build_join_state(
         &mut self,
         join: &HashJoin,
-        id: IndexType,
         merge_into_target_table_index: IndexType,
         merge_into_is_distributed: bool,
     ) -> Result<Arc<HashJoinState>> {
@@ -152,7 +121,6 @@ impl PipelineBuilder {
             &join.build_projections,
             HashJoinDesc::create(join)?,
             &join.probe_to_build,
-            id,
             merge_into_target_table_index,
             merge_into_is_distributed,
         )
@@ -169,8 +137,6 @@ impl PipelineBuilder {
             self.func_ctx.clone(),
             self.settings.clone(),
             build_side_context,
-            self.enable_profiling,
-            self.proc_profs.clone(),
             self.main_pipeline.get_scopes(),
         );
         build_side_builder.cte_state = self.cte_state.clone();
@@ -201,18 +167,12 @@ impl PipelineBuilder {
             } else {
                 None
             };
-            let transform =
-                TransformHashJoinBuild::try_create(input, build_state.clone(), spill_state)?;
 
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    hash_join_plan.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            Ok(ProcessorPtr::create(TransformHashJoinBuild::try_create(
+                input,
+                build_state.clone(),
+                spill_state,
+            )?))
         };
         // for distributed merge into when source as build side.
         if hash_join_plan.need_hold_hash_table {
@@ -257,7 +217,8 @@ impl PipelineBuilder {
             } else {
                 None
             };
-            let transform = TransformHashJoinProbe::create(
+
+            Ok(ProcessorPtr::create(TransformHashJoinProbe::create(
                 input,
                 output,
                 join.projections.clone(),
@@ -268,32 +229,9 @@ impl PipelineBuilder {
                 &join.join_type,
                 !join.non_equi_conditions.is_empty(),
                 has_string_column,
-            )?;
-
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    join.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            )?))
         })?;
 
-        if self.enable_profiling {
-            // Add a stub after the probe processor to accumulate the output rows.
-            self.main_pipeline.add_transform(|input, output| {
-                Ok(ProcessorPtr::create(Transformer::create(
-                    input,
-                    output,
-                    ProfileStub::new(join.plan_id, self.proc_profs.clone())
-                        .accumulate_output_rows()
-                        .accumulate_output_bytes(),
-                )))
-            })?;
-        }
-
         if join.need_hold_hash_table {
             let mut projected_probe_fields = vec![];
             for (i, field) in probe_state.probe_schema.fields().iter().enumerate() {
@@ -332,8 +270,6 @@ impl PipelineBuilder {
             self.func_ctx.clone(),
             self.settings.clone(),
             left_side_ctx,
-            self.enable_profiling,
-            self.proc_profs.clone(),
             self.main_pipeline.get_scopes(),
         );
         left_side_builder.cte_state = self.cte_state.clone();
diff --git a/src/query/service/src/pipelines/builders/builder_limit.rs b/src/query/service/src/pipelines/builders/builder_limit.rs
index 7b150531fadc..e509274cc82c 100644
--- a/src/query/service/src/pipelines/builders/builder_limit.rs
+++ b/src/query/service/src/pipelines/builders/builder_limit.rs
@@ -14,7 +14,6 @@
 
 use databend_common_exception::Result;
 use databend_common_pipeline_core::processors::ProcessorPtr;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
 use databend_common_sql::executor::physical_plans::Limit;
 
 use crate::pipelines::processors::TransformLimit;
@@ -27,18 +26,12 @@ impl PipelineBuilder {
         if limit.limit.is_some() || limit.offset != 0 {
             self.main_pipeline.try_resize(1)?;
             return self.main_pipeline.add_transform(|input, output| {
-                let transform =
-                    TransformLimit::try_create(limit.limit, limit.offset, input, output)?;
-
-                if self.enable_profiling {
-                    Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                        transform,
-                        limit.plan_id,
-                        self.proc_profs.clone(),
-                    )))
-                } else {
-                    Ok(ProcessorPtr::create(transform))
-                }
+                Ok(ProcessorPtr::create(TransformLimit::try_create(
+                    limit.limit,
+                    limit.offset,
+                    input,
+                    output,
+                )?))
             });
         }
         Ok(())
diff --git a/src/query/service/src/pipelines/builders/builder_project.rs b/src/query/service/src/pipelines/builders/builder_project.rs
index b5baacc8569b..ef8433182d1d 100644
--- a/src/query/service/src/pipelines/builders/builder_project.rs
+++ b/src/query/service/src/pipelines/builders/builder_project.rs
@@ -19,7 +19,6 @@ use databend_common_functions::BUILTIN_FUNCTIONS;
 use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_core::Pipeline;
 use databend_common_pipeline_sinks::EmptySink;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
 use databend_common_sql::evaluator::BlockOperator;
 use databend_common_sql::evaluator::CompoundBlockOperator;
 use databend_common_sql::executor::physical_plans::Project;
@@ -90,24 +89,14 @@ impl PipelineBuilder {
         let max_block_size = self.settings.get_max_block_size()? as usize;
 
         self.main_pipeline.add_transform(|input, output| {
-            let transform = TransformSRF::try_create(
+            Ok(ProcessorPtr::create(TransformSRF::try_create(
                 input,
                 output,
                 self.func_ctx.clone(),
                 project_set.projections.clone(),
                 srf_exprs.clone(),
                 max_block_size,
-            );
-
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    project_set.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            )))
         })
     }
 }
diff --git a/src/query/service/src/pipelines/builders/builder_scalar.rs b/src/query/service/src/pipelines/builders/builder_scalar.rs
index a7bd3113a04b..7cbe9b62002d 100644
--- a/src/query/service/src/pipelines/builders/builder_scalar.rs
+++ b/src/query/service/src/pipelines/builders/builder_scalar.rs
@@ -15,7 +15,6 @@
 use databend_common_exception::Result;
 use databend_common_functions::BUILTIN_FUNCTIONS;
 use databend_common_pipeline_core::processors::ProcessorPtr;
-use databend_common_pipeline_transforms::processors::TransformProfileWrapper;
 use databend_common_pipeline_transforms::processors::Transformer;
 use databend_common_sql::evaluator::BlockOperator;
 use databend_common_sql::evaluator::CompoundBlockOperator;
@@ -46,25 +45,15 @@ impl PipelineBuilder {
         let num_input_columns = input_schema.num_fields();
 
         self.main_pipeline.add_transform(|input, output| {
-            let transform = CompoundBlockOperator::new(
-                vec![op.clone()],
-                self.func_ctx.clone(),
-                num_input_columns,
-            );
-
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(TransformProfileWrapper::create(
-                    transform,
-                    input,
-                    output,
-                    eval_scalar.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(Transformer::create(
-                    input, output, transform,
-                )))
-            }
+            Ok(ProcessorPtr::create(Transformer::create(
+                input,
+                output,
+                CompoundBlockOperator::new(
+                    vec![op.clone()],
+                    self.func_ctx.clone(),
+                    num_input_columns,
+                ),
+            )))
         })?;
 
         Ok(())
diff --git a/src/query/service/src/pipelines/builders/builder_scan.rs b/src/query/service/src/pipelines/builders/builder_scan.rs
index 3c4d87a93c2c..c1432a2ded6c 100644
--- a/src/query/service/src/pipelines/builders/builder_scan.rs
+++ b/src/query/service/src/pipelines/builders/builder_scan.rs
@@ -12,17 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::Arc;
-use std::sync::Mutex;
-use std::time::Instant;
-
 use databend_common_catalog::table_context::TableContext;
 use databend_common_exception::Result;
 use databend_common_expression::DataBlock;
 use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_sources::OneBlockSource;
-use databend_common_pipeline_transforms::processors::ProfileStub;
-use databend_common_pipeline_transforms::processors::Transformer;
 use databend_common_sql::evaluator::BlockOperator;
 use databend_common_sql::evaluator::CompoundBlockOperator;
 use databend_common_sql::executor::physical_plans::ConstantTableScan;
@@ -44,31 +38,6 @@ impl PipelineBuilder {
             true,
         )?;
 
-        if self.enable_profiling {
-            self.main_pipeline.add_transform(|input, output| {
-                // shared timer between `on_start` and `on_finish`
-                let start_timer = Arc::new(Mutex::new(Instant::now()));
-                let finish_timer = Arc::new(Mutex::new(Instant::now()));
-                Ok(ProcessorPtr::create(Transformer::create(
-                    input,
-                    output,
-                    ProfileStub::new(scan.plan_id, self.proc_profs.clone())
-                        .on_start(move |v| {
-                            *start_timer.lock().unwrap() = Instant::now();
-                            *v
-                        })
-                        .on_finish(move |prof| {
-                            let elapsed = finish_timer.lock().unwrap().elapsed();
-                            let mut prof = *prof;
-                            prof.wait_time = elapsed;
-                            prof
-                        })
-                        .accumulate_output_bytes()
-                        .accumulate_output_rows(),
-                )))
-            })?;
-        }
-
         // Fill internal columns if needed.
         if let Some(internal_columns) = &scan.internal_column {
             self.main_pipeline.add_transform(|input, output| {
diff --git a/src/query/service/src/pipelines/builders/builder_sort.rs b/src/query/service/src/pipelines/builders/builder_sort.rs
index 68edb97bf100..9990f1f4ca99 100644
--- a/src/query/service/src/pipelines/builders/builder_sort.rs
+++ b/src/query/service/src/pipelines/builders/builder_sort.rs
@@ -22,10 +22,8 @@ use databend_common_pipeline_core::query_spill_prefix;
 use databend_common_pipeline_core::Pipeline;
 use databend_common_pipeline_transforms::processors::sort::utils::add_order_field;
 use databend_common_pipeline_transforms::processors::try_add_multi_sort_merge;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
 use databend_common_pipeline_transforms::processors::TransformSortMergeBuilder;
 use databend_common_pipeline_transforms::processors::TransformSortPartial;
-use databend_common_profile::SharedProcessorProfiles;
 use databend_common_sql::evaluator::BlockOperator;
 use databend_common_sql::evaluator::CompoundBlockOperator;
 use databend_common_sql::executor::physical_plans::Sort;
@@ -89,20 +87,13 @@ impl PipelineBuilder {
             })
             .collect::<Result<Vec<_>>>()?;
 
-        self.build_sort_pipeline(
-            plan_schema,
-            sort_desc,
-            sort.plan_id,
-            sort.limit,
-            sort.after_exchange,
-        )
+        self.build_sort_pipeline(plan_schema, sort_desc, sort.limit, sort.after_exchange)
     }
 
     pub(crate) fn build_sort_pipeline(
         &mut self,
         plan_schema: DataSchemaRef,
         sort_desc: Vec<SortColumnDescription>,
-        plan_id: u32,
         limit: Option<usize>,
         after_exchange: Option<bool>,
     ) -> Result<()> {
@@ -114,18 +105,12 @@ impl PipelineBuilder {
         if self.main_pipeline.output_len() == 1 || max_threads == 1 {
             self.main_pipeline.try_resize(max_threads)?;
         }
-        let prof_info = if self.enable_profiling {
-            Some((plan_id, self.proc_profs.clone()))
-        } else {
-            None
-        };
 
         let mut builder =
             SortPipelineBuilder::create(self.ctx.clone(), plan_schema.clone(), sort_desc.clone())
                 .with_partial_block_size(block_size)
                 .with_final_block_size(block_size)
-                .with_limit(limit)
-                .with_prof_info(prof_info.clone());
+                .with_limit(limit);
 
         match after_exchange {
             Some(true) => {
@@ -140,7 +125,6 @@ impl PipelineBuilder {
                         block_size,
                         limit,
                         sort_desc,
-                        prof_info,
                         true,
                     )
                 } else {
@@ -171,7 +155,6 @@ pub struct SortPipelineBuilder {
     limit: Option<usize>,
     partial_block_size: usize,
     final_block_size: usize,
-    prof_info: Option<(u32, SharedProcessorProfiles)>,
     remove_order_col_at_last: bool,
 }
 
@@ -188,7 +171,6 @@ impl SortPipelineBuilder {
             limit: None,
             partial_block_size: 0,
             final_block_size: 0,
-            prof_info: None,
             remove_order_col_at_last: false,
         }
     }
@@ -208,11 +190,6 @@ impl SortPipelineBuilder {
         self
     }
 
-    pub fn with_prof_info(mut self, prof_info: Option<(u32, SharedProcessorProfiles)>) -> Self {
-        self.prof_info = prof_info;
-        self
-    }
-
     pub fn remove_order_col_at_last(mut self) -> Self {
         self.remove_order_col_at_last = true;
         self
@@ -221,21 +198,12 @@ impl SortPipelineBuilder {
     pub fn build_full_sort_pipeline(self, pipeline: &mut Pipeline) -> Result<()> {
         // Partial sort
         pipeline.add_transform(|input, output| {
-            let transform = TransformSortPartial::try_create(
+            Ok(ProcessorPtr::create(TransformSortPartial::try_create(
                 input,
                 output,
                 self.limit,
                 self.sort_desc.clone(),
-            )?;
-            if let Some((plan_id, prof)) = &self.prof_info {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    *plan_id,
-                    prof.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            )?))
         })?;
 
         self.build_merge_sort_pipeline(pipeline, false)
@@ -310,16 +278,7 @@ impl SortPipelineBuilder {
             .with_max_memory_usage(max_memory_usage)
             .with_spilling_bytes_threshold_per_core(bytes_limit_per_proc);
 
-            let transform = builder.build()?;
-            if let Some((plan_id, prof)) = &self.prof_info {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    *plan_id,
-                    prof.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            Ok(ProcessorPtr::create(builder.build()?))
         })?;
 
         if may_spill {
@@ -329,7 +288,7 @@ impl SortPipelineBuilder {
                 let op = DataOperator::instance().operator();
                 let spiller =
                     Spiller::create(self.ctx.clone(), op, config.clone(), SpillerType::OrderBy);
-                let transform = create_transform_sort_spill(
+                Ok(ProcessorPtr::create(create_transform_sort_spill(
                     input,
                     output,
                     schema.clone(),
@@ -337,16 +296,7 @@ impl SortPipelineBuilder {
                     self.limit,
                     spiller,
                     output_order_col,
-                );
-                if let Some((plan_id, prof)) = &self.prof_info {
-                    Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                        transform,
-                        *plan_id,
-                        prof.clone(),
-                    )))
-                } else {
-                    Ok(ProcessorPtr::create(transform))
-                }
+                )))
             })?;
         }
 
@@ -358,7 +308,6 @@ impl SortPipelineBuilder {
                 self.final_block_size,
                 self.limit,
                 self.sort_desc,
-                self.prof_info.clone(),
                 self.remove_order_col_at_last,
             )?;
         }
diff --git a/src/query/service/src/pipelines/builders/builder_udf.rs b/src/query/service/src/pipelines/builders/builder_udf.rs
index e82e17799537..9511ca35b687 100644
--- a/src/query/service/src/pipelines/builders/builder_udf.rs
+++ b/src/query/service/src/pipelines/builders/builder_udf.rs
@@ -14,7 +14,6 @@
 
 use databend_common_exception::Result;
 use databend_common_pipeline_core::processors::ProcessorPtr;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
 use databend_common_sql::executor::physical_plans::Udf;
 
 use crate::pipelines::processors::transforms::TransformUdf;
@@ -25,21 +24,12 @@ impl PipelineBuilder {
         self.build_pipeline(&udf.input)?;
 
         self.main_pipeline.add_transform(|input, output| {
-            let transform = TransformUdf::try_create(
+            Ok(ProcessorPtr::create(TransformUdf::try_create(
                 self.func_ctx.clone(),
                 udf.udf_funcs.clone(),
                 input,
                 output,
-            )?;
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    udf.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            )?))
         })
     }
 }
diff --git a/src/query/service/src/pipelines/builders/builder_union_all.rs b/src/query/service/src/pipelines/builders/builder_union_all.rs
index 7b21330757a3..7ed40bc87e80 100644
--- a/src/query/service/src/pipelines/builders/builder_union_all.rs
+++ b/src/query/service/src/pipelines/builders/builder_union_all.rs
@@ -17,7 +17,6 @@ use databend_common_exception::Result;
 use databend_common_expression::DataBlock;
 use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_sinks::UnionReceiveSink;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
 use databend_common_sql::executor::physical_plans::UnionAll;
 use databend_common_sql::executor::PhysicalPlan;
 
@@ -28,43 +27,27 @@ use crate::sessions::QueryContext;
 impl PipelineBuilder {
     pub fn build_union_all(&mut self, union_all: &UnionAll) -> Result<()> {
         self.build_pipeline(&union_all.left)?;
-        let union_all_receiver = self.expand_union_all(&union_all.right, union_all)?;
+        let union_all_receiver = self.expand_union_all(&union_all.right)?;
         self.main_pipeline
             .add_transform(|transform_input_port, transform_output_port| {
-                let transform = TransformMergeBlock::try_create(
+                Ok(ProcessorPtr::create(TransformMergeBlock::try_create(
                     transform_input_port,
                     transform_output_port,
                     union_all.left.output_schema()?,
                     union_all.right.output_schema()?,
                     union_all.pairs.clone(),
                     union_all_receiver.clone(),
-                )?;
-
-                if self.enable_profiling {
-                    Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                        transform,
-                        union_all.plan_id,
-                        self.proc_profs.clone(),
-                    )))
-                } else {
-                    Ok(ProcessorPtr::create(transform))
-                }
+                )?))
             })?;
         Ok(())
     }
 
-    fn expand_union_all(
-        &mut self,
-        input: &PhysicalPlan,
-        union_plan: &UnionAll,
-    ) -> Result<Receiver<DataBlock>> {
+    fn expand_union_all(&mut self, input: &PhysicalPlan) -> Result<Receiver<DataBlock>> {
         let union_ctx = QueryContext::create_from(self.ctx.clone());
         let mut pipeline_builder = PipelineBuilder::create(
             self.func_ctx.clone(),
             self.settings.clone(),
             union_ctx,
-            self.enable_profiling,
-            self.proc_profs.clone(),
             self.main_pipeline.get_scopes(),
         );
         pipeline_builder.cte_state = self.cte_state.clone();
@@ -76,18 +59,11 @@ impl PipelineBuilder {
         let (tx, rx) = async_channel::unbounded();
 
         build_res.main_pipeline.add_sink(|input_port| {
-            let transform =
-                UnionReceiveSink::create(Some(tx.clone()), input_port, self.ctx.clone());
-
-            if self.enable_profiling {
-                Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                    transform,
-                    union_plan.plan_id,
-                    self.proc_profs.clone(),
-                )))
-            } else {
-                Ok(ProcessorPtr::create(transform))
-            }
+            Ok(ProcessorPtr::create(UnionReceiveSink::create(
+                Some(tx.clone()),
+                input_port,
+                self.ctx.clone(),
+            )))
         })?;
 
         self.pipelines.push(build_res.main_pipeline.finalize());
diff --git a/src/query/service/src/pipelines/builders/builder_window.rs b/src/query/service/src/pipelines/builders/builder_window.rs
index 82af7ab49986..41043661a364 100644
--- a/src/query/service/src/pipelines/builders/builder_window.rs
+++ b/src/query/service/src/pipelines/builders/builder_window.rs
@@ -70,7 +70,7 @@ impl PipelineBuilder {
 
             sort_desc.extend(order_by.clone());
 
-            self.build_sort_pipeline(input_schema.clone(), sort_desc, window.plan_id, None, None)?;
+            self.build_sort_pipeline(input_schema.clone(), sort_desc, window.limit, None)?;
         }
         // `TransformWindow` is a pipeline breaker.
         self.main_pipeline.try_resize(1)?;
diff --git a/src/query/service/src/pipelines/executor/executor_graph.rs b/src/query/service/src/pipelines/executor/executor_graph.rs
index 134fb505cc1e..e2d416fba33e 100644
--- a/src/query/service/src/pipelines/executor/executor_graph.rs
+++ b/src/query/service/src/pipelines/executor/executor_graph.rs
@@ -23,8 +23,8 @@ use databend_common_base::runtime::TrackedFuture;
 use databend_common_base::runtime::TrySpawn;
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
-use databend_common_pipeline_core::processors::profile::Profile;
 use databend_common_pipeline_core::processors::EventCause;
+use databend_common_pipeline_core::processors::Profile;
 use databend_common_pipeline_core::Pipeline;
 use databend_common_pipeline_core::PlanScope;
 use log::debug;
@@ -207,6 +207,12 @@ impl ExecutingGraph {
                     let output_trigger = graph[source_node].create_trigger(edge_index);
                     graph[source_node].outputs_port[source_port].set_trigger(output_trigger);
 
+                    if graph[source_node].profile.plan_id.is_some()
+                        && graph[source_node].profile.plan_id != graph[target_node].profile.plan_id
+                    {
+                        graph[source_node].outputs_port[source_port].record_profile();
+                    }
+
                     connect(
                         &graph[target_node].inputs_port[target_port],
                         &graph[source_node].outputs_port[source_port],
@@ -276,6 +282,8 @@ impl ExecutingGraph {
             if let Some(schedule_index) = need_schedule_nodes.pop_front() {
                 let node = &locker.graph[schedule_index];
 
+                Profile::track_profile(&node.profile);
+
                 if state_guard_cache.is_none() {
                     state_guard_cache = Some(node.state.lock().unwrap());
                 }
@@ -386,6 +394,7 @@ impl ScheduleQueue {
         unsafe {
             workers_condvar.inc_active_async_worker();
             let weak_executor = Arc::downgrade(executor);
+            let node_profile = executor.graph.get_node_profile(proc.id()).clone();
             let process_future = proc.async_process();
             executor.async_runtime.spawn(
                 query_id.as_ref().clone(),
@@ -396,6 +405,7 @@ impl ScheduleQueue {
                     global_queue,
                     workers_condvar,
                     weak_executor,
+                    node_profile,
                     process_future,
                 ))
                 .in_span(Span::enter_with_local_parent(std::any::type_name::<
@@ -443,8 +453,8 @@ impl RunningGraph {
         Ok(schedule_queue)
     }
 
-    pub(crate) fn get_node(&self, pid: NodeIndex) -> &Node {
-        &self.0.graph[pid]
+    pub(crate) fn get_node_profile(&self, pid: NodeIndex) -> &Arc<Profile> {
+        &self.0.graph[pid].profile
     }
 
     pub fn get_proc_profiles(&self) -> Vec<Arc<Profile>> {
diff --git a/src/query/service/src/pipelines/executor/executor_settings.rs b/src/query/service/src/pipelines/executor/executor_settings.rs
index 1275791cc92f..b45bc8465582 100644
--- a/src/query/service/src/pipelines/executor/executor_settings.rs
+++ b/src/query/service/src/pipelines/executor/executor_settings.rs
@@ -21,16 +21,13 @@ use databend_common_settings::Settings;
 #[derive(Clone)]
 pub struct ExecutorSettings {
     pub query_id: Arc<String>,
-    pub enable_profiling: bool,
     pub max_execute_time_in_seconds: Duration,
 }
 
 impl ExecutorSettings {
     pub fn try_create(settings: &Settings, query_id: String) -> Result<ExecutorSettings> {
-        let enable_profiling = settings.get_enable_query_profiling()?;
         let max_execute_time_in_seconds = settings.get_max_execute_time_in_seconds()?;
         Ok(ExecutorSettings {
-            enable_profiling,
             query_id: Arc::new(query_id),
             max_execute_time_in_seconds: Duration::from_secs(max_execute_time_in_seconds),
         })
diff --git a/src/query/service/src/pipelines/executor/executor_tasks.rs b/src/query/service/src/pipelines/executor/executor_tasks.rs
index 3168cce975a1..cc70fcd199ee 100644
--- a/src/query/service/src/pipelines/executor/executor_tasks.rs
+++ b/src/query/service/src/pipelines/executor/executor_tasks.rs
@@ -16,7 +16,6 @@ use std::collections::VecDeque;
 use std::sync::atomic::AtomicBool;
 use std::sync::atomic::Ordering;
 use std::sync::Arc;
-use std::time::Duration;
 
 use databend_common_exception::Result;
 use parking_lot::Mutex;
@@ -198,22 +197,11 @@ pub struct CompletedAsyncTask {
     pub id: NodeIndex,
     pub worker_id: usize,
     pub res: Result<()>,
-    pub elapsed: Option<Duration>,
 }
 
 impl CompletedAsyncTask {
-    pub fn create(
-        id: NodeIndex,
-        worker_id: usize,
-        res: Result<()>,
-        elapsed: Option<Duration>,
-    ) -> Self {
-        CompletedAsyncTask {
-            id,
-            worker_id,
-            res,
-            elapsed,
-        }
+    pub fn create(id: NodeIndex, worker_id: usize, res: Result<()>) -> Self {
+        CompletedAsyncTask { id, worker_id, res }
     }
 }
 
diff --git a/src/query/service/src/pipelines/executor/executor_worker_context.rs b/src/query/service/src/pipelines/executor/executor_worker_context.rs
index deb86134fa3a..9b430dc302e3 100644
--- a/src/query/service/src/pipelines/executor/executor_worker_context.rs
+++ b/src/query/service/src/pipelines/executor/executor_worker_context.rs
@@ -14,15 +14,18 @@
 
 use std::fmt::Debug;
 use std::fmt::Formatter;
+use std::intrinsics::assume;
 use std::sync::Arc;
-use std::time::Duration;
 use std::time::Instant;
 
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use petgraph::prelude::NodeIndex;
 
 use crate::pipelines::executor::CompletedAsyncTask;
+use crate::pipelines::executor::RunningGraph;
 use crate::pipelines::executor::WorkersCondvar;
 use crate::pipelines::processors::ProcessorPtr;
 
@@ -70,35 +73,33 @@ impl ExecutorWorkerContext {
     }
 
     /// # Safety
-    pub unsafe fn execute_task<const ENABLE_PROFILING: bool>(
-        &mut self,
-    ) -> Result<(NodeIndex, bool, Option<Duration>)> {
+    pub unsafe fn execute_task(&mut self, graph: &RunningGraph) -> Result<NodeIndex> {
         match std::mem::replace(&mut self.task, ExecutorTask::None) {
             ExecutorTask::None => Err(ErrorCode::Internal("Execute none task.")),
-            ExecutorTask::Sync(processor) => self.execute_sync_task::<ENABLE_PROFILING>(processor),
+            ExecutorTask::Sync(processor) => self.execute_sync_task(processor, graph),
             ExecutorTask::AsyncCompleted(task) => match task.res {
-                Ok(_) => Ok((task.id, true, task.elapsed)),
+                Ok(_) => Ok(task.id),
                 Err(cause) => Err(cause),
             },
         }
     }
 
     /// # Safety
-    unsafe fn execute_sync_task<const ENABLE_PROFILING: bool>(
+    unsafe fn execute_sync_task(
         &mut self,
         proc: ProcessorPtr,
-    ) -> Result<(NodeIndex, bool, Option<Duration>)> {
-        match ENABLE_PROFILING {
-            true => {
-                let instant = Instant::now();
-                proc.process()?;
-                Ok((proc.id(), false, Some(instant.elapsed())))
-            }
-            false => {
-                proc.process()?;
-                Ok((proc.id(), false, None))
-            }
-        }
+        graph: &RunningGraph,
+    ) -> Result<NodeIndex> {
+        Profile::track_profile(graph.get_node_profile(proc.id()));
+
+        let instant = Instant::now();
+
+        proc.process()?;
+
+        let nanos = instant.elapsed().as_nanos();
+        assume(nanos < 18446744073709551615_u128);
+        Profile::record_usize_profile(ProfileStatisticsName::CpuTime, nanos as usize);
+        Ok(proc.id())
     }
 
     pub fn get_workers_condvar(&self) -> &Arc<WorkersCondvar> {
diff --git a/src/query/service/src/pipelines/executor/pipeline_executor.rs b/src/query/service/src/pipelines/executor/pipeline_executor.rs
index 7d8c077d5a87..027a865bbb17 100644
--- a/src/query/service/src/pipelines/executor/pipeline_executor.rs
+++ b/src/query/service/src/pipelines/executor/pipeline_executor.rs
@@ -12,8 +12,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::intrinsics::assume;
-use std::sync::atomic::Ordering;
 use std::sync::Arc;
 use std::time::Instant;
 
@@ -27,7 +25,7 @@ use databend_common_base::runtime::TrySpawn;
 use databend_common_base::GLOBAL_TASK;
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
-use databend_common_pipeline_core::processors::profile::Profile;
+use databend_common_pipeline_core::processors::Profile;
 use databend_common_pipeline_core::LockGuard;
 use databend_common_pipeline_core::Pipeline;
 use futures::future::select;
@@ -357,14 +355,8 @@ impl PipelineExecutor {
             thread_join_handles.push(Thread::named_spawn(Some(name), move || unsafe {
                 let _g = span.set_local_parent();
                 let this_clone = this.clone();
-                let enable_profiling = this.settings.enable_profiling;
                 let try_result = catch_unwind(move || -> Result<()> {
-                    let res = match enable_profiling {
-                        true => this_clone.execute_single_thread::<true>(thread_num),
-                        false => this_clone.execute_single_thread::<false>(thread_num),
-                    };
-
-                    match res {
+                    match this_clone.execute_single_thread(thread_num) {
                         Ok(_) => Ok(()),
                         Err(cause) => {
                             if log::max_level() == LevelFilter::Trace {
@@ -393,10 +385,7 @@ impl PipelineExecutor {
     /// # Safety
     ///
     /// Method is thread unsafe and require thread safe call
-    pub unsafe fn execute_single_thread<const ENABLE_PROFILING: bool>(
-        self: &Arc<Self>,
-        thread_num: usize,
-    ) -> Result<()> {
+    pub unsafe fn execute_single_thread(self: &Arc<Self>, thread_num: usize) -> Result<()> {
         let workers_condvar = self.workers_condvar.clone();
         let mut context = ExecutorWorkerContext::create(
             thread_num,
@@ -411,28 +400,7 @@ impl PipelineExecutor {
             }
 
             while !self.global_tasks_queue.is_finished() && context.has_task() {
-                let (executed_pid, is_async, elapsed) =
-                    context.execute_task::<ENABLE_PROFILING>()?;
-
-                if ENABLE_PROFILING {
-                    let node = self.graph.get_node(executed_pid);
-                    if let Some(elapsed) = elapsed {
-                        let nanos = elapsed.as_nanos();
-                        assume(nanos < 18446744073709551615_u128);
-
-                        if is_async {
-                            node.profile
-                                .wait_time
-                                .fetch_add(nanos as u64, Ordering::Relaxed);
-                        } else {
-                            node.profile
-                                .cpu_time
-                                .fetch_add(nanos as u64, Ordering::Relaxed);
-                        }
-                    }
-
-                    node.processor.record_profile(&node.profile);
-                }
+                let executed_pid = context.execute_task(&self.graph)?;
 
                 // Not scheduled graph if pipeline is finished.
                 if !self.global_tasks_queue.is_finished() {
diff --git a/src/query/service/src/pipelines/executor/processor_async_task.rs b/src/query/service/src/pipelines/executor/processor_async_task.rs
index 11bd17b0f801..696acd10cfc0 100644
--- a/src/query/service/src/pipelines/executor/processor_async_task.rs
+++ b/src/query/service/src/pipelines/executor/processor_async_task.rs
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 use std::future::Future;
+use std::intrinsics::assume;
 use std::pin::Pin;
 use std::sync::Arc;
 use std::sync::Weak;
@@ -26,6 +27,8 @@ use databend_common_base::runtime::catch_unwind;
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use futures_util::future::BoxFuture;
 use futures_util::future::Either;
 use futures_util::FutureExt;
@@ -43,7 +46,10 @@ pub struct ProcessorAsyncTask {
     processor_id: NodeIndex,
     queue: Arc<ExecutorTasksQueue>,
     workers_condvar: Arc<WorkersCondvar>,
-    inner: BoxFuture<'static, (Duration, Result<()>)>,
+    profile: Arc<Profile>,
+    instant: Instant,
+    last_nanos: usize,
+    inner: BoxFuture<'static, Result<()>>,
 }
 
 impl ProcessorAsyncTask {
@@ -54,6 +60,7 @@ impl ProcessorAsyncTask {
         queue: Arc<ExecutorTasksQueue>,
         workers_condvar: Arc<WorkersCondvar>,
         weak_executor: Weak<PipelineExecutor>,
+        profile: Arc<Profile>,
         inner: Inner,
     ) -> ProcessorAsyncTask {
         let finished_notify = queue.get_finished_notify();
@@ -111,17 +118,21 @@ impl ProcessorAsyncTask {
                         };
                     }
                     Either::Right((res, _)) => {
-                        return (start.elapsed(), res);
+                        return res;
                     }
                 }
             }
         };
 
+        let instant = Instant::now();
         ProcessorAsyncTask {
             worker_id,
             processor_id,
             queue,
             workers_condvar,
+            profile,
+            last_nanos: instant.elapsed().as_nanos() as usize,
+            instant,
             inner: inner.boxed(),
         }
     }
@@ -135,26 +146,40 @@ impl Future for ProcessorAsyncTask {
             return Poll::Ready(());
         }
 
+        Profile::track_profile(&self.profile);
+
+        let last_nanos = self.last_nanos;
+        let last_instant = self.instant;
         let inner = self.inner.as_mut();
 
-        match catch_unwind(move || inner.poll(cx)) {
-            Ok(Poll::Pending) => Poll::Pending,
-            Ok(Poll::Ready((elapsed, res))) => {
+        let before_poll_nanos = elapsed_nanos(last_instant);
+        let wait_nanos = before_poll_nanos - last_nanos;
+        Profile::record_usize_profile(ProfileStatisticsName::WaitTime, wait_nanos);
+
+        let poll_res = catch_unwind(move || inner.poll(cx));
+
+        let after_poll_nanos = elapsed_nanos(last_instant);
+        Profile::record_usize_profile(
+            ProfileStatisticsName::CpuTime,
+            after_poll_nanos - before_poll_nanos,
+        );
+
+        match poll_res {
+            Ok(Poll::Pending) => {
+                self.last_nanos = after_poll_nanos;
+                Poll::Pending
+            }
+            Ok(Poll::Ready(res)) => {
                 self.queue.completed_async_task(
                     self.workers_condvar.clone(),
-                    CompletedAsyncTask::create(
-                        self.processor_id,
-                        self.worker_id,
-                        res,
-                        Some(elapsed),
-                    ),
+                    CompletedAsyncTask::create(self.processor_id, self.worker_id, res),
                 );
                 Poll::Ready(())
             }
             Err(cause) => {
                 self.queue.completed_async_task(
                     self.workers_condvar.clone(),
-                    CompletedAsyncTask::create(self.processor_id, self.worker_id, Err(cause), None),
+                    CompletedAsyncTask::create(self.processor_id, self.worker_id, Err(cause)),
                 );
 
                 Poll::Ready(())
@@ -162,3 +187,11 @@ impl Future for ProcessorAsyncTask {
         }
     }
 }
+
+fn elapsed_nanos(instant: Instant) -> usize {
+    let nanos = (Instant::now() - instant).as_nanos();
+    unsafe {
+        assume(nanos < 18446744073709551615_u128);
+    }
+    nanos as usize
+}
diff --git a/src/query/service/src/pipelines/pipeline_build_res.rs b/src/query/service/src/pipelines/pipeline_build_res.rs
index b08c0b0635dc..ad2fb9ca72ff 100644
--- a/src/query/service/src/pipelines/pipeline_build_res.rs
+++ b/src/query/service/src/pipelines/pipeline_build_res.rs
@@ -21,7 +21,6 @@ use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::Pipeline;
 use databend_common_pipeline_core::SourcePipeBuilder;
 use databend_common_pipeline_sources::OneBlockSource;
-use databend_common_profile::SharedProcessorProfiles;
 
 use crate::api::DefaultExchangeInjector;
 use crate::api::ExchangeInjector;
@@ -38,10 +37,6 @@ pub struct PipelineBuildResult {
     // Containing some sub queries pipelines, must be complete pipeline
     pub sources_pipelines: Vec<Pipeline>,
 
-    /// Set of profiling spans for the query.
-    /// Will be empty if profiling is disabled.
-    pub prof_span_set: SharedProcessorProfiles,
-
     pub exchange_injector: Arc<dyn ExchangeInjector>,
     /// for local fragment data sharing
     pub builder_data: PipelineBuilderData,
@@ -52,7 +47,6 @@ impl PipelineBuildResult {
         PipelineBuildResult {
             main_pipeline: Pipeline::create(),
             sources_pipelines: vec![],
-            prof_span_set: SharedProcessorProfiles::default(),
             exchange_injector: DefaultExchangeInjector::create(),
             builder_data: PipelineBuilderData {
                 input_join_state: None,
@@ -75,7 +69,6 @@ impl PipelineBuildResult {
         Ok(PipelineBuildResult {
             main_pipeline,
             sources_pipelines: vec![],
-            prof_span_set: SharedProcessorProfiles::default(),
             exchange_injector: DefaultExchangeInjector::create(),
             builder_data: PipelineBuilderData {
                 input_join_state: None,
diff --git a/src/query/service/src/pipelines/pipeline_builder.rs b/src/query/service/src/pipelines/pipeline_builder.rs
index d5ca7c0af540..075426da8960 100644
--- a/src/query/service/src/pipelines/pipeline_builder.rs
+++ b/src/query/service/src/pipelines/pipeline_builder.rs
@@ -19,10 +19,10 @@ use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_expression::DataField;
 use databend_common_expression::FunctionContext;
+use databend_common_pipeline_core::processors::profile::ProfileLabel;
 use databend_common_pipeline_core::Pipeline;
 use databend_common_pipeline_core::PlanScope;
 use databend_common_pipeline_core::PlanScopeGuard;
-use databend_common_profile::SharedProcessorProfiles;
 use databend_common_settings::Settings;
 use databend_common_sql::executor::PhysicalPlan;
 use databend_common_sql::IndexType;
@@ -50,8 +50,6 @@ pub struct PipelineBuilder {
     // Cte -> state, each cte has it's own state
     pub cte_state: HashMap<IndexType, Arc<MaterializedCteState>>,
 
-    pub(crate) enable_profiling: bool,
-    pub(crate) proc_profs: SharedProcessorProfiles,
     pub(crate) exchange_injector: Arc<dyn ExchangeInjector>,
 }
 
@@ -60,18 +58,14 @@ impl PipelineBuilder {
         func_ctx: FunctionContext,
         settings: Arc<Settings>,
         ctx: Arc<QueryContext>,
-        enable_profiling: bool,
-        prof_span_set: SharedProcessorProfiles,
         scopes: Vec<PlanScope>,
     ) -> PipelineBuilder {
         PipelineBuilder {
-            enable_profiling,
             ctx,
             func_ctx,
             settings,
             pipelines: vec![],
             main_pipeline: Pipeline::with_scopes(scopes),
-            proc_profs: prof_span_set,
             exchange_injector: DefaultExchangeInjector::create(),
             cte_state: HashMap::new(),
             merge_into_probe_data_fields: None,
@@ -93,7 +87,6 @@ impl PipelineBuilder {
         Ok(PipelineBuildResult {
             main_pipeline: self.main_pipeline,
             sources_pipelines: self.pipelines,
-            prof_span_set: self.proc_profs,
             exchange_injector: self.exchange_injector,
             builder_data: PipelineBuilderData {
                 input_join_state: self.join_state,
@@ -102,18 +95,30 @@ impl PipelineBuilder {
         })
     }
 
-    pub(crate) fn add_plan_scope(&mut self, plan: &PhysicalPlan) -> Option<PlanScopeGuard> {
+    pub(crate) fn add_plan_scope(&mut self, plan: &PhysicalPlan) -> Result<Option<PlanScopeGuard>> {
         match plan {
-            PhysicalPlan::EvalScalar(v) if v.exprs.is_empty() => None,
+            PhysicalPlan::EvalScalar(v) if v.exprs.is_empty() => Ok(None),
             _ => {
-                let scope = PlanScope::create(plan.get_id(), plan.name());
-                Some(self.main_pipeline.add_plan_scope(scope))
+                let desc = plan.get_desc()?;
+                let plan_labels = plan.get_labels()?;
+                let mut profile_labels = Vec::with_capacity(plan_labels.len());
+                for (name, value) in plan_labels {
+                    profile_labels.push(ProfileLabel::create(name, value));
+                }
+
+                let scope = PlanScope::create(
+                    plan.get_id(),
+                    plan.name(),
+                    Arc::new(desc),
+                    Arc::new(profile_labels),
+                );
+                Ok(Some(self.main_pipeline.add_plan_scope(scope)))
             }
         }
     }
 
     pub(crate) fn build_pipeline(&mut self, plan: &PhysicalPlan) -> Result<()> {
-        let _guard = self.add_plan_scope(plan);
+        let _guard = self.add_plan_scope(plan)?;
         match plan {
             PhysicalPlan::TableScan(scan) => self.build_table_scan(scan),
             PhysicalPlan::CteScan(scan) => self.build_cte_scan(scan),
diff --git a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_aggregate_spill_writer.rs b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_aggregate_spill_writer.rs
index c53da0393099..cfb2be722203 100644
--- a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_aggregate_spill_writer.rs
+++ b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_aggregate_spill_writer.rs
@@ -30,6 +30,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use futures_util::future::BoxFuture;
 use log::info;
 use opendal::Operator;
@@ -258,6 +260,13 @@ pub fn spilling_aggregate_payload<Method: HashMethodBounds>(
             metrics_inc_aggregate_spill_write_count();
             metrics_inc_aggregate_spill_write_bytes(write_bytes as u64);
             metrics_inc_aggregate_spill_write_milliseconds(instant.elapsed().as_millis() as u64);
+
+            Profile::record_usize_profile(ProfileStatisticsName::SpillWriteCount, 1);
+            Profile::record_usize_profile(ProfileStatisticsName::SpillWriteBytes, write_bytes);
+            Profile::record_usize_profile(
+                ProfileStatisticsName::SpillWriteTime,
+                instant.elapsed().as_millis() as usize,
+            );
         }
 
         {
diff --git a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_aggregate_serializer.rs b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_aggregate_serializer.rs
index 183d12e8629f..823721b4666e 100644
--- a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_aggregate_serializer.rs
+++ b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_aggregate_serializer.rs
@@ -39,6 +39,8 @@ use databend_common_metrics::transform::*;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_transforms::processors::BlockMetaTransform;
 use databend_common_pipeline_transforms::processors::BlockMetaTransformer;
 use databend_common_settings::FlightCompression;
@@ -261,6 +263,13 @@ fn spilling_aggregate_payload<Method: HashMethodBounds>(
                 metrics_inc_aggregate_spill_write_count();
                 metrics_inc_aggregate_spill_write_bytes(write_bytes as u64);
                 metrics_inc_aggregate_spill_write_milliseconds(instant.elapsed().as_millis() as u64);
+
+                Profile::record_usize_profile(ProfileStatisticsName::SpillWriteCount, 1);
+                Profile::record_usize_profile(ProfileStatisticsName::SpillWriteBytes, write_bytes);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::SpillWriteTime,
+                    instant.elapsed().as_millis() as usize,
+                );
             }
 
             {
diff --git a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_group_by_serializer.rs b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_group_by_serializer.rs
index 53c51c5a005f..afc4fa7a333c 100644
--- a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_group_by_serializer.rs
+++ b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_exchange_group_by_serializer.rs
@@ -44,6 +44,8 @@ use databend_common_metrics::transform::*;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_transforms::processors::BlockMetaTransform;
 use databend_common_pipeline_transforms::processors::BlockMetaTransformer;
 use databend_common_pipeline_transforms::processors::UnknownMode;
@@ -313,6 +315,13 @@ fn spilling_group_by_payload<Method: HashMethodBounds>(
                 metrics_inc_group_by_spill_write_count();
                 metrics_inc_group_by_spill_write_bytes(write_bytes as u64);
                 metrics_inc_group_by_spill_write_milliseconds(instant.elapsed().as_millis() as u64);
+
+                Profile::record_usize_profile(ProfileStatisticsName::SpillWriteCount, 1);
+                Profile::record_usize_profile(ProfileStatisticsName::SpillWriteBytes, write_bytes);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::SpillWriteTime,
+                    instant.elapsed().as_millis() as usize,
+                );
             }
 
             {
diff --git a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_group_by_spill_writer.rs b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_group_by_spill_writer.rs
index 78905671f23e..1ddb32550494 100644
--- a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_group_by_spill_writer.rs
+++ b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_group_by_spill_writer.rs
@@ -30,6 +30,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use futures_util::future::BoxFuture;
 use log::info;
 use opendal::Operator;
@@ -251,6 +253,13 @@ pub fn spilling_group_by_payload<Method: HashMethodBounds>(
             metrics_inc_group_by_spill_write_count();
             metrics_inc_group_by_spill_write_bytes(write_bytes as u64);
             metrics_inc_group_by_spill_write_milliseconds(instant.elapsed().as_millis() as u64);
+
+            Profile::record_usize_profile(ProfileStatisticsName::SpillWriteCount, 1);
+            Profile::record_usize_profile(ProfileStatisticsName::SpillWriteBytes, write_bytes);
+            Profile::record_usize_profile(
+                ProfileStatisticsName::SpillWriteTime,
+                instant.elapsed().as_millis() as usize,
+            );
         }
 
         {
diff --git a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_spill_reader.rs b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_spill_reader.rs
index eb52e34c3d97..f0cfc0a1b1ab 100644
--- a/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_spill_reader.rs
+++ b/src/query/service/src/pipelines/processors/transforms/aggregator/serde/transform_spill_reader.rs
@@ -29,6 +29,8 @@ use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use itertools::Itertools;
 use log::info;
 use opendal::Operator;
@@ -212,6 +214,19 @@ impl<Method: HashMethodBounds, V: Send + Sync + 'static> Processor
                                         metrics_inc_aggregate_spill_read_milliseconds(
                                             instant.elapsed().as_millis() as u64,
                                         );
+
+                                        Profile::record_usize_profile(
+                                            ProfileStatisticsName::SpillReadCount,
+                                            1,
+                                        );
+                                        Profile::record_usize_profile(
+                                            ProfileStatisticsName::SpillReadBytes,
+                                            data.len(),
+                                        );
+                                        Profile::record_usize_profile(
+                                            ProfileStatisticsName::SpillReadTime,
+                                            instant.elapsed().as_millis() as usize,
+                                        );
                                     }
 
                                     info!(
diff --git a/src/query/service/src/pipelines/processors/transforms/aggregator/transform_partition_bucket.rs b/src/query/service/src/pipelines/processors/transforms/aggregator/transform_partition_bucket.rs
index 550dc19610f9..8f1b414a7997 100644
--- a/src/query/service/src/pipelines/processors/transforms/aggregator/transform_partition_bucket.rs
+++ b/src/query/service/src/pipelines/processors/transforms/aggregator/transform_partition_bucket.rs
@@ -33,10 +33,6 @@ use databend_common_pipeline_core::processors::ProcessorPtr;
 use databend_common_pipeline_core::Pipe;
 use databend_common_pipeline_core::PipeItem;
 use databend_common_pipeline_core::Pipeline;
-use databend_common_pipeline_transforms::processors::ProcessorProfileWrapper;
-use databend_common_pipeline_transforms::processors::ProfileStub;
-use databend_common_pipeline_transforms::processors::Transformer;
-use databend_common_profile::SharedProcessorProfiles;
 use databend_common_storage::DataOperator;
 
 use crate::pipelines::processors::transforms::aggregator::aggregate_meta::AggregateMeta;
@@ -443,9 +439,6 @@ pub fn build_partition_bucket<Method: HashMethodBounds, V: Copy + Send + Sync +
     method: Method,
     pipeline: &mut Pipeline,
     params: Arc<AggregatorParams>,
-    enable_profiling: bool,
-    prof_id: u32,
-    proc_profs: SharedProcessorProfiles,
 ) -> Result<()> {
     let input_nums = pipeline.output_len();
     let transform = TransformPartitionBucket::<Method, V>::create(method.clone(), input_nums)?;
@@ -471,36 +464,23 @@ pub fn build_partition_bucket<Method: HashMethodBounds, V: Copy + Send + Sync +
     })?;
 
     pipeline.add_transform(|input, output| {
-        let transform = match params.aggregate_functions.is_empty() {
-            true => {
-                TransformFinalGroupBy::try_create(input, output, method.clone(), params.clone())?
-            }
-            false => {
-                TransformFinalAggregate::try_create(input, output, method.clone(), params.clone())?
-            }
-        };
-        if enable_profiling {
-            Ok(ProcessorPtr::create(ProcessorProfileWrapper::create(
-                transform,
-                prof_id,
-                proc_profs.clone(),
-            )))
-        } else {
-            Ok(ProcessorPtr::create(transform))
-        }
+        Ok(ProcessorPtr::create(
+            match params.aggregate_functions.is_empty() {
+                true => TransformFinalGroupBy::try_create(
+                    input,
+                    output,
+                    method.clone(),
+                    params.clone(),
+                )?,
+                false => TransformFinalAggregate::try_create(
+                    input,
+                    output,
+                    method.clone(),
+                    params.clone(),
+                )?,
+            },
+        ))
     })?;
-    // Append a profile stub to record the output rows and bytes
-    if enable_profiling {
-        pipeline.add_transform(|input, output| {
-            Ok(ProcessorPtr::create(Transformer::create(
-                input,
-                output,
-                ProfileStub::new(prof_id, proc_profs.clone())
-                    .accumulate_output_rows()
-                    .accumulate_output_bytes(),
-            )))
-        })?;
-    }
 
     Ok(())
 }
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/desc.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/desc.rs
index 3f7ea0c3ef1a..21244bb01fb5 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/desc.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/desc.rs
@@ -18,6 +18,7 @@ use databend_common_expression::Expr;
 use databend_common_expression::RemoteExpr;
 use databend_common_functions::BUILTIN_FUNCTIONS;
 use databend_common_sql::executor::physical_plans::HashJoin;
+use databend_common_sql::IndexType;
 use parking_lot::RwLock;
 
 use crate::sql::plans::JoinType;
@@ -35,6 +36,7 @@ pub struct HashJoinDesc {
     pub(crate) build_keys: Vec<Expr>,
     pub(crate) probe_keys: Vec<Expr>,
     pub(crate) join_type: JoinType,
+    pub(crate) single_to_inner: Option<JoinType>,
     /// when we have non-equal conditions for hash join,
     /// for example `a = b and c = d and e > f`, we will use `and_filters`
     /// to wrap `e > f` as a other_predicate to do next step's check.
@@ -42,9 +44,11 @@ pub struct HashJoinDesc {
     pub(crate) marker_join_desc: MarkJoinDesc,
     /// Whether the Join are derived from correlated subquery.
     pub(crate) from_correlated_subquery: bool,
-    pub(crate) probe_keys_rt: Vec<Option<Expr<String>>>,
+    pub(crate) probe_keys_rt: Vec<Option<(Expr<String>, IndexType)>>,
     // Under cluster, mark if the join is broadcast join.
     pub broadcast: bool,
+    // If enable bloom runtime filter
+    pub enable_bloom_runtime_filter: bool,
 }
 
 impl HashJoinDesc {
@@ -62,10 +66,14 @@ impl HashJoinDesc {
             .map(|k| k.as_expr(&BUILTIN_FUNCTIONS))
             .collect();
 
-        let probe_keys_rt: Vec<Option<Expr<String>>> = join
+        let probe_keys_rt: Vec<Option<(Expr<String>, IndexType)>> = join
             .probe_keys_rt
             .iter()
-            .map(|k| k.as_ref().map(|v| v.as_expr(&BUILTIN_FUNCTIONS)))
+            .map(|probe_key_rt| {
+                probe_key_rt
+                    .as_ref()
+                    .map(|(expr, idx)| (expr.as_expr(&BUILTIN_FUNCTIONS), *idx))
+            })
             .collect();
 
         Ok(HashJoinDesc {
@@ -80,6 +88,8 @@ impl HashJoinDesc {
             from_correlated_subquery: join.from_correlated_subquery,
             probe_keys_rt,
             broadcast: join.broadcast,
+            single_to_inner: join.single_to_inner.clone(),
+            enable_bloom_runtime_filter: join.enable_bloom_runtime_filter,
         })
     }
 
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_build_state.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_build_state.rs
index 34c3d2122e4b..39eda78b1fc3 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_build_state.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_build_state.rs
@@ -152,7 +152,9 @@ impl HashJoinBuildState {
             if !is_cluster || is_broadcast_join {
                 enable_inlist_runtime_filter = true;
                 enable_min_max_runtime_filter = true;
-                if ctx.get_settings().get_runtime_filter()? {
+                if ctx.get_settings().get_bloom_runtime_filter()?
+                    && hash_join_state.hash_join_desc.enable_bloom_runtime_filter
+                {
                     enable_bloom_runtime_filter = true;
                 }
             }
@@ -211,7 +213,11 @@ impl HashJoinBuildState {
 
     // Add `data_block` for build table to `row_space`
     pub(crate) fn add_build_block(&self, data_block: DataBlock) -> Result<()> {
-        let block_outer_scan_map = if self.hash_join_state.need_outer_scan() {
+        let block_outer_scan_map = if self.hash_join_state.need_outer_scan()
+            || matches!(
+                self.hash_join_state.hash_join_desc.single_to_inner,
+                Some(JoinType::RightSingle)
+            ) {
             vec![false; data_block.num_rows()]
         } else {
             vec![]
@@ -227,7 +233,12 @@ impl HashJoinBuildState {
             // Acquire lock in current scope
             let _lock = self.mutex.lock();
             let build_state = unsafe { &mut *self.hash_join_state.build_state.get() };
-            if self.hash_join_state.need_outer_scan() {
+            if self.hash_join_state.need_outer_scan()
+                || matches!(
+                    self.hash_join_state.hash_join_desc.single_to_inner,
+                    Some(JoinType::RightSingle)
+                )
+            {
                 build_state.outer_scan_map.push(block_outer_scan_map);
             }
             if self.hash_join_state.need_mark_scan() {
@@ -280,24 +291,7 @@ impl HashJoinBuildState {
                     .clone()
             };
 
-            let mut runtime_filter = RuntimeFilterInfo::default();
-            if self.enable_inlist_runtime_filter && build_num_rows < INLIST_RUNTIME_FILTER_THRESHOLD
-            {
-                self.inlist_runtime_filter(&mut runtime_filter, &build_chunks)?;
-            }
-            // If enable bloom runtime filter, collect hashes for build keys
-            if self.enable_bloom_runtime_filter {
-                self.bloom_runtime_filter(&self.func_ctx, &build_chunks, &mut runtime_filter)?;
-            }
-
-            if self.enable_min_max_runtime_filter {
-                self.min_max_runtime_filter(&self.func_ctx, &build_chunks, &mut runtime_filter)?;
-            }
-
-            if !runtime_filter.is_empty() {
-                self.ctx
-                    .set_runtime_filter((self.hash_join_state.table_index, runtime_filter));
-            }
+            self.add_runtime_filter(&build_chunks, build_num_rows)?;
 
             if self.hash_join_state.hash_join_desc.join_type == JoinType::Cross {
                 return Ok(());
@@ -797,58 +791,88 @@ impl HashJoinBuildState {
         Ok(())
     }
 
-    fn bloom_runtime_filter(
-        &self,
-        func_ctx: &FunctionContext,
-        data_blocks: &[DataBlock],
-        runtime_filter: &mut RuntimeFilterInfo,
-    ) -> Result<()> {
-        for (build_key, probe_key) in self
+    fn add_runtime_filter(&self, build_chunks: &[DataBlock], build_num_rows: usize) -> Result<()> {
+        for (build_key, probe_key, table_index) in self
             .hash_join_state
             .hash_join_desc
             .build_keys
             .iter()
             .zip(self.hash_join_state.hash_join_desc.probe_keys_rt.iter())
+            .filter_map(|(b, p)| p.as_ref().map(|(p, index)| (b, p, index)))
         {
-            if !build_key.data_type().remove_nullable().is_numeric()
-                && !build_key.data_type().remove_nullable().is_string()
+            let mut runtime_filter = RuntimeFilterInfo::default();
+            if self.enable_inlist_runtime_filter && build_num_rows < INLIST_RUNTIME_FILTER_THRESHOLD
             {
-                return Ok(());
+                self.inlist_runtime_filter(
+                    &mut runtime_filter,
+                    build_chunks,
+                    build_key,
+                    probe_key,
+                )?;
             }
-            if let Some(Expr::ColumnRef { id, .. }) = probe_key {
-                let mut columns = Vec::with_capacity(data_blocks.len());
-                for block in data_blocks.iter() {
-                    if block.num_columns() == 0 {
-                        continue;
-                    }
-                    let evaluator = Evaluator::new(block, func_ctx, &BUILTIN_FUNCTIONS);
-                    let column = evaluator
-                        .run(build_key)?
-                        .convert_to_full_column(build_key.data_type(), block.num_rows());
-                    columns.push(column);
-                }
-                if columns.is_empty() {
-                    return Ok(());
-                }
-                let build_key_column = Column::concat_columns(columns.into_iter())?;
-                // Generate bloom filter using build column
-                let data_type = build_key.data_type().clone();
-                let num_rows = build_key_column.len();
-                let method = DataBlock::choose_hash_method_with_types(&[data_type.clone()], false)?;
-                let mut hashes = HashSet::with_capacity(num_rows);
-                hash_by_method(
-                    &method,
-                    &[(build_key_column, data_type)],
-                    num_rows,
-                    &mut hashes,
+            if self.enable_bloom_runtime_filter {
+                self.bloom_runtime_filter(build_chunks, &mut runtime_filter, build_key, probe_key)?;
+            }
+            if self.enable_min_max_runtime_filter {
+                self.min_max_runtime_filter(
+                    build_chunks,
+                    &mut runtime_filter,
+                    build_key,
+                    probe_key,
                 )?;
-                let mut hashes_vec = Vec::with_capacity(num_rows);
-                hashes.into_iter().for_each(|hash| {
-                    hashes_vec.push(hash);
-                });
-                let filter = BinaryFuse16::try_from(&hashes_vec)?;
-                runtime_filter.add_bloom((id.to_string(), filter));
             }
+            if !runtime_filter.is_empty() {
+                self.ctx.set_runtime_filter((*table_index, runtime_filter));
+            }
+        }
+        Ok(())
+    }
+
+    fn bloom_runtime_filter(
+        &self,
+        data_blocks: &[DataBlock],
+        runtime_filter: &mut RuntimeFilterInfo,
+        build_key: &Expr,
+        probe_key: &Expr<String>,
+    ) -> Result<()> {
+        if !build_key.data_type().remove_nullable().is_numeric()
+            && !build_key.data_type().remove_nullable().is_string()
+        {
+            return Ok(());
+        }
+        if let Expr::ColumnRef { id, .. } = probe_key {
+            let mut columns = Vec::with_capacity(data_blocks.len());
+            for block in data_blocks.iter() {
+                if block.num_columns() == 0 {
+                    continue;
+                }
+                let evaluator = Evaluator::new(block, &self.func_ctx, &BUILTIN_FUNCTIONS);
+                let column = evaluator
+                    .run(build_key)?
+                    .convert_to_full_column(build_key.data_type(), block.num_rows());
+                columns.push(column);
+            }
+            if columns.is_empty() {
+                return Ok(());
+            }
+            let build_key_column = Column::concat_columns(columns.into_iter())?;
+            // Generate bloom filter using build column
+            let data_type = build_key.data_type().clone();
+            let num_rows = build_key_column.len();
+            let method = DataBlock::choose_hash_method_with_types(&[data_type.clone()], false)?;
+            let mut hashes = HashSet::with_capacity(num_rows);
+            hash_by_method(
+                &method,
+                &[(build_key_column, data_type)],
+                num_rows,
+                &mut hashes,
+            )?;
+            let mut hashes_vec = Vec::with_capacity(num_rows);
+            hashes.into_iter().for_each(|hash| {
+                hashes_vec.push(hash);
+            });
+            let filter = BinaryFuse16::try_from(&hashes_vec)?;
+            runtime_filter.add_bloom((id.to_string(), filter));
         }
         Ok(())
     }
@@ -857,22 +881,14 @@ impl HashJoinBuildState {
         &self,
         runtime_filter: &mut RuntimeFilterInfo,
         data_blocks: &[DataBlock],
+        build_key: &Expr,
+        probe_key: &Expr<String>,
     ) -> Result<()> {
-        for (build_key, probe_key) in self
-            .hash_join_state
-            .hash_join_desc
-            .build_keys
-            .iter()
-            .zip(self.hash_join_state.hash_join_desc.probe_keys_rt.iter())
+        if let Some(distinct_build_column) =
+            dedup_build_key_column(&self.func_ctx, data_blocks, build_key)?
         {
-            if let Some(distinct_build_column) =
-                dedup_build_key_column(&self.func_ctx, data_blocks, build_key)?
-            {
-                if let Some(probe_key) = probe_key {
-                    if let Some(filter) = inlist_filter(probe_key, distinct_build_column.clone())? {
-                        runtime_filter.add_inlist(filter);
-                    }
-                }
+            if let Some(filter) = inlist_filter(probe_key, distinct_build_column.clone())? {
+                runtime_filter.add_inlist(filter);
             }
         }
         Ok(())
@@ -880,107 +896,99 @@ impl HashJoinBuildState {
 
     fn min_max_runtime_filter(
         &self,
-        func_ctx: &FunctionContext,
         data_blocks: &[DataBlock],
         runtime_filter: &mut RuntimeFilterInfo,
+        build_key: &Expr,
+        probe_key: &Expr<String>,
     ) -> Result<()> {
-        for (build_key, probe_key) in self
-            .hash_join_state
-            .hash_join_desc
-            .build_keys
-            .iter()
-            .zip(self.hash_join_state.hash_join_desc.probe_keys_rt.iter())
-            .filter_map(|(b, p)| p.as_ref().map(|p| (b, p)))
+        if !build_key.data_type().remove_nullable().is_numeric()
+            && !build_key.data_type().remove_nullable().is_string()
         {
-            if !build_key.data_type().remove_nullable().is_numeric()
-                && !build_key.data_type().remove_nullable().is_string()
-            {
+            return Ok(());
+        }
+        if let Expr::ColumnRef { .. } = probe_key {
+            let mut columns = Vec::with_capacity(data_blocks.len());
+            for block in data_blocks.iter() {
+                if block.num_columns() == 0 {
+                    continue;
+                }
+                let evaluator = Evaluator::new(block, &self.func_ctx, &BUILTIN_FUNCTIONS);
+                let column = evaluator
+                    .run(build_key)?
+                    .convert_to_full_column(build_key.data_type(), block.num_rows());
+                columns.push(column);
+            }
+            if columns.is_empty() {
+                return Ok(());
+            }
+            let build_key_column = Column::concat_columns(columns.into_iter())?;
+            if build_key_column.len() == 0 {
                 return Ok(());
             }
-            if let Expr::ColumnRef { .. } = probe_key {
-                let mut columns = Vec::with_capacity(data_blocks.len());
-                for block in data_blocks.iter() {
-                    if block.num_columns() == 0 {
-                        continue;
+            // Generate min max filter using build column
+            let min_max = build_key_column.remove_nullable().domain();
+            let min_max_filter = match min_max {
+                Domain::Number(domain) => match domain {
+                    NumberDomain::UInt8(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
                     }
-                    let evaluator = Evaluator::new(block, func_ctx, &BUILTIN_FUNCTIONS);
-                    let column = evaluator
-                        .run(build_key)?
-                        .convert_to_full_column(build_key.data_type(), block.num_rows());
-                    columns.push(column);
-                }
-                if columns.is_empty() {
-                    return Ok(());
-                }
-                let build_key_column = Column::concat_columns(columns.into_iter())?;
-                if build_key_column.len() == 0 {
-                    return Ok(());
-                }
-                // Generate min max filter using build column
-                let min_max = build_key_column.remove_nullable().domain();
-                let min_max_filter = match min_max {
-                    Domain::Number(domain) => match domain {
-                        NumberDomain::UInt8(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::UInt16(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::UInt32(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::UInt64(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::Int8(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::Int16(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::Int32(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::Int64(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::Float32(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                        NumberDomain::Float64(simple_domain) => {
-                            let min = Scalar::Number(NumberScalar::from(simple_domain.min));
-                            let max = Scalar::Number(NumberScalar::from(simple_domain.max));
-                            min_max_filter(min, max, probe_key)?
-                        }
-                    },
-                    Domain::String(domain) => {
-                        let min = Scalar::String(domain.min);
-                        let max = Scalar::String(domain.max.unwrap());
+                    NumberDomain::UInt16(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
                         min_max_filter(min, max, probe_key)?
                     }
-                    _ => unreachable!(),
-                };
-                if let Some(min_max_filter) = min_max_filter {
-                    runtime_filter.add_min_max(min_max_filter);
+                    NumberDomain::UInt32(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::UInt64(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::Int8(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::Int16(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::Int32(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::Int64(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::Float32(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                    NumberDomain::Float64(simple_domain) => {
+                        let min = Scalar::Number(NumberScalar::from(simple_domain.min));
+                        let max = Scalar::Number(NumberScalar::from(simple_domain.max));
+                        min_max_filter(min, max, probe_key)?
+                    }
+                },
+                Domain::String(domain) => {
+                    let min = Scalar::String(domain.min);
+                    let max = Scalar::String(domain.max.unwrap());
+                    min_max_filter(min, max, probe_key)?
                 }
+                _ => unreachable!(),
+            };
+            if let Some(min_max_filter) = min_max_filter {
+                runtime_filter.add_min_max(min_max_filter);
             }
         }
         Ok(())
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_probe_state.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_probe_state.rs
index 534386ad8655..bfb74cf90240 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_probe_state.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_probe_state.rs
@@ -176,7 +176,7 @@ impl HashJoinProbeState {
 
     pub fn probe_join(
         &self,
-        input: DataBlock,
+        mut input: DataBlock,
         probe_state: &mut ProbeState,
     ) -> Result<Vec<DataBlock>> {
         let input_num_rows = input.num_rows();
@@ -259,7 +259,9 @@ impl HashJoinProbeState {
             *ty = ty.remove_nullable();
         }
 
-        let input = input.project(&self.probe_projections);
+        if self.hash_join_state.hash_join_desc.join_type != JoinType::LeftMark {
+            input = input.project(&self.probe_projections);
+        }
         probe_state.generation_state.is_probe_projected = input.num_columns() > 0;
 
         if self.hash_join_state.fast_return.load(Ordering::Relaxed)
@@ -289,7 +291,7 @@ impl HashJoinProbeState {
 
         // Probe:
         // (1) INNER / RIGHT / RIGHT SINGLE / RIGHT SEMI / RIGHT ANTI / RIGHT MARK / LEFT SEMI / LEFT MARK
-        //        prefer_early_filtering is true  => early_filtering_probe_with_selection
+        //        prefer_early_filtering is true  => early_filtering_matched_probe
         //        prefer_early_filtering is false => probe
         // (2) LEFT / LEFT SINGLE / LEFT ANTI / FULL
         //        prefer_early_filtering is true  => early_filtering_probe
@@ -306,36 +308,42 @@ impl HashJoinProbeState {
                     .build_keys_accessor_and_hashes(keys_state, &mut probe_state.hashes)?;
 
                 // Perform a round of hash table probe.
-                if Self::check_for_selection(&self.hash_join_state.hash_join_desc.join_type) {
-                    probe_state.selection_count = if prefer_early_filtering {
+                probe_state.probe_with_selection = prefer_early_filtering;
+                probe_state.selection_count = if !Self::need_unmatched_selection(
+                    &self.hash_join_state.hash_join_desc.join_type,
+                    probe_state.with_conjunction,
+                ) {
+                    if prefer_early_filtering {
                         // Early filtering, use selection to get better performance.
-                        probe_state.probe_with_selection = true;
-
-                        table.hash_table.early_filtering_probe_with_selection(
+                        table.hash_table.early_filtering_matched_probe(
                             &mut probe_state.hashes,
                             valids,
                             &mut probe_state.selection,
                         )
                     } else {
                         // If don't do early filtering, don't use selection.
-                        probe_state.probe_with_selection = false;
-
                         table.hash_table.probe(&mut probe_state.hashes, valids)
-                    };
-                    probe_state.num_keys_hash_matched += probe_state.selection_count as u64;
+                    }
                 } else {
-                    // For left join, left single join, full join and left anti join, don't use selection.
-                    probe_state.probe_with_selection = false;
-
-                    let count = if prefer_early_filtering {
-                        table
-                            .hash_table
-                            .early_filtering_probe(&mut probe_state.hashes, valids)
+                    if prefer_early_filtering {
+                        // Early filtering, use matched selection and unmatched selection to get better performance.
+                        let unmatched_selection =
+                            probe_state.probe_unmatched_indexes.as_mut().unwrap();
+                        let (matched_count, unmatched_count) =
+                            table.hash_table.early_filtering_probe(
+                                &mut probe_state.hashes,
+                                valids,
+                                &mut probe_state.selection,
+                                unmatched_selection,
+                            );
+                        probe_state.probe_unmatched_indexes_count = unmatched_count;
+                        matched_count
                     } else {
+                        // If don't do early filtering, don't use selection.
                         table.hash_table.probe(&mut probe_state.hashes, valids)
-                    };
-                    probe_state.num_keys_hash_matched += count as u64;
-                }
+                    }
+                };
+                probe_state.num_keys_hash_matched += probe_state.selection_count as u64;
 
                 // Continue to probe hash table and process data blocks.
                 self.result_blocks(&input, keys, &table.hash_table, probe_state)
@@ -367,19 +375,12 @@ impl HashJoinProbeState {
         )
     }
 
-    /// Checks if a join type can use selection.
-    pub fn check_for_selection(join_type: &JoinType) -> bool {
+    /// Checks if the join type need to use unmatched selection.
+    pub fn need_unmatched_selection(join_type: &JoinType, with_conjunction: bool) -> bool {
         matches!(
             join_type,
-            JoinType::Inner
-                | JoinType::Right
-                | JoinType::RightSingle
-                | JoinType::RightSemi
-                | JoinType::RightAnti
-                | JoinType::RightMark
-                | JoinType::LeftSemi
-                | JoinType::LeftMark
-        )
+            JoinType::Left | JoinType::LeftSingle | JoinType::Full | JoinType::LeftAnti
+        ) && !with_conjunction
     }
 
     pub fn probe_attach(&self) -> Result<usize> {
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_state.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_state.rs
index 706de842b75b..47268c34e97f 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_state.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/hash_join_state.rs
@@ -120,9 +120,6 @@ pub struct HashJoinState {
     pub(crate) partition_id: AtomicI8,
     pub(crate) enable_spill: bool,
 
-    /// If the join node generate runtime filters, the scan node will use it to do prune.
-    pub(crate) table_index: IndexType,
-
     pub(crate) merge_into_state: Option<SyncUnsafeCell<MergeIntoState>>,
 }
 
@@ -133,7 +130,6 @@ impl HashJoinState {
         build_projections: &ColumnSet,
         hash_join_desc: HashJoinDesc,
         probe_to_build: &[(usize, (bool, bool))],
-        table_index: IndexType,
         merge_into_target_table_index: IndexType,
         merge_into_is_distributed: bool,
     ) -> Result<Arc<HashJoinState>> {
@@ -170,7 +166,6 @@ impl HashJoinState {
             _continue_build_dummy_receiver,
             partition_id: AtomicI8::new(-2),
             enable_spill,
-            table_index,
             merge_into_state: MergeIntoState::try_create_merge_into_state(
                 merge_into_target_table_index,
                 merge_into_is_distributed,
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/inner_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/inner_join.rs
index 0b6245fc1311..9885931245a1 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/inner_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/inner_join.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::sync::atomic::AtomicBool;
 use std::sync::atomic::Ordering;
 
 use databend_common_exception::ErrorCode;
@@ -33,7 +34,12 @@ use crate::pipelines::processors::transforms::hash_join::HashJoinProbeState;
 use crate::pipelines::processors::transforms::hash_join::ProbeState;
 
 impl HashJoinProbeState {
-    pub(crate) fn inner_join<'a, H: HashJoinHashtableLike>(
+    pub(crate) fn inner_join<
+        'a,
+        H: HashJoinHashtableLike,
+        const FROM_LEFT_SINGLE: bool,
+        const FROM_RIGHT_SINGLE: bool,
+    >(
         &self,
         input: &DataBlock,
         keys: Box<(dyn KeyAccessor<Key = H::Key>)>,
@@ -52,7 +58,18 @@ impl HashJoinProbeState {
         let pointers = probe_state.hashes.as_slice();
 
         // Build states.
-        let build_state = unsafe { &*self.hash_join_state.build_state.get() };
+        let build_state = unsafe { &mut *self.hash_join_state.build_state.get() };
+        let outer_scan_map = &mut build_state.outer_scan_map;
+        let mut right_single_scan_map = if FROM_RIGHT_SINGLE {
+            outer_scan_map
+                .iter_mut()
+                .map(|sp| unsafe {
+                    std::mem::transmute::<*mut bool, *mut AtomicBool>(sp.as_mut_ptr())
+                })
+                .collect::<Vec<_>>()
+        } else {
+            vec![]
+        };
 
         // Results.
         let mut matched_idx = 0;
@@ -66,12 +83,18 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
                 }
 
+                if FROM_LEFT_SINGLE && match_count > 1 {
+                    return Err(ErrorCode::Internal(
+                        "Scalar subquery can't return more than one row",
+                    ));
+                }
+
                 // Fill `probe_indexes`.
                 for _ in 0..match_count {
                     unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
@@ -79,26 +102,25 @@ impl HashJoinProbeState {
                 }
 
                 while matched_idx == max_block_size {
-                    result_blocks.push(self.process_inner_join_block(
+                    result_blocks.push(self.process_inner_join_block::<FROM_RIGHT_SINGLE>(
                         matched_idx,
                         input,
                         probe_indexes,
                         build_indexes,
                         &mut probe_state.generation_state,
                         &build_state.generation_state,
+                        &mut right_single_scan_map,
                     )?);
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
-                        key,
-                        incomplete_ptr,
-                        build_indexes_ptr,
-                        matched_idx,
-                        max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
-                        matched_idx += 1;
-                    }
+                    (matched_idx, incomplete_ptr) = self
+                        .fill_probe_and_build_indexes::<_, FROM_LEFT_SINGLE>(
+                            hash_table,
+                            key,
+                            incomplete_ptr,
+                            *idx,
+                            probe_indexes,
+                            build_indexes_ptr,
+                            max_block_size,
+                        )?;
                 }
             }
         } else {
@@ -107,12 +129,18 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
                 }
 
+                if FROM_LEFT_SINGLE && match_count > 1 {
+                    return Err(ErrorCode::Internal(
+                        "Scalar subquery can't return more than one row",
+                    ));
+                }
+
                 // Fill `probe_indexes`.
                 for _ in 0..match_count {
                     unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
@@ -120,38 +148,38 @@ impl HashJoinProbeState {
                 }
 
                 while matched_idx == max_block_size {
-                    result_blocks.push(self.process_inner_join_block(
+                    result_blocks.push(self.process_inner_join_block::<FROM_RIGHT_SINGLE>(
                         matched_idx,
                         input,
                         probe_indexes,
                         build_indexes,
                         &mut probe_state.generation_state,
                         &build_state.generation_state,
+                        &mut right_single_scan_map,
                     )?);
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
-                        key,
-                        incomplete_ptr,
-                        build_indexes_ptr,
-                        matched_idx,
-                        max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                        matched_idx += 1;
-                    }
+                    (matched_idx, incomplete_ptr) = self
+                        .fill_probe_and_build_indexes::<_, FROM_LEFT_SINGLE>(
+                            hash_table,
+                            key,
+                            incomplete_ptr,
+                            idx as u32,
+                            probe_indexes,
+                            build_indexes_ptr,
+                            max_block_size,
+                        )?;
                 }
             }
         }
 
         if matched_idx > 0 {
-            result_blocks.push(self.process_inner_join_block(
+            result_blocks.push(self.process_inner_join_block::<FROM_RIGHT_SINGLE>(
                 matched_idx,
                 input,
                 probe_indexes,
                 build_indexes,
                 &mut probe_state.generation_state,
                 &build_state.generation_state,
+                &mut right_single_scan_map,
             )?);
         }
 
@@ -188,7 +216,7 @@ impl HashJoinProbeState {
     }
 
     #[inline]
-    fn process_inner_join_block(
+    fn process_inner_join_block<const FROM_RIGHT_SINGLE: bool>(
         &self,
         matched_idx: usize,
         input: &DataBlock,
@@ -196,6 +224,7 @@ impl HashJoinProbeState {
         build_indexes: &[RowPtr],
         probe_state: &mut ProbeBlockGenerationState,
         build_state: &BuildBlockGenerationState,
+        right_single_scan_map: &mut [*mut AtomicBool],
     ) -> Result<DataBlock> {
         if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
             return Err(ErrorCode::AbortedQuery(
@@ -242,6 +271,53 @@ impl HashJoinProbeState {
                 result_block.add_column(entry);
             }
         }
+
+        if FROM_RIGHT_SINGLE {
+            self.update_right_single_scan_map(
+                &build_indexes[0..matched_idx],
+                right_single_scan_map,
+                None,
+            )?;
+        }
+
         Ok(result_block)
     }
+
+    #[inline(always)]
+    #[allow(clippy::too_many_arguments)]
+    pub(crate) fn fill_probe_and_build_indexes<
+        'a,
+        H: HashJoinHashtableLike,
+        const FROM_LEFT_SINGLE: bool,
+    >(
+        &self,
+        hash_table: &H,
+        key: &H::Key,
+        incomplete_ptr: u64,
+        idx: u32,
+        probe_indexes: &mut [u32],
+        build_indexes_ptr: *mut RowPtr,
+        max_block_size: usize,
+    ) -> Result<(usize, u64)>
+    where
+        H::Key: 'a,
+    {
+        let (match_count, ptr) =
+            hash_table.next_probe(key, incomplete_ptr, build_indexes_ptr, 0, max_block_size);
+        if match_count == 0 {
+            return Ok((0, 0));
+        }
+
+        if FROM_LEFT_SINGLE {
+            return Err(ErrorCode::Internal(
+                "Scalar subquery can't return more than one row",
+            ));
+        }
+
+        for i in 0..match_count {
+            unsafe { *probe_indexes.get_unchecked_mut(i) = idx };
+        }
+
+        Ok((match_count, ptr))
+    }
 }
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_anti_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_anti_join.rs
index caea33882c97..9a6e709a7307 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_anti_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_anti_join.rs
@@ -39,23 +39,37 @@ impl HashJoinProbeState {
         H::Key: 'a,
     {
         // Probe states.
-        let mutable_indexes = &mut probe_state.mutable_indexes;
-        let probe_indexes = &mut mutable_indexes.probe_indexes;
         let pointers = probe_state.hashes.as_slice();
 
-        // Results.
-        let mut matched_idx = 0;
-        let mut result_blocks = vec![];
-
         // Probe hash table and generate data blocks.
-        for idx in 0..input.num_rows() {
-            let key = unsafe { keys.key_unchecked(idx) };
-            let ptr = unsafe { *pointers.get_unchecked(idx) };
-            if !hash_table.next_contains(key, ptr) {
-                unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                matched_idx += 1;
+        let (probe_indexes, count) = if probe_state.probe_with_selection {
+            // Safe to unwrap.
+            let probe_unmatched_indexes = probe_state.probe_unmatched_indexes.as_mut().unwrap();
+            let mut unmatched_idx = probe_state.probe_unmatched_indexes_count;
+            let selection = &probe_state.selection.as_slice()[0..probe_state.selection_count];
+            for idx in selection.iter() {
+                let key = unsafe { keys.key_unchecked(*idx as usize) };
+                let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
+                if !hash_table.next_contains(key, ptr) {
+                    unsafe { *probe_unmatched_indexes.get_unchecked_mut(unmatched_idx) = *idx };
+                    unmatched_idx += 1;
+                }
             }
-        }
+            (probe_unmatched_indexes, unmatched_idx)
+        } else {
+            let mutable_indexes = &mut probe_state.mutable_indexes;
+            let probe_indexes = &mut mutable_indexes.probe_indexes;
+            let mut unmatched_idx = 0;
+            for idx in 0..input.num_rows() {
+                let key = unsafe { keys.key_unchecked(idx) };
+                let ptr = unsafe { *pointers.get_unchecked(idx) };
+                if !hash_table.next_contains(key, ptr) {
+                    unsafe { *probe_indexes.get_unchecked_mut(unmatched_idx) = idx as u32 };
+                    unmatched_idx += 1;
+                }
+            }
+            (probe_indexes, unmatched_idx)
+        };
 
         if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
             return Err(ErrorCode::AbortedQuery(
@@ -63,15 +77,16 @@ impl HashJoinProbeState {
             ));
         }
 
-        if matched_idx > 0 {
-            result_blocks.push(DataBlock::take(
-                input,
-                &probe_indexes[0..matched_idx],
-                &mut probe_state.generation_state.string_items_buf,
-            )?);
+        let result_block = DataBlock::take(
+            input,
+            &probe_indexes[0..count],
+            &mut probe_state.generation_state.string_items_buf,
+        )?;
+        if result_block.is_empty() {
+            Ok(vec![])
+        } else {
+            Ok(vec![result_block])
         }
-
-        Ok(result_blocks)
     }
 
     pub(crate) fn left_anti_join_with_conjunct<'a, H: HashJoinHashtableLike>(
@@ -109,47 +124,88 @@ impl HashJoinProbeState {
         let mut result_blocks = vec![];
 
         // Probe hash table and generate data blocks.
-        for idx in 0..input.num_rows() {
-            let key = unsafe { keys.key_unchecked(idx) };
-            let ptr = unsafe { *pointers.get_unchecked(idx) };
+        if probe_state.probe_with_selection {
+            let selection = &probe_state.selection.as_slice()[0..probe_state.selection_count];
+            for idx in selection.iter() {
+                let key = unsafe { keys.key_unchecked(*idx as usize) };
+                let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
-            // Probe hash table and fill `build_indexes`.
-            let (mut match_count, mut incomplete_ptr) =
-                hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
+                // Probe hash table and fill `build_indexes`.
+                let (match_count, mut incomplete_ptr) =
+                    hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
 
-            if match_count == 0 {
-                continue;
-            }
+                if match_count == 0 {
+                    continue;
+                }
+
+                // Fill `probe_indexes`.
+                for _ in 0..match_count {
+                    unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
+                    matched_idx += 1;
+                }
 
-            // Fill `probe_indexes`.
-            for _ in 0..match_count {
-                unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                matched_idx += 1;
+                while matched_idx == max_block_size {
+                    self.process_left_anti_join_block(
+                        matched_idx,
+                        input,
+                        probe_indexes,
+                        build_indexes,
+                        &mut probe_state.generation_state,
+                        &build_state.generation_state,
+                        other_predicate,
+                        &mut row_state,
+                    )?;
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
+                        key,
+                        incomplete_ptr,
+                        *idx,
+                        probe_indexes,
+                        build_indexes_ptr,
+                        max_block_size,
+                    )?;
+                }
             }
+        } else {
+            for idx in 0..input.num_rows() {
+                let key = unsafe { keys.key_unchecked(idx) };
+                let ptr = unsafe { *pointers.get_unchecked(idx) };
+
+                // Probe hash table and fill `build_indexes`.
+                let (match_count, mut incomplete_ptr) =
+                    hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
+
+                if match_count == 0 {
+                    continue;
+                }
 
-            while matched_idx == max_block_size {
-                self.process_left_anti_join_block(
-                    matched_idx,
-                    input,
-                    probe_indexes,
-                    build_indexes,
-                    &mut probe_state.generation_state,
-                    &build_state.generation_state,
-                    other_predicate,
-                    &mut row_state,
-                )?;
-                matched_idx = 0;
-                (match_count, incomplete_ptr) = hash_table.next_probe(
-                    key,
-                    incomplete_ptr,
-                    build_indexes_ptr,
-                    matched_idx,
-                    max_block_size,
-                );
+                // Fill `probe_indexes`.
                 for _ in 0..match_count {
                     unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
                     matched_idx += 1;
                 }
+
+                while matched_idx == max_block_size {
+                    self.process_left_anti_join_block(
+                        matched_idx,
+                        input,
+                        probe_indexes,
+                        build_indexes,
+                        &mut probe_state.generation_state,
+                        &build_state.generation_state,
+                        other_predicate,
+                        &mut row_state,
+                    )?;
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
+                        key,
+                        incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
+                        build_indexes_ptr,
+                        max_block_size,
+                    )?;
+                }
             }
         }
 
@@ -167,17 +223,17 @@ impl HashJoinProbeState {
         }
 
         // Find all unmatched indexes and generate the result `DataBlock`.
-        matched_idx = 0;
+        let mut unmatched_idx = 0;
         for (i, state) in row_state.iter().enumerate() {
             if !*state {
-                unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = i as u32 };
-                matched_idx += 1;
+                unsafe { *probe_indexes.get_unchecked_mut(unmatched_idx) = i as u32 };
+                unmatched_idx += 1;
             }
         }
-        if matched_idx > 0 {
+        if unmatched_idx > 0 {
             result_blocks.push(DataBlock::take(
                 input,
-                &probe_indexes[0..matched_idx],
+                &probe_indexes[0..unmatched_idx],
                 &mut probe_state.generation_state.string_items_buf,
             )?);
         }
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_join.rs
index bdf26a1dc984..1335e4fd8877 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_join.rs
@@ -33,7 +33,7 @@ use crate::pipelines::processors::transforms::hash_join::ProbeState;
 use crate::sql::plans::JoinType;
 
 impl HashJoinProbeState {
-    pub(crate) fn left_join<'a, H: HashJoinHashtableLike>(
+    pub(crate) fn left_join<'a, H: HashJoinHashtableLike, const LEFT_SINGLE: bool>(
         &self,
         input: &DataBlock,
         keys: Box<(dyn KeyAccessor<Key = H::Key>)>,
@@ -63,69 +63,114 @@ impl HashJoinProbeState {
         let mut unmatched_idx = 0;
         let mut result_blocks = vec![];
 
-        // Probe hash table and generate data blocks.
-        for idx in 0..input_rows {
-            let key = unsafe { keys.key_unchecked(idx) };
-            let ptr = unsafe { *pointers.get_unchecked(idx) };
-
-            // Probe hash table and fill `build_indexes`.
-            let (mut match_count, mut incomplete_ptr) =
-                hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
-
-            let mut total_probe_matched = 0;
-            if match_count > 0 {
-                total_probe_matched += match_count;
-                if self.hash_join_state.hash_join_desc.join_type == JoinType::LeftSingle
-                    && total_probe_matched > 1
-                {
-                    return Err(ErrorCode::Internal(
-                        "Scalar subquery can't return more than one row",
-                    ));
+        if probe_state.probe_with_selection {
+            unmatched_idx = probe_state.probe_unmatched_indexes_count;
+            let selection = &probe_state.selection.as_slice()[0..probe_state.selection_count];
+            for idx in selection.iter() {
+                let key = unsafe { keys.key_unchecked(*idx as usize) };
+                let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
+                // Probe hash table and fill `build_indexes`.
+                let (match_count, mut incomplete_ptr) =
+                    hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
+
+                if match_count > 0 {
+                    if LEFT_SINGLE && match_count > 1 {
+                        return Err(ErrorCode::Internal(
+                            "Scalar subquery can't return more than one row",
+                        ));
+                    }
+
+                    for _ in 0..match_count {
+                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
+                        matched_idx += 1;
+                    }
+                } else {
+                    unsafe { *probe_unmatched_indexes.get_unchecked_mut(unmatched_idx) = *idx };
+                    unmatched_idx += 1;
                 }
-                for _ in 0..match_count {
-                    unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                    matched_idx += 1;
+
+                while matched_idx == max_block_size {
+                    self.process_left_or_full_join_block(
+                        matched_idx,
+                        input,
+                        probe_indexes,
+                        build_indexes,
+                        &mut probe_state.generation_state,
+                        &build_state.generation_state,
+                        outer_scan_map,
+                        &mut result_blocks,
+                        None,
+                        None,
+                        None,
+                    )?;
+                    (matched_idx, incomplete_ptr) = self.fill_left_outer_states::<_, LEFT_SINGLE>(
+                        hash_table,
+                        key,
+                        incomplete_ptr,
+                        *idx,
+                        probe_indexes,
+                        build_indexes_ptr,
+                        max_block_size,
+                        false,
+                        None,
+                        None,
+                    )?;
                 }
-            } else {
-                unsafe { *probe_unmatched_indexes.get_unchecked_mut(unmatched_idx) = idx as u32 };
-                unmatched_idx += 1;
             }
+        } else {
+            // Probe hash table and generate data blocks.
+            for idx in 0..input_rows {
+                let key = unsafe { keys.key_unchecked(idx) };
+                let ptr = unsafe { *pointers.get_unchecked(idx) };
+
+                // Probe hash table and fill `build_indexes`.
+                let (match_count, mut incomplete_ptr) =
+                    hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
 
-            while matched_idx == max_block_size {
-                self.process_left_or_full_join_block(
-                    matched_idx,
-                    input,
-                    probe_indexes,
-                    build_indexes,
-                    &mut probe_state.generation_state,
-                    &build_state.generation_state,
-                    outer_scan_map,
-                    &mut result_blocks,
-                    None,
-                    None,
-                    None,
-                )?;
-                matched_idx = 0;
-                (match_count, incomplete_ptr) = hash_table.next_probe(
-                    key,
-                    incomplete_ptr,
-                    build_indexes_ptr,
-                    matched_idx,
-                    max_block_size,
-                );
                 if match_count > 0 {
-                    total_probe_matched += match_count;
-                    if self.hash_join_state.hash_join_desc.join_type == JoinType::LeftSingle
-                        && total_probe_matched > 1
-                    {
+                    if LEFT_SINGLE && match_count > 1 {
                         return Err(ErrorCode::Internal(
                             "Scalar subquery can't return more than one row",
                         ));
                     }
+
                     for _ in 0..match_count {
                         unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
                         matched_idx += 1;
                     }
+                } else {
+                    unsafe {
+                        *probe_unmatched_indexes.get_unchecked_mut(unmatched_idx) = idx as u32
+                    };
+                    unmatched_idx += 1;
+                }
+
+                while matched_idx == max_block_size {
+                    self.process_left_or_full_join_block(
+                        matched_idx,
+                        input,
+                        probe_indexes,
+                        build_indexes,
+                        &mut probe_state.generation_state,
+                        &build_state.generation_state,
+                        outer_scan_map,
+                        &mut result_blocks,
+                        None,
+                        None,
+                        None,
+                    )?;
+                    (matched_idx, incomplete_ptr) = self.fill_left_outer_states::<_, LEFT_SINGLE>(
+                        hash_table,
+                        key,
+                        incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
+                        build_indexes_ptr,
+                        max_block_size,
+                        false,
+                        None,
+                        None,
+                    )?;
                 }
             }
         }
@@ -159,7 +204,7 @@ impl HashJoinProbeState {
         Ok(result_blocks)
     }
 
-    pub(crate) fn left_join_with_conjunct<'a, H: HashJoinHashtableLike>(
+    pub(crate) fn left_join_with_conjunct<'a, H: HashJoinHashtableLike, const LEFT_SINGLE: bool>(
         &self,
         input: &DataBlock,
         keys: Box<(dyn KeyAccessor<Key = H::Key>)>,
@@ -200,62 +245,72 @@ impl HashJoinProbeState {
         let mut result_blocks = vec![];
 
         // Probe hash table and generate data blocks.
-        for idx in 0..input_rows {
-            let key = unsafe { keys.key_unchecked(idx) };
-            let ptr = unsafe { *pointers.get_unchecked(idx) };
-
-            // Probe hash table and fill `build_indexes`.
-            let (mut match_count, mut incomplete_ptr) =
-                hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
-            // `total_probe_matched` is used to record the matched rows count for current `idx` row from probe_block
-            let mut total_probe_matched = 0;
-            if match_count > 0 {
-                total_probe_matched += match_count;
-                if self.hash_join_state.hash_join_desc.join_type == JoinType::LeftSingle
-                    && total_probe_matched > 1
-                {
-                    return Err(ErrorCode::Internal(
-                        "Scalar subquery can't return more than one row",
-                    ));
-                }
+        if probe_state.probe_with_selection {
+            let selection = &probe_state.selection.as_slice()[0..probe_state.selection_count];
+            for idx in selection.iter() {
+                let key = unsafe { keys.key_unchecked(*idx as usize) };
+                let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
-                unsafe {
-                    *row_state.get_unchecked_mut(idx) += match_count;
-                    for _ in 0..match_count {
-                        *row_state_indexes.get_unchecked_mut(matched_idx) = idx;
-                        *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32;
-                        matched_idx += 1;
+                // Probe hash table and fill `build_indexes`.
+                let (match_count, mut incomplete_ptr) =
+                    hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
+
+                if match_count > 0 {
+                    if LEFT_SINGLE && match_count > 1 {
+                        return Err(ErrorCode::Internal(
+                            "Scalar subquery can't return more than one row",
+                        ));
+                    }
+
+                    unsafe {
+                        *row_state.get_unchecked_mut(*idx as usize) += match_count;
+                        for _ in 0..match_count {
+                            *row_state_indexes.get_unchecked_mut(matched_idx) = *idx as usize;
+                            *probe_indexes.get_unchecked_mut(matched_idx) = *idx;
+                            matched_idx += 1;
+                        }
                     }
                 }
+
+                while matched_idx == max_block_size {
+                    self.process_left_or_full_join_block(
+                        matched_idx,
+                        input,
+                        probe_indexes,
+                        build_indexes,
+                        &mut probe_state.generation_state,
+                        &build_state.generation_state,
+                        outer_scan_map,
+                        &mut result_blocks,
+                        Some(other_predicate),
+                        Some(row_state),
+                        Some(row_state_indexes),
+                    )?;
+                    (matched_idx, incomplete_ptr) = self.fill_left_outer_states::<_, LEFT_SINGLE>(
+                        hash_table,
+                        key,
+                        incomplete_ptr,
+                        *idx,
+                        probe_indexes,
+                        build_indexes_ptr,
+                        max_block_size,
+                        true,
+                        Some(row_state),
+                        Some(row_state_indexes),
+                    )?;
+                }
             }
+        } else {
+            for idx in 0..input_rows {
+                let key = unsafe { keys.key_unchecked(idx) };
+                let ptr = unsafe { *pointers.get_unchecked(idx) };
+
+                // Probe hash table and fill `build_indexes`.
+                let (match_count, mut incomplete_ptr) =
+                    hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
 
-            while matched_idx == max_block_size {
-                self.process_left_or_full_join_block(
-                    matched_idx,
-                    input,
-                    probe_indexes,
-                    build_indexes,
-                    &mut probe_state.generation_state,
-                    &build_state.generation_state,
-                    outer_scan_map,
-                    &mut result_blocks,
-                    Some(other_predicate),
-                    Some(row_state),
-                    Some(row_state_indexes),
-                )?;
-                matched_idx = 0;
-                (match_count, incomplete_ptr) = hash_table.next_probe(
-                    key,
-                    incomplete_ptr,
-                    build_indexes_ptr,
-                    matched_idx,
-                    max_block_size,
-                );
                 if match_count > 0 {
-                    total_probe_matched += match_count;
-                    if self.hash_join_state.hash_join_desc.join_type == JoinType::LeftSingle
-                        && total_probe_matched > 1
-                    {
+                    if LEFT_SINGLE && match_count > 1 {
                         return Err(ErrorCode::Internal(
                             "Scalar subquery can't return more than one row",
                         ));
@@ -270,6 +325,34 @@ impl HashJoinProbeState {
                         }
                     }
                 }
+
+                while matched_idx == max_block_size {
+                    self.process_left_or_full_join_block(
+                        matched_idx,
+                        input,
+                        probe_indexes,
+                        build_indexes,
+                        &mut probe_state.generation_state,
+                        &build_state.generation_state,
+                        outer_scan_map,
+                        &mut result_blocks,
+                        Some(other_predicate),
+                        Some(row_state),
+                        Some(row_state_indexes),
+                    )?;
+                    (matched_idx, incomplete_ptr) = self.fill_left_outer_states::<_, LEFT_SINGLE>(
+                        hash_table,
+                        key,
+                        incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
+                        build_indexes_ptr,
+                        max_block_size,
+                        true,
+                        Some(row_state),
+                        Some(row_state_indexes),
+                    )?;
+                }
             }
         }
 
@@ -538,4 +621,53 @@ impl HashJoinProbeState {
 
         Ok(())
     }
+
+    #[inline(always)]
+    #[allow(clippy::too_many_arguments)]
+    fn fill_left_outer_states<'a, H: HashJoinHashtableLike, const LEFT_SINGLE: bool>(
+        &self,
+        hash_table: &H,
+        key: &H::Key,
+        incomplete_ptr: u64,
+        idx: u32,
+        probe_indexes: &mut [u32],
+        build_indexes_ptr: *mut RowPtr,
+        max_block_size: usize,
+        with_conjunct: bool,
+        row_state: Option<&mut Vec<usize>>,
+        row_state_indexes: Option<&mut Vec<usize>>,
+    ) -> Result<(usize, u64)>
+    where
+        H::Key: 'a,
+    {
+        let (match_count, ptr) =
+            hash_table.next_probe(key, incomplete_ptr, build_indexes_ptr, 0, max_block_size);
+        if match_count == 0 {
+            return Ok((0, 0));
+        }
+
+        if LEFT_SINGLE {
+            return Err(ErrorCode::Internal(
+                "Scalar subquery can't return more than one row",
+            ));
+        }
+
+        if !with_conjunct {
+            for i in 0..match_count {
+                unsafe { *probe_indexes.get_unchecked_mut(i) = idx };
+            }
+        } else {
+            let row_state = row_state.unwrap();
+            let row_state_indexes = row_state_indexes.unwrap();
+            unsafe {
+                *row_state.get_unchecked_mut(idx as usize) += match_count;
+                for i in 0..match_count {
+                    *row_state_indexes.get_unchecked_mut(i) = idx as usize;
+                    *probe_indexes.get_unchecked_mut(i) = idx;
+                }
+            }
+        }
+
+        Ok((match_count, ptr))
+    }
 }
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_mark_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_mark_join.rs
index 163c77290888..6dc6498cd47a 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_mark_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_mark_join.rs
@@ -59,6 +59,7 @@ impl HashJoinProbeState {
                 .write();
             *has_null = true;
         }
+
         let mutable_indexes = &mut probe_state.mutable_indexes;
         let build_indexes = &mut mutable_indexes.build_indexes;
         let build_indexes_ptr = build_indexes.as_mut_ptr();
@@ -80,7 +81,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -89,28 +90,14 @@ impl HashJoinProbeState {
                 matched_idx += match_count;
 
                 while matched_idx == max_block_size {
-                    if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
-                        return Err(ErrorCode::AbortedQuery(
-                            "Aborted query, because the server is shutting down or the query was killed.",
-                        ));
-                    }
-                    for probed_row in build_indexes.iter() {
-                        unsafe {
-                            *mark_scan_map
-                                .get_unchecked_mut(probed_row.chunk_index as usize)
-                                .get_unchecked_mut(probed_row.row_index as usize) =
-                                MARKER_KIND_TRUE;
-                        }
-                    }
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    self.process_left_mark_join_block(build_indexes, mark_scan_map)?;
+                    (matched_idx, incomplete_ptr) = hash_table.next_probe(
                         key,
                         incomplete_ptr,
                         build_indexes_ptr,
-                        matched_idx,
+                        0,
                         max_block_size,
                     );
-                    matched_idx += match_count;
                 }
             }
         } else {
@@ -119,7 +106,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill build_indexes.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -128,32 +115,22 @@ impl HashJoinProbeState {
                 matched_idx += match_count;
 
                 while matched_idx == max_block_size {
-                    if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
-                        return Err(ErrorCode::AbortedQuery(
-                            "Aborted query, because the server is shutting down or the query was killed.",
-                        ));
-                    }
-                    for probed_row in build_indexes.iter() {
-                        unsafe {
-                            *mark_scan_map
-                                .get_unchecked_mut(probed_row.chunk_index as usize)
-                                .get_unchecked_mut(probed_row.row_index as usize) =
-                                MARKER_KIND_TRUE;
-                        }
-                    }
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    self.process_left_mark_join_block(build_indexes, mark_scan_map)?;
+                    (matched_idx, incomplete_ptr) = hash_table.next_probe(
                         key,
                         incomplete_ptr,
                         build_indexes_ptr,
-                        matched_idx,
+                        0,
                         max_block_size,
                     );
-                    matched_idx += match_count;
                 }
             }
         }
 
+        if matched_idx > 0 {
+            self.process_left_mark_join_block(&build_indexes[0..matched_idx], mark_scan_map)?;
+        }
+
         Ok(vec![])
     }
 
@@ -210,7 +187,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -223,7 +200,7 @@ impl HashJoinProbeState {
                 }
 
                 while matched_idx == max_block_size {
-                    self.process_left_mark_join_block(
+                    self.process_left_mark_join_with_conjunct_block(
                         matched_idx,
                         input,
                         probe_indexes,
@@ -233,18 +210,15 @@ impl HashJoinProbeState {
                         other_predicate,
                         mark_scan_map,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        *idx,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         } else {
@@ -253,7 +227,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill build_indexes.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -266,7 +240,7 @@ impl HashJoinProbeState {
                 }
 
                 while matched_idx == max_block_size {
-                    self.process_left_mark_join_block(
+                    self.process_left_mark_join_with_conjunct_block(
                         matched_idx,
                         input,
                         probe_indexes,
@@ -276,18 +250,15 @@ impl HashJoinProbeState {
                         other_predicate,
                         mark_scan_map,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         }
@@ -305,7 +276,7 @@ impl HashJoinProbeState {
         }
 
         if matched_idx > 0 {
-            self.process_left_mark_join_block(
+            self.process_left_mark_join_with_conjunct_block(
                 matched_idx,
                 input,
                 probe_indexes,
@@ -323,6 +294,30 @@ impl HashJoinProbeState {
     #[inline]
     #[allow(clippy::too_many_arguments)]
     fn process_left_mark_join_block(
+        &self,
+        build_indexes: &[RowPtr],
+        mark_scan_map: &mut [Vec<u8>],
+    ) -> Result<()> {
+        if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
+            return Err(ErrorCode::AbortedQuery(
+                "Aborted query, because the server is shutting down or the query was killed.",
+            ));
+        }
+
+        for probed_row in build_indexes.iter() {
+            unsafe {
+                *mark_scan_map
+                    .get_unchecked_mut(probed_row.chunk_index as usize)
+                    .get_unchecked_mut(probed_row.row_index as usize) = MARKER_KIND_TRUE;
+            }
+        }
+
+        Ok(())
+    }
+
+    #[inline]
+    #[allow(clippy::too_many_arguments)]
+    fn process_left_mark_join_with_conjunct_block(
         &self,
         matched_idx: usize,
         input: &DataBlock,
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_semi_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_semi_join.rs
index 56dd6fcb8efd..3a280f5b7554 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_semi_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/left_semi_join.rs
@@ -128,7 +128,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -151,18 +151,15 @@ impl HashJoinProbeState {
                         other_predicate,
                         &mut row_state,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        *idx,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         } else {
@@ -171,7 +168,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill build_indexes.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -194,18 +191,15 @@ impl HashJoinProbeState {
                         other_predicate,
                         &mut row_state,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         }
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_join.rs
index 2f5413bf968c..809476118049 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_join.rs
@@ -76,7 +76,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -100,18 +100,15 @@ impl HashJoinProbeState {
                         outer_scan_map,
                         &mut right_single_scan_map,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        *idx,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         } else {
@@ -120,7 +117,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill build_indexes.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -144,18 +141,15 @@ impl HashJoinProbeState {
                         outer_scan_map,
                         &mut right_single_scan_map,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         }
@@ -177,7 +171,7 @@ impl HashJoinProbeState {
         Ok(result_blocks)
     }
 
-    fn update_right_single_scan_map(
+    pub(crate) fn update_right_single_scan_map(
         &self,
         build_indexes: &[RowPtr],
         right_single_scan_map: &[*mut AtomicBool],
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_mark_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_mark_join.rs
index 45d8a23be659..fed7c8d44cb5 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_mark_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_mark_join.rs
@@ -140,7 +140,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -163,18 +163,15 @@ impl HashJoinProbeState {
                         markers,
                         other_predicate,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        *idx,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         } else {
@@ -183,7 +180,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill build_indexes.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -206,18 +203,15 @@ impl HashJoinProbeState {
                         markers,
                         other_predicate,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         }
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_semi_anti_join.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_semi_anti_join.rs
index 4bc5d6a08bc3..4c190d4dc454 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_semi_anti_join.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_join/right_semi_anti_join.rs
@@ -60,7 +60,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -69,29 +69,14 @@ impl HashJoinProbeState {
                 matched_idx += match_count;
 
                 while matched_idx == max_block_size {
-                    if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
-                        return Err(ErrorCode::AbortedQuery(
-                            "Aborted query, because the server is shutting down or the query was killed.",
-                        ));
-                    }
-
-                    for row_ptr in build_indexes.iter() {
-                        unsafe {
-                            *outer_scan_map
-                                .get_unchecked_mut(row_ptr.chunk_index as usize)
-                                .get_unchecked_mut(row_ptr.row_index as usize) = true
-                        };
-                    }
-
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    self.process_right_semi_anti_join_block(build_indexes, outer_scan_map)?;
+                    (matched_idx, incomplete_ptr) = hash_table.next_probe(
                         key,
                         incomplete_ptr,
                         build_indexes_ptr,
-                        matched_idx,
+                        0,
                         max_block_size,
                     );
-                    matched_idx += match_count;
                 }
             }
         } else {
@@ -100,7 +85,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -108,39 +93,23 @@ impl HashJoinProbeState {
 
                 matched_idx += match_count;
                 while matched_idx == max_block_size {
-                    if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
-                        return Err(ErrorCode::AbortedQuery(
-                            "Aborted query, because the server is shutting down or the query was killed.",
-                        ));
-                    }
-
-                    for row_ptr in build_indexes.iter() {
-                        unsafe {
-                            *outer_scan_map
-                                .get_unchecked_mut(row_ptr.chunk_index as usize)
-                                .get_unchecked_mut(row_ptr.row_index as usize) = true
-                        };
-                    }
-
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    self.process_right_semi_anti_join_block(build_indexes, outer_scan_map)?;
+                    (matched_idx, incomplete_ptr) = hash_table.next_probe(
                         key,
                         incomplete_ptr,
                         build_indexes_ptr,
-                        matched_idx,
+                        0,
                         max_block_size,
                     );
-                    matched_idx += match_count;
                 }
             }
         }
 
-        for row_ptr in build_indexes[0..matched_idx].iter() {
-            unsafe {
-                *outer_scan_map
-                    .get_unchecked_mut(row_ptr.chunk_index as usize)
-                    .get_unchecked_mut(row_ptr.row_index as usize) = true
-            };
+        if matched_idx > 0 {
+            self.process_right_semi_anti_join_block(
+                &build_indexes[0..matched_idx],
+                outer_scan_map,
+            )?;
         }
 
         Ok(vec![])
@@ -185,7 +154,7 @@ impl HashJoinProbeState {
                 let ptr = unsafe { *pointers.get_unchecked(*idx as usize) };
 
                 // Probe hash table and fill `build_indexes`.
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -198,7 +167,7 @@ impl HashJoinProbeState {
                 }
 
                 while matched_idx == max_block_size {
-                    self.process_right_semi_anti_join_block(
+                    self.process_right_semi_anti_join_with_conjunct_block(
                         matched_idx,
                         input,
                         probe_indexes,
@@ -208,25 +177,22 @@ impl HashJoinProbeState {
                         outer_scan_map,
                         other_predicate,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        *idx,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = *idx };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         } else {
             for idx in 0..input.num_rows() {
                 let key = unsafe { keys.key_unchecked(idx) };
                 let ptr = unsafe { *pointers.get_unchecked(idx) };
-                let (mut match_count, mut incomplete_ptr) =
+                let (match_count, mut incomplete_ptr) =
                     hash_table.next_probe(key, ptr, build_indexes_ptr, matched_idx, max_block_size);
                 if match_count == 0 {
                     continue;
@@ -238,7 +204,7 @@ impl HashJoinProbeState {
                 }
 
                 while matched_idx == max_block_size {
-                    self.process_right_semi_anti_join_block(
+                    self.process_right_semi_anti_join_with_conjunct_block(
                         matched_idx,
                         input,
                         probe_indexes,
@@ -248,24 +214,21 @@ impl HashJoinProbeState {
                         outer_scan_map,
                         other_predicate,
                     )?;
-                    matched_idx = 0;
-                    (match_count, incomplete_ptr) = hash_table.next_probe(
+                    (matched_idx, incomplete_ptr) = self.fill_probe_and_build_indexes::<_, false>(
+                        hash_table,
                         key,
                         incomplete_ptr,
+                        idx as u32,
+                        probe_indexes,
                         build_indexes_ptr,
-                        matched_idx,
                         max_block_size,
-                    );
-                    for _ in 0..match_count {
-                        unsafe { *probe_indexes.get_unchecked_mut(matched_idx) = idx as u32 };
-                        matched_idx += 1;
-                    }
+                    )?;
                 }
             }
         }
 
         if matched_idx > 0 {
-            self.process_right_semi_anti_join_block(
+            self.process_right_semi_anti_join_with_conjunct_block(
                 matched_idx,
                 input,
                 probe_indexes,
@@ -283,6 +246,30 @@ impl HashJoinProbeState {
     #[inline]
     #[allow(clippy::too_many_arguments)]
     fn process_right_semi_anti_join_block(
+        &self,
+        build_indexes: &[RowPtr],
+        outer_scan_map: &mut [Vec<bool>],
+    ) -> Result<()> {
+        if self.hash_join_state.interrupt.load(Ordering::Relaxed) {
+            return Err(ErrorCode::AbortedQuery(
+                "Aborted query, because the server is shutting down or the query was killed.",
+            ));
+        }
+
+        for row_ptr in build_indexes.iter() {
+            unsafe {
+                *outer_scan_map
+                    .get_unchecked_mut(row_ptr.chunk_index as usize)
+                    .get_unchecked_mut(row_ptr.row_index as usize) = true
+            };
+        }
+
+        Ok(())
+    }
+
+    #[inline]
+    #[allow(clippy::too_many_arguments)]
+    fn process_right_semi_anti_join_with_conjunct_block(
         &self,
         matched_idx: usize,
         input: &DataBlock,
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_state.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_state.rs
index f4215a541e8d..8be9a294a40d 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/probe_state.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/probe_state.rs
@@ -41,18 +41,23 @@ pub struct ProbeState {
     pub(crate) probe_unmatched_indexes: Option<Vec<u32>>,
     // The `markers` is used for right mark join.
     pub(crate) markers: Option<Vec<u8>>,
+    // If hash join other condition is not empty.
+    pub(crate) with_conjunction: bool,
 
     // Early filtering.
     // 1.The `selection` is used to store the indexes of input which matched by hash.
     pub(crate) selection: Vec<u32>,
     // 2.The indexes of [0, selection_count) in `selection` are valid.
     pub(crate) selection_count: usize,
-    // 3.Statistics for **adaptive** early filtering, the `num_keys` indicates the number of valid keys in probe side,
-    // the `num_keys_hash_matched` indicates the number of keys which matched by hash.
+    // 3.Statistics for **adaptive** early filtering, the `num_keys` indicates the number of valid keys
+    // in probe side, the `num_keys_hash_matched` indicates the number of keys which matched by hash.
     pub(crate) num_keys: u64,
     pub(crate) num_keys_hash_matched: u64,
     // 4.Whether to probe with selection.
     pub(crate) probe_with_selection: bool,
+    // 5.If join type is LEFT / LEFT SINGLE / LEFT ANTI / FULL, we use it to store unmatched indexes
+    // count during early filtering.
+    pub(crate) probe_unmatched_indexes_count: usize,
 }
 
 impl ProbeState {
@@ -64,27 +69,28 @@ impl ProbeState {
     pub fn create(
         max_block_size: usize,
         join_type: &JoinType,
-        with_conjunct: bool,
+        with_conjunction: bool,
         has_string_column: bool,
         func_ctx: FunctionContext,
     ) -> Self {
-        let (row_state, row_state_indexes, probe_unmatched_indexes) = match &join_type {
+        let (row_state, row_state_indexes) = match &join_type {
             JoinType::Left | JoinType::LeftSingle | JoinType::Full => {
-                if with_conjunct {
-                    (
-                        Some(vec![0; max_block_size]),
-                        Some(vec![0; max_block_size]),
-                        None,
-                    )
+                if with_conjunction {
+                    (Some(vec![0; max_block_size]), Some(vec![0; max_block_size]))
                 } else {
-                    (
-                        Some(vec![0; max_block_size]),
-                        None,
-                        Some(vec![0; max_block_size]),
-                    )
+                    (Some(vec![0; max_block_size]), None)
                 }
             }
-            _ => (None, None, None),
+            _ => (None, None),
+        };
+        let probe_unmatched_indexes = if matches!(
+            &join_type,
+            JoinType::Left | JoinType::LeftSingle | JoinType::Full | JoinType::LeftAnti
+        ) && !with_conjunction
+        {
+            Some(vec![0; max_block_size])
+        } else {
+            None
         };
         let markers = if matches!(&join_type, JoinType::RightMark) {
             Some(vec![MARKER_KIND_FALSE; max_block_size])
@@ -106,6 +112,8 @@ impl ProbeState {
             row_state_indexes,
             probe_unmatched_indexes,
             markers,
+            probe_unmatched_indexes_count: 0,
+            with_conjunction,
         }
     }
 
@@ -138,17 +146,11 @@ impl MutableIndexes {
 }
 
 pub struct ProbeBlockGenerationState {
-    /// in fact, it means whether we need to output some probe blocks's columns,
-    /// we use probe_projections to check whether we can get a non-empty result
-    /// block.
+    // The is_probe_projected means whether we need to output probe columns.
     pub(crate) is_probe_projected: bool,
-    /// for Right/Full/RightSingle we use true_validity to reduce memory, because
-    /// we need to wrap probe block's all column type as nullable(if they are not).
-    /// But when we need to wrap this way, the validity is all true, so we use this
-    /// one to share the memory.
+    // When we need a bitmap that is all true, we can directly slice it to reduce memory usage.
     pub(crate) true_validity: Bitmap,
-    /// we use `string_items_buf` for Binary/String/Bitmap/Variant Column
-    /// to store the (pointer,length). So we can reuse the memory for all take.
+    // The string_items_buf is used as a buffer to reduce memory allocation when taking [u8] Columns.
     pub(crate) string_items_buf: Option<Vec<(u64, usize)>>,
 }
 
diff --git a/src/query/service/src/pipelines/processors/transforms/hash_join/result_blocks.rs b/src/query/service/src/pipelines/processors/transforms/hash_join/result_blocks.rs
index cf4af55b6b0c..3676f45c35c9 100644
--- a/src/query/service/src/pipelines/processors/transforms/hash_join/result_blocks.rs
+++ b/src/query/service/src/pipelines/processors/transforms/hash_join/result_blocks.rs
@@ -56,10 +56,26 @@ impl HashJoinProbeState {
             .other_predicate
             .is_none();
         match self.hash_join_state.hash_join_desc.join_type {
-            JoinType::Inner => self.inner_join(input, keys, hash_table, probe_state),
-            JoinType::Left | JoinType::LeftSingle | JoinType::Full => match has_other_predicate {
-                true => self.left_join(input, keys, hash_table, probe_state),
-                false => self.left_join_with_conjunct(input, keys, hash_table, probe_state),
+            JoinType::Inner => match self.hash_join_state.hash_join_desc.single_to_inner {
+                Some(JoinType::LeftSingle) => {
+                    self.inner_join::<_, true, false>(input, keys, hash_table, probe_state)
+                }
+                Some(JoinType::RightSingle) => {
+                    self.inner_join::<_, false, true>(input, keys, hash_table, probe_state)
+                }
+                _ => self.inner_join::<_, false, false>(input, keys, hash_table, probe_state),
+            },
+            JoinType::Left | JoinType::Full => match has_other_predicate {
+                true => self.left_join::<_, false>(input, keys, hash_table, probe_state),
+                false => {
+                    self.left_join_with_conjunct::<_, false>(input, keys, hash_table, probe_state)
+                }
+            },
+            JoinType::LeftSingle => match has_other_predicate {
+                true => self.left_join::<_, true>(input, keys, hash_table, probe_state),
+                false => {
+                    self.left_join_with_conjunct::<_, true>(input, keys, hash_table, probe_state)
+                }
             },
             JoinType::LeftSemi => match has_other_predicate {
                 true => self.left_semi_join(input, keys, hash_table, probe_state),
diff --git a/src/query/service/src/pipelines/processors/transforms/transform_sort_spill.rs b/src/query/service/src/pipelines/processors/transforms/transform_sort_spill.rs
index c29a193f766f..c96f454accf4 100644
--- a/src/query/service/src/pipelines/processors/transforms/transform_sort_spill.rs
+++ b/src/query/service/src/pipelines/processors/transforms/transform_sort_spill.rs
@@ -38,6 +38,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_transforms::processors::sort::CommonRows;
 use databend_common_pipeline_transforms::processors::sort::DateRows;
 use databend_common_pipeline_transforms::processors::sort::HeapMerger;
@@ -283,6 +285,13 @@ where R: Rows + Sync + Send + 'static
             metrics_inc_sort_spill_write_count();
             metrics_inc_sort_spill_write_bytes(bytes);
             metrics_inc_sort_spill_write_milliseconds(ins.elapsed().as_millis() as u64);
+
+            Profile::record_usize_profile(ProfileStatisticsName::SpillWriteCount, 1);
+            Profile::record_usize_profile(ProfileStatisticsName::SpillWriteBytes, bytes as usize);
+            Profile::record_usize_profile(
+                ProfileStatisticsName::SpillWriteTime,
+                ins.elapsed().as_millis() as usize,
+            );
         }
 
         self.unmerged_blocks.push_back(vec![location].into());
@@ -342,6 +351,16 @@ where R: Rows + Sync + Send + 'static
                 metrics_inc_sort_spill_read_count();
                 metrics_inc_sort_spill_read_bytes(bytes);
                 metrics_inc_sort_spill_read_milliseconds(ins.elapsed().as_millis() as u64);
+
+                Profile::record_usize_profile(ProfileStatisticsName::SpillReadCount, 1);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::SpillReadBytes,
+                    bytes as usize,
+                );
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::SpillReadTime,
+                    ins.elapsed().as_millis() as usize,
+                );
             }
 
             self.output_data = Some(block);
@@ -376,6 +395,16 @@ where R: Rows + Sync + Send + 'static
                 metrics_inc_sort_spill_write_count();
                 metrics_inc_sort_spill_write_bytes(bytes);
                 metrics_inc_sort_spill_write_milliseconds(ins.elapsed().as_millis() as u64);
+
+                Profile::record_usize_profile(ProfileStatisticsName::SpillWriteCount, 1);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::SpillWriteBytes,
+                    bytes as usize,
+                );
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::SpillWriteTime,
+                    ins.elapsed().as_millis() as usize,
+                );
             }
 
             spilled.push_back(location);
@@ -408,6 +437,16 @@ impl SortedStream for BlockStream {
                         metrics_inc_sort_spill_read_count();
                         metrics_inc_sort_spill_read_bytes(bytes);
                         metrics_inc_sort_spill_read_milliseconds(ins.elapsed().as_millis() as u64);
+
+                        Profile::record_usize_profile(ProfileStatisticsName::SpillReadCount, 1);
+                        Profile::record_usize_profile(
+                            ProfileStatisticsName::SpillReadBytes,
+                            bytes as usize,
+                        );
+                        Profile::record_usize_profile(
+                            ProfileStatisticsName::SpillReadTime,
+                            ins.elapsed().as_millis() as usize,
+                        );
                     }
 
                     Some(block)
diff --git a/src/query/service/src/schedulers/fragments/fragmenter.rs b/src/query/service/src/schedulers/fragments/fragmenter.rs
index 20bebb05bc3b..b7e974b59d01 100644
--- a/src/query/service/src/schedulers/fragments/fragmenter.rs
+++ b/src/query/service/src/schedulers/fragments/fragmenter.rs
@@ -251,7 +251,9 @@ impl PhysicalPlanReplacer for Fragmenter {
             need_hold_hash_table: plan.need_hold_hash_table,
             stat_info: plan.stat_info.clone(),
             probe_keys_rt: plan.probe_keys_rt.clone(),
+            enable_bloom_runtime_filter: plan.enable_bloom_runtime_filter,
             broadcast: plan.broadcast,
+            single_to_inner: plan.single_to_inner.clone(),
         }))
     }
 
diff --git a/src/query/service/src/schedulers/fragments/query_fragment_actions.rs b/src/query/service/src/schedulers/fragments/query_fragment_actions.rs
index 0884578eef5b..69daa9f491ba 100644
--- a/src/query/service/src/schedulers/fragments/query_fragment_actions.rs
+++ b/src/query/service/src/schedulers/fragments/query_fragment_actions.rs
@@ -105,15 +105,13 @@ impl QueryFragmentActions {
 
 pub struct QueryFragmentsActions {
     ctx: Arc<QueryContext>,
-    enable_profiling: bool,
     pub fragments_actions: Vec<QueryFragmentActions>,
 }
 
 impl QueryFragmentsActions {
-    pub fn create(ctx: Arc<QueryContext>, enable_profiling: bool) -> QueryFragmentsActions {
+    pub fn create(ctx: Arc<QueryContext>) -> QueryFragmentsActions {
         QueryFragmentsActions {
             ctx,
-            enable_profiling,
             fragments_actions: Vec::new(),
         }
     }
@@ -183,7 +181,6 @@ impl QueryFragmentsActions {
             nodes_info.clone(),
             settings.clone(),
             cluster.local_id(),
-            self.enable_profiling,
         );
 
         for (executor, fragments) in fragments_packets.into_iter() {
@@ -199,7 +196,6 @@ impl QueryFragmentsActions {
                 executors_info,
                 settings.clone(),
                 cluster.local_id(),
-                self.enable_profiling,
             ));
         }
 
diff --git a/src/query/service/src/schedulers/fragments/query_fragment_actions_display.rs b/src/query/service/src/schedulers/fragments/query_fragment_actions_display.rs
index 647bd1b2ba0a..f715a8a70675 100644
--- a/src/query/service/src/schedulers/fragments/query_fragment_actions_display.rs
+++ b/src/query/service/src/schedulers/fragments/query_fragment_actions_display.rs
@@ -15,7 +15,6 @@
 use std::fmt::Display;
 use std::fmt::Formatter;
 
-use databend_common_profile::SharedProcessorProfiles;
 use databend_common_sql::MetadataRef;
 
 use crate::api::DataExchange;
@@ -80,7 +79,7 @@ impl<'a> Display for QueryFragmentActionsWrap<'a> {
             let fragment_action = &self.inner.fragment_actions[0];
             let plan_display_string = fragment_action
                 .physical_plan
-                .format(self.metadata.clone(), SharedProcessorProfiles::default())
+                .format(self.metadata.clone(), Default::default())
                 .and_then(|node| node.format_pretty_with_prefix("    "))
                 .unwrap();
             write!(f, "{}", plan_display_string)?;
diff --git a/src/query/service/src/schedulers/scheduler.rs b/src/query/service/src/schedulers/scheduler.rs
index dd1830218e7c..2bd3a8f427c7 100644
--- a/src/query/service/src/schedulers/scheduler.rs
+++ b/src/query/service/src/schedulers/scheduler.rs
@@ -15,7 +15,6 @@
 use std::sync::Arc;
 
 use databend_common_exception::Result;
-use databend_common_profile::SharedProcessorProfiles;
 
 use crate::pipelines::PipelineBuildResult;
 use crate::pipelines::PipelineBuilder;
@@ -36,9 +35,7 @@ pub async fn build_query_pipeline(
     plan: &PhysicalPlan,
     ignore_result: bool,
 ) -> Result<PipelineBuildResult> {
-    let enable_profile = ctx.get_settings().get_enable_query_profiling()?;
-    let mut build_res =
-        build_query_pipeline_without_render_result_set(ctx, plan, enable_profile).await?;
+    let mut build_res = build_query_pipeline_without_render_result_set(ctx, plan).await?;
 
     let input_schema = plan.output_schema()?;
     PipelineBuilder::build_result_projection(
@@ -55,12 +52,11 @@ pub async fn build_query_pipeline(
 pub async fn build_query_pipeline_without_render_result_set(
     ctx: &Arc<QueryContext>,
     plan: &PhysicalPlan,
-    enable_profiling: bool,
 ) -> Result<PipelineBuildResult> {
     let build_res = if !plan.is_distributed_plan() {
-        build_local_pipeline(ctx, plan, enable_profiling).await
+        build_local_pipeline(ctx, plan).await
     } else {
-        build_distributed_pipeline(ctx, plan, false).await
+        build_distributed_pipeline(ctx, plan).await
     }?;
     Ok(build_res)
 }
@@ -70,14 +66,11 @@ pub async fn build_query_pipeline_without_render_result_set(
 pub async fn build_local_pipeline(
     ctx: &Arc<QueryContext>,
     plan: &PhysicalPlan,
-    enable_profiling: bool,
 ) -> Result<PipelineBuildResult> {
     let pipeline = PipelineBuilder::create(
         ctx.get_function_context()?,
         ctx.get_settings(),
         ctx.clone(),
-        enable_profiling,
-        SharedProcessorProfiles::default(),
         vec![],
     );
     let mut build_res = pipeline.finalize(plan)?;
@@ -92,18 +85,17 @@ pub async fn build_local_pipeline(
 pub async fn build_distributed_pipeline(
     ctx: &Arc<QueryContext>,
     plan: &PhysicalPlan,
-    enable_profiling: bool,
 ) -> Result<PipelineBuildResult> {
     let fragmenter = Fragmenter::try_create(ctx.clone())?;
 
     let root_fragment = fragmenter.build_fragment(plan)?;
-    let mut fragments_actions = QueryFragmentsActions::create(ctx.clone(), enable_profiling);
+    let mut fragments_actions = QueryFragmentsActions::create(ctx.clone());
     root_fragment.get_actions(ctx.clone(), &mut fragments_actions)?;
 
     let exchange_manager = ctx.get_exchange_manager();
 
     let mut build_res = exchange_manager
-        .commit_actions(ctx.clone(), enable_profiling, fragments_actions)
+        .commit_actions(ctx.clone(), fragments_actions)
         .await?;
 
     let settings = ctx.get_settings();
diff --git a/src/query/service/src/servers/flight_sql/flight_sql_service/query.rs b/src/query/service/src/servers/flight_sql/flight_sql_service/query.rs
index 99a56cfbd17a..7d49ce80e2cd 100644
--- a/src/query/service/src/servers/flight_sql/flight_sql_service/query.rs
+++ b/src/query/service/src/servers/flight_sql/flight_sql_service/query.rs
@@ -26,6 +26,7 @@ use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_expression::DataBlock;
 use databend_common_expression::DataSchema;
+use databend_common_sql::get_query_kind;
 use databend_common_sql::plans::Plan;
 use databend_common_sql::PlanExtras;
 use databend_common_sql::Planner;
@@ -95,7 +96,10 @@ impl FlightSqlServiceImpl {
             .await
             .map_err(|e| status!("Could not create_query_context", e))?;
 
-        context.attach_query_str(plan.kind(), plan_extras.statement.to_mask_sql());
+        context.attach_query_str(
+            get_query_kind(&plan_extras.statement),
+            plan_extras.statement.to_mask_sql(),
+        );
         let interpreter = InterpreterFactory::get(context.clone(), plan).await?;
 
         let mut blocks = interpreter.execute(context.clone()).await?;
@@ -120,7 +124,10 @@ impl FlightSqlServiceImpl {
             .await
             .map_err(|e| status!("Could not create_query_context", e))?;
 
-        context.attach_query_str(plan.kind(), plan_extras.statement.to_mask_sql());
+        context.attach_query_str(
+            get_query_kind(&plan_extras.statement),
+            plan_extras.statement.to_mask_sql(),
+        );
         let interpreter = InterpreterFactory::get(context.clone(), plan).await?;
 
         let data_schema = plan.schema();
diff --git a/src/query/service/src/servers/http/clickhouse_handler.rs b/src/query/service/src/servers/http/clickhouse_handler.rs
index 661a07a4354f..6d7b364fdabf 100644
--- a/src/query/service/src/servers/http/clickhouse_handler.rs
+++ b/src/query/service/src/servers/http/clickhouse_handler.rs
@@ -267,8 +267,6 @@ pub async fn clickhouse_handler_get(
             .map_err(|err| err.display_with_sql(&sql))
             .map_err(BadRequest)?;
         let format = get_format_with_default(extras.format, default_format)?;
-
-        context.attach_query_str(plan.kind(), extras.statement.to_mask_sql());
         let interpreter = InterpreterFactory::get(context.clone(), &plan)
             .await
             .map_err(|err| err.display_with_sql(&sql))
@@ -346,7 +344,6 @@ pub async fn clickhouse_handler_post(
             .map_err(|err| err.display_with_sql(&sql))
             .map_err(BadRequest)?;
         let schema = plan.schema();
-        ctx.attach_query_str(plan.kind(), extras.statement.to_mask_sql());
         let mut handle = None;
         if let Plan::Insert(insert) = &mut plan {
             if let InsertInputSource::StreamingWithFormat(format, start, input_context_ref) =
diff --git a/src/query/service/src/servers/http/v1/load.rs b/src/query/service/src/servers/http/v1/load.rs
index 5cd523103686..19ca53bd7143 100644
--- a/src/query/service/src/servers/http/v1/load.rs
+++ b/src/query/service/src/servers/http/v1/load.rs
@@ -128,12 +128,11 @@ pub async fn streaming_load(
     }
 
     let mut planner = Planner::new(context.clone());
-    let (mut plan, extras) = planner
+    let (mut plan, _) = planner
         .plan_sql(insert_sql)
         .await
         .map_err(|err| err.display_with_sql(insert_sql))
         .map_err(InternalServerError)?;
-    context.attach_query_str(plan.kind(), extras.statement.to_mask_sql());
 
     let schema = plan.schema();
     match &mut plan {
diff --git a/src/query/service/src/servers/http/v1/query/execute_state.rs b/src/query/service/src/servers/http/v1/query/execute_state.rs
index 7c6b2867d0e8..0f1c96b57e38 100644
--- a/src/query/service/src/servers/http/v1/query/execute_state.rs
+++ b/src/query/service/src/servers/http/v1/query/execute_state.rs
@@ -272,12 +272,10 @@ impl ExecuteState {
     pub(crate) async fn try_start_query(
         executor: Arc<RwLock<Executor>>,
         plan: Plan,
-        extras: PlanExtras,
         session: Arc<Session>,
         ctx: Arc<QueryContext>,
         block_sender: SizedChannelSender<DataBlock>,
     ) -> Result<()> {
-        ctx.attach_query_str(plan.kind(), extras.statement.to_mask_sql());
         let interpreter = InterpreterFactory::get(ctx.clone(), &plan).await?;
         let running_state = ExecuteRunning {
             session,
@@ -325,7 +323,7 @@ async fn execute(
         // duplicate codes, but there is an async call
         let data = BlockEntry::new(
             DataType::String,
-            databend_common_expression::Value::Scalar(Scalar::String(err.to_string().into_bytes())),
+            databend_common_expression::Value::Scalar(Scalar::String(err.to_string())),
         );
         block_sender.send(DataBlock::new(vec![data], 1), 1).await;
         return Err(err);
@@ -342,9 +340,7 @@ async fn execute(
             // duplicate codes, but there is an async call
             let data = BlockEntry::new(
                 DataType::String,
-                databend_common_expression::Value::Scalar(Scalar::String(
-                    err.to_string().into_bytes(),
-                )),
+                databend_common_expression::Value::Scalar(Scalar::String(err.to_string())),
             );
             block_sender.send(DataBlock::new(vec![data], 1), 1).await;
             Executor::stop(&executor, Err(err), false).await;
@@ -363,7 +359,7 @@ async fn execute(
                         let data = BlockEntry::new(
                             DataType::String,
                             databend_common_expression::Value::Scalar(Scalar::String(
-                                err.to_string().into_bytes(),
+                                err.to_string(),
                             )),
                         );
                         block_sender.send(DataBlock::new(vec![data], 1), 1).await;
diff --git a/src/query/service/src/servers/http/v1/query/http_query.rs b/src/query/service/src/servers/http/v1/query/http_query.rs
index 48e064f631ff..6d33f62fe33c 100644
--- a/src/query/service/src/servers/http/v1/query/http_query.rs
+++ b/src/query/service/src/servers/http/v1/query/http_query.rs
@@ -341,7 +341,7 @@ impl HttpQuery {
         let sql = request.sql.clone();
         let query_id_clone = query_id.clone();
 
-        let (plan, plan_extras) = ExecuteState::plan_sql(&sql, ctx.clone()).await?;
+        let (plan, _) = ExecuteState::plan_sql(&sql, ctx.clone()).await?;
         let schema = plan.schema();
 
         let span = if let Some(parent) = SpanContext::current_local_parent() {
@@ -358,7 +358,6 @@ impl HttpQuery {
                 if let Err(e) = ExecuteState::try_start_query(
                     state,
                     plan,
-                    plan_extras,
                     session,
                     ctx_clone.clone(),
                     block_sender,
diff --git a/src/query/service/src/servers/mysql/mysql_federated.rs b/src/query/service/src/servers/mysql/mysql_federated.rs
index 88723083246c..bc040afbb1f2 100644
--- a/src/query/service/src/servers/mysql/mysql_federated.rs
+++ b/src/query/service/src/servers/mysql/mysql_federated.rs
@@ -43,9 +43,8 @@ impl MySQLFederated {
     // |value|
     fn select_function_block(name: &str, value: &str) -> Option<(TableSchemaRef, DataBlock)> {
         let schema = TableSchemaRefExt::create(vec![TableField::new(name, TableDataType::String)]);
-        let block = DataBlock::new_from_columns(vec![StringType::from_data(vec![
-            value.as_bytes().to_vec(),
-        ])]);
+        let block =
+            DataBlock::new_from_columns(vec![StringType::from_data(vec![value.to_string()])]);
         Some((schema, block))
     }
 
@@ -59,8 +58,8 @@ impl MySQLFederated {
             TableField::new("Value", TableDataType::String),
         ]);
         let block = DataBlock::new_from_columns(vec![
-            StringType::from_data(vec![name.as_bytes().to_vec()]),
-            StringType::from_data(vec![value.as_bytes().to_vec()]),
+            StringType::from_data(vec![name.to_string()]),
+            StringType::from_data(vec![value.to_string()]),
         ]);
         Some((schema, block))
     }
@@ -103,7 +102,7 @@ impl MySQLFederated {
                     // var is 'cc'.
                     let var = vars_as[0];
                     let value = default_map.get(var).unwrap_or(&"0").to_string();
-                    values.push(StringType::from_data(vec![value.as_bytes().to_vec()]));
+                    values.push(StringType::from_data(vec![value]));
                 } else {
                     // @@aa
                     // var is 'aa'
@@ -113,7 +112,7 @@ impl MySQLFederated {
                     ));
 
                     let value = default_map.get(var).unwrap_or(&"0").to_string();
-                    values.push(StringType::from_data(vec![value.as_bytes().to_vec()]));
+                    values.push(StringType::from_data(vec![value]));
                 }
             }
         }
diff --git a/src/query/service/src/servers/mysql/mysql_interactive_worker.rs b/src/query/service/src/servers/mysql/mysql_interactive_worker.rs
index dbc5a2fad137..79b0147a2c12 100644
--- a/src/query/service/src/servers/mysql/mysql_interactive_worker.rs
+++ b/src/query/service/src/servers/mysql/mysql_interactive_worker.rs
@@ -354,9 +354,7 @@ impl InteractiveWorkerBase {
                 let context = self.session.create_query_context().await?;
 
                 let mut planner = Planner::new(context.clone());
-                let (plan, extras) = planner.plan_sql(query).await?;
-
-                context.attach_query_str(plan.kind(), extras.statement.to_mask_sql());
+                let (plan, _) = planner.plan_sql(query).await?;
                 let interpreter = InterpreterFactory::get(context.clone(), &plan).await;
 
                 let has_result_set = plan.has_result_set();
diff --git a/src/query/service/src/sessions/query_ctx.rs b/src/query/service/src/sessions/query_ctx.rs
index 49a1981fec0b..722bba610a6b 100644
--- a/src/query/service/src/sessions/query_ctx.rs
+++ b/src/query/service/src/sessions/query_ctx.rs
@@ -71,6 +71,7 @@ use databend_common_meta_app::schema::TableInfo;
 use databend_common_metrics::storage::*;
 use databend_common_pipeline_core::processors::profile::PlanProfile;
 use databend_common_pipeline_core::processors::profile::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_pipeline_core::InputError;
 use databend_common_settings::Settings;
 use databend_common_sql::IndexType;
@@ -406,15 +407,20 @@ impl TableContext for QueryContext {
     }
 
     fn get_partition(&self) -> Option<PartInfoPtr> {
-        self.partition_queue.write().pop_front()
+        if let Some(part) = self.partition_queue.write().pop_front() {
+            Profile::record_usize_profile(ProfileStatisticsName::ScanPartitions, 1);
+            return Some(part);
+        }
+
+        None
     }
 
     fn get_partitions(&self, num: usize) -> Vec<PartInfoPtr> {
         let mut res = Vec::with_capacity(num);
-        let mut partition_queue = self.partition_queue.write();
+        let mut queue_guard = self.partition_queue.write();
 
         for _index in 0..num {
-            match partition_queue.pop_front() {
+            match queue_guard.pop_front() {
                 None => {
                     break;
                 }
@@ -424,6 +430,8 @@ impl TableContext for QueryContext {
             };
         }
 
+        Profile::record_usize_profile(ProfileStatisticsName::ScanPartitions, res.len());
+
         res
     }
 
diff --git a/src/query/service/src/table_functions/async_crash_me.rs b/src/query/service/src/table_functions/async_crash_me.rs
index b2408e5760be..a3345542d89d 100644
--- a/src/query/service/src/table_functions/async_crash_me.rs
+++ b/src/query/service/src/table_functions/async_crash_me.rs
@@ -61,10 +61,10 @@ impl AsyncCrashMeTable {
         let args = table_args.expect_all_positioned(table_func_name, None)?;
         if args.len() == 1 {
             let arg = args[0].clone();
-            panic_message =
-                Some(String::from_utf8(arg.into_string().map_err(|_| {
-                    ErrorCode::BadArguments("Expected string argument")
-                })?)?);
+            panic_message = Some(
+                arg.into_string()
+                    .map_err(|_| ErrorCode::BadArguments("Expected string argument"))?,
+            );
         }
 
         let table_info = TableInfo {
@@ -119,7 +119,7 @@ impl Table for AsyncCrashMeTable {
 
     fn table_args(&self) -> Option<TableArgs> {
         let args = match &self.panic_message {
-            Some(s) => vec![Scalar::String(s.as_bytes().to_vec())],
+            Some(s) => vec![Scalar::String(s.clone())],
             None => vec![],
         };
         Some(TableArgs::new_positioned(args))
diff --git a/src/query/service/src/table_functions/cloud/task_dependents.rs b/src/query/service/src/table_functions/cloud/task_dependents.rs
index 11344caf1b24..8273f8cbc257 100644
--- a/src/query/service/src/table_functions/cloud/task_dependents.rs
+++ b/src/query/service/src/table_functions/cloud/task_dependents.rs
@@ -194,38 +194,30 @@ impl TaskDependentsSource {
     }
     fn to_block(&self, tasks: &Vec<Task>) -> databend_common_exception::Result<DataBlock> {
         let mut created_on: Vec<i64> = Vec::with_capacity(tasks.len());
-        let mut name: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-        let mut owner: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-        let mut comment: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
-        let mut warehouse: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
-        let mut schedule: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
-        let mut predecessors: Vec<Vec<Vec<u8>>> = Vec::with_capacity(tasks.len());
+        let mut name: Vec<String> = Vec::with_capacity(tasks.len());
+        let mut owner: Vec<String> = Vec::with_capacity(tasks.len());
+        let mut comment: Vec<Option<String>> = Vec::with_capacity(tasks.len());
+        let mut warehouse: Vec<Option<String>> = Vec::with_capacity(tasks.len());
+        let mut schedule: Vec<Option<String>> = Vec::with_capacity(tasks.len());
+        let mut predecessors: Vec<Vec<String>> = Vec::with_capacity(tasks.len());
 
-        let mut state: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-        let mut definition: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-        let mut condition_text: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
+        let mut state: Vec<String> = Vec::with_capacity(tasks.len());
+        let mut definition: Vec<String> = Vec::with_capacity(tasks.len());
+        let mut condition_text: Vec<String> = Vec::with_capacity(tasks.len());
 
         for task in tasks {
             let task = task.clone();
             let tsk: task_utils::Task = task.try_into()?;
             created_on.push(tsk.created_at.timestamp_micros());
-            name.push(tsk.task_name.into_bytes());
-            owner.push(tsk.owner.into_bytes());
-            comment.push(tsk.comment.map(|s| s.into_bytes()));
-            warehouse.push(
-                tsk.warehouse_options
-                    .and_then(|s| s.warehouse.map(|v| v.into_bytes())),
-            );
-            schedule.push(tsk.schedule_options.map(|s| s.into_bytes()));
-            predecessors.push(
-                tsk.after
-                    .into_iter()
-                    .map(|s| s.into_bytes())
-                    .collect::<Vec<_>>(),
-            );
-            state.push(tsk.status.to_string().into_bytes());
-            definition.push(tsk.query_text.into_bytes());
-            condition_text.push(tsk.condition_text.into_bytes());
+            name.push(tsk.task_name.clone());
+            owner.push(tsk.owner.clone());
+            comment.push(tsk.comment.clone());
+            warehouse.push(tsk.warehouse_options.and_then(|s| s.warehouse.clone()));
+            schedule.push(tsk.schedule_options.clone());
+            predecessors.push(tsk.after.clone());
+            state.push(tsk.status.to_string());
+            definition.push(tsk.query_text.clone());
+            condition_text.push(tsk.condition_text.clone());
         }
 
         Ok(DataBlock::new_from_columns(vec![
@@ -302,17 +294,12 @@ impl TaskDependentsParsed {
     pub fn parse(table_args: &TableArgs) -> databend_common_exception::Result<Self> {
         let args = table_args.expect_all_named("task_dependents")?;
 
-        let mut task_name = String::from("");
-        let mut recursive = false;
+        let mut task_name = None;
+        let mut recursive = None;
         for (k, v) in &args {
             match k.to_lowercase().as_str() {
-                "task_name" => {
-                    task_name =
-                        String::from_utf8_lossy(v.as_string().unwrap().as_slice()).to_string();
-                }
-                "recursive" => {
-                    recursive = *v.as_boolean().unwrap();
-                }
+                "task_name" => task_name = v.as_string().cloned(),
+                "recursive" => recursive = v.as_boolean().cloned(),
                 _ => {
                     return Err(ErrorCode::BadArguments(format!(
                         "unknown param {} for {}",
@@ -321,7 +308,8 @@ impl TaskDependentsParsed {
                 }
             }
         }
-        if task_name.is_empty() {
+
+        if task_name.is_none() {
             return Err(ErrorCode::BadArguments(format!(
                 "task_name must be specified for {}",
                 "task_dependents"
@@ -329,8 +317,8 @@ impl TaskDependentsParsed {
         }
 
         Ok(Self {
-            task_name,
-            recursive,
+            task_name: task_name.unwrap(),
+            recursive: recursive.unwrap_or_default(),
         })
     }
 }
diff --git a/src/query/service/src/table_functions/cloud/task_dependents_enable.rs b/src/query/service/src/table_functions/cloud/task_dependents_enable.rs
index 81b267944591..875287d8036a 100644
--- a/src/query/service/src/table_functions/cloud/task_dependents_enable.rs
+++ b/src/query/service/src/table_functions/cloud/task_dependents_enable.rs
@@ -58,8 +58,7 @@ impl TaskDependentsEnableTable {
         table_args: TableArgs,
     ) -> Result<Arc<dyn TableFunction>> {
         let args = table_args.expect_all_positioned(table_func_name, Some(1))?;
-        let task_name =
-            String::from_utf8_lossy(args[0].as_string().unwrap().as_slice()).to_string();
+        let task_name = args[0].as_string().unwrap();
 
         let table_info = TableInfo {
             ident: TableIdent::new(table_id, 0),
@@ -81,7 +80,7 @@ impl TaskDependentsEnableTable {
 
         Ok(Arc::new(TaskDependentsEnableTable {
             table_info,
-            task_name,
+            task_name: task_name.to_string(),
         }))
     }
 }
@@ -113,7 +112,7 @@ impl Table for TaskDependentsEnableTable {
 
     fn table_args(&self) -> Option<TableArgs> {
         Some(TableArgs::new_positioned(vec![Scalar::String(
-            self.task_name.as_bytes().to_vec(),
+            self.task_name.clone(),
         )]))
     }
 
diff --git a/src/query/service/src/table_functions/infer_schema/infer_schema_table.rs b/src/query/service/src/table_functions/infer_schema/infer_schema_table.rs
index 3e3dbaf66735..d668290062c9 100644
--- a/src/query/service/src/table_functions/infer_schema/infer_schema_table.rs
+++ b/src/query/service/src/table_functions/infer_schema/infer_schema_table.rs
@@ -262,15 +262,15 @@ impl AsyncSource for InferSchemaSource {
             }
         };
 
-        let mut names: Vec<Vec<u8>> = vec![];
-        let mut types: Vec<Vec<u8>> = vec![];
+        let mut names: Vec<String> = vec![];
+        let mut types: Vec<String> = vec![];
         let mut nulls: Vec<bool> = vec![];
 
         for field in schema.fields().iter() {
-            names.push(field.name().to_string().as_bytes().to_vec());
+            names.push(field.name().to_string());
 
             let non_null_type = field.data_type().remove_recursive_nullable();
-            types.push(non_null_type.sql_name().as_bytes().to_vec());
+            types.push(non_null_type.sql_name());
             nulls.push(field.is_nullable());
         }
 
diff --git a/src/query/service/src/table_functions/inspect_parquet/inspect_parquet_table.rs b/src/query/service/src/table_functions/inspect_parquet/inspect_parquet_table.rs
index e105f801505a..cbf20d85cd02 100644
--- a/src/query/service/src/table_functions/inspect_parquet/inspect_parquet_table.rs
+++ b/src/query/service/src/table_functions/inspect_parquet/inspect_parquet_table.rs
@@ -69,12 +69,10 @@ impl InspectParquetTable {
         table_args: TableArgs,
     ) -> Result<Arc<dyn TableFunction>> {
         let args = table_args.expect_all_positioned(table_func_name, Some(1))?;
-        let file_path = String::from_utf8(
-            args[0]
-                .clone()
-                .into_string()
-                .map_err(|_| ErrorCode::BadArguments("Expected string argument."))?,
-        )?;
+        let file_path = args[0]
+            .clone()
+            .into_string()
+            .map_err(|_| ErrorCode::BadArguments("Expected string argument."))?;
         if !file_path.starts_with('@') {
             return Err(ErrorCode::BadArguments(format!(
                 "stage path must start with @, but got {}",
@@ -265,7 +263,7 @@ impl AsyncSource for InspectParquetSource {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(StringType::upcast_scalar(created.into())),
+                    Value::Scalar(StringType::upcast_scalar(created)),
                 ),
                 BlockEntry::new(
                     DataType::Number(NumberDataType::UInt64),
diff --git a/src/query/service/src/table_functions/list_stage/list_stage_table.rs b/src/query/service/src/table_functions/list_stage/list_stage_table.rs
index 69055fe7b043..045c71dbf94a 100644
--- a/src/query/service/src/table_functions/list_stage/list_stage_table.rs
+++ b/src/query/service/src/table_functions/list_stage/list_stage_table.rs
@@ -211,28 +211,24 @@ impl AsyncSource for ListStagesSource {
 
         let files = files_info.list(&op, false, None).await?;
 
-        let names: Vec<Vec<u8>> = files
-            .iter()
-            .map(|file| file.path.to_string().into_bytes())
-            .collect();
+        let names: Vec<String> = files.iter().map(|file| file.path.to_string()).collect();
 
         let sizes: Vec<u64> = files.iter().map(|file| file.size).collect();
-        let etags: Vec<Option<Vec<u8>>> = files
+        let etags: Vec<Option<String>> = files
             .iter()
-            .map(|file| file.etag.as_ref().map(|f| f.to_string().into_bytes()))
+            .map(|file| file.etag.as_ref().map(|f| f.to_string()))
             .collect();
-        let last_modifieds: Vec<Vec<u8>> = files
+        let last_modifieds: Vec<String> = files
             .iter()
             .map(|file| {
                 file.last_modified
                     .format("%Y-%m-%d %H:%M:%S.%3f %z")
                     .to_string()
-                    .into_bytes()
             })
             .collect();
-        let creators: Vec<Option<Vec<u8>>> = files
+        let creators: Vec<Option<String>> = files
             .iter()
-            .map(|file| file.creator.as_ref().map(|c| c.to_string().into_bytes()))
+            .map(|file| file.creator.as_ref().map(|c| c.to_string()))
             .collect();
 
         let block = DataBlock::new_from_columns(vec![
diff --git a/src/query/service/src/table_functions/openai/ai_to_sql.rs b/src/query/service/src/table_functions/openai/ai_to_sql.rs
index 078f4e46415d..3bb83a5e33e8 100644
--- a/src/query/service/src/table_functions/openai/ai_to_sql.rs
+++ b/src/query/service/src/table_functions/openai/ai_to_sql.rs
@@ -64,12 +64,10 @@ impl GPT2SQLTable {
     ) -> Result<Arc<dyn TableFunction>> {
         // Check args.
         let args = table_args.expect_all_positioned(table_func_name, Some(1))?;
-        let prompt = String::from_utf8(
-            args[0]
-                .clone()
-                .into_string()
-                .map_err(|_| ErrorCode::BadArguments("Expected string argument."))?,
-        )?;
+        let prompt = args[0]
+            .clone()
+            .into_string()
+            .map_err(|_| ErrorCode::BadArguments("Expected string argument."))?;
 
         let schema = TableSchema::new(vec![
             TableField::new("database", TableDataType::String),
@@ -248,8 +246,8 @@ impl AsyncSource for GPT2SQLSource {
         let sql = format!("SELECT {}", sql);
         info!("openai response sql: {}", sql);
         let database = self.ctx.get_current_database();
-        let database: Vec<Vec<u8>> = vec![database.into_bytes()];
-        let sql: Vec<Vec<u8>> = vec![sql.into_bytes()];
+        let database: Vec<String> = vec![database];
+        let sql: Vec<String> = vec![sql];
 
         // Mark done.
         self.finished = true;
diff --git a/src/query/service/src/table_functions/others/execute_background_job.rs b/src/query/service/src/table_functions/others/execute_background_job.rs
index 22e8705a6942..bc903c16fdb7 100644
--- a/src/query/service/src/table_functions/others/execute_background_job.rs
+++ b/src/query/service/src/table_functions/others/execute_background_job.rs
@@ -53,7 +53,7 @@ impl ExecuteBackgroundJobTable {
         table_args: TableArgs,
     ) -> Result<Arc<dyn TableFunction>> {
         let args = table_args.expect_all_positioned(table_func_name, Some(1))?;
-        let job_name = String::from_utf8_lossy(args[0].as_string().unwrap().as_slice()).to_string();
+        let job_name = args[0].as_string().unwrap().clone();
 
         let table_info = TableInfo {
             ident: TableIdent::new(table_id, 0),
@@ -107,7 +107,7 @@ impl Table for ExecuteBackgroundJobTable {
 
     fn table_args(&self) -> Option<TableArgs> {
         Some(TableArgs::new_positioned(vec![Scalar::String(
-            self.job_name.as_bytes().to_vec(),
+            self.job_name.clone(),
         )]))
     }
 
diff --git a/src/query/service/src/table_functions/others/license_info.rs b/src/query/service/src/table_functions/others/license_info.rs
index df62b0207e40..228476c67ecc 100644
--- a/src/query/service/src/table_functions/others/license_info.rs
+++ b/src/query/service/src/table_functions/others/license_info.rs
@@ -161,33 +161,19 @@ impl LicenseInfoSource {
                 BlockEntry::new(
                     DataType::String,
                     Value::Scalar(Scalar::String(
-                        info.issuer
-                            .clone()
-                            .unwrap_or("".to_string())
-                            .into_bytes()
-                            .to_vec(),
+                        info.issuer.clone().unwrap_or("".to_string()),
                     )),
                 ),
                 BlockEntry::new(
                     DataType::String,
                     Value::Scalar(Scalar::String(
-                        info.custom
-                            .r#type
-                            .clone()
-                            .unwrap_or("".to_string())
-                            .into_bytes()
-                            .to_vec(),
+                        info.custom.r#type.clone().unwrap_or("".to_string()),
                     )),
                 ),
                 BlockEntry::new(
                     DataType::String,
                     Value::Scalar(Scalar::String(
-                        info.custom
-                            .org
-                            .clone()
-                            .unwrap_or("".to_string())
-                            .into_bytes()
-                            .to_vec(),
+                        info.custom.org.clone().unwrap_or("".to_string()),
                     )),
                 ),
                 BlockEntry::new(
@@ -204,9 +190,7 @@ impl LicenseInfoSource {
                 ),
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(
-                        human_readable_available_time.into_bytes().to_vec(),
-                    )),
+                    Value::Scalar(Scalar::String(human_readable_available_time)),
                 ),
             ],
             1,
diff --git a/src/query/service/src/table_functions/others/suggested_background_tasks.rs b/src/query/service/src/table_functions/others/suggested_background_tasks.rs
index 6032138024b0..2f96e5f96902 100644
--- a/src/query/service/src/table_functions/others/suggested_background_tasks.rs
+++ b/src/query/service/src/table_functions/others/suggested_background_tasks.rs
@@ -164,8 +164,7 @@ impl SuggestedBackgroundTasksSource {
         sql: String,
     ) -> Result<Option<RecordBatch>> {
         let mut planner = Planner::new(ctx.clone());
-        let (plan, plan_extras) = planner.plan_sql(sql.as_str()).await?;
-        ctx.attach_query_str(plan.kind(), plan_extras.statement.to_mask_sql());
+        let (plan, _) = planner.plan_sql(sql.as_str()).await?;
         let data_schema = plan.schema();
         let interpreter = InterpreterFactory::get(ctx.clone(), &plan).await?;
         let stream = interpreter.execute(ctx.clone()).await?;
@@ -220,11 +219,11 @@ impl SuggestedBackgroundTasksSource {
                     table_stats,
                     ..
                 } => {
-                    suggestion_type.push("compaction".to_string().into_bytes().to_vec());
+                    suggestion_type.push("compaction".to_string());
                     should_do_segment_compact.push(Some(need_compact_segment));
                     should_do_compact.push(Some(need_compact_block));
-                    database_name.push(db_name.into_bytes().to_vec());
-                    table_names.push(table_name.into_bytes().to_vec());
+                    database_name.push(db_name);
+                    table_names.push(table_name);
                     table_statistics.push(serde_json::to_vec(&table_stats).unwrap());
                 }
             }
diff --git a/src/query/service/src/table_functions/others/tenant_quota.rs b/src/query/service/src/table_functions/others/tenant_quota.rs
index a1cd50241efd..2cb900527464 100644
--- a/src/query/service/src/table_functions/others/tenant_quota.rs
+++ b/src/query/service/src/table_functions/others/tenant_quota.rs
@@ -137,7 +137,7 @@ impl Table for TenantQuotaTable {
         let args = self
             .args
             .iter()
-            .map(|s| Scalar::String(s.as_bytes().to_vec()))
+            .map(|s| Scalar::String(s.clone()))
             .collect();
         Some(TableArgs::new_positioned(args))
     }
diff --git a/src/query/service/src/table_functions/sync_crash_me.rs b/src/query/service/src/table_functions/sync_crash_me.rs
index 131fa7999839..e968a962d1a4 100644
--- a/src/query/service/src/table_functions/sync_crash_me.rs
+++ b/src/query/service/src/table_functions/sync_crash_me.rs
@@ -61,10 +61,10 @@ impl SyncCrashMeTable {
         let args = table_args.expect_all_positioned(table_func_name, None)?;
         if args.len() == 1 {
             let arg = args[0].clone();
-            panic_message =
-                Some(String::from_utf8(arg.into_string().map_err(|_| {
-                    ErrorCode::BadArguments("Expected string argument.")
-                })?)?);
+            panic_message = Some(
+                arg.into_string()
+                    .map_err(|_| ErrorCode::BadArguments("Expected string argument."))?,
+            );
         }
 
         let table_info = TableInfo {
@@ -119,7 +119,7 @@ impl Table for SyncCrashMeTable {
 
     fn table_args(&self) -> Option<TableArgs> {
         let args = match &self.panic_message {
-            Some(s) => vec![Scalar::String(s.as_bytes().to_vec())],
+            Some(s) => vec![Scalar::String(s.clone())],
             None => vec![],
         };
         Some(TableArgs::new_positioned(args))
diff --git a/src/query/service/src/test_kits/config.rs b/src/query/service/src/test_kits/config.rs
index e90f1cf8a1a4..fa64f2fcf2d9 100644
--- a/src/query/service/src/test_kits/config.rs
+++ b/src/query/service/src/test_kits/config.rs
@@ -43,6 +43,7 @@ impl ConfigBuilder {
         let tmp_dir = TempDir::new().expect("create tmp dir failed");
         let root = tmp_dir.path().to_str().unwrap().to_string();
         conf.storage.params = StorageParams::Fs(StorageFsConfig { root });
+        conf.storage.allow_insecure = true;
 
         ConfigBuilder { conf }
     }
diff --git a/src/query/service/src/test_kits/fixture.rs b/src/query/service/src/test_kits/fixture.rs
index 7143e4505dd9..0326c300c091 100644
--- a/src/query/service/src/test_kits/fixture.rs
+++ b/src/query/service/src/test_kits/fixture.rs
@@ -638,7 +638,7 @@ impl TestFixture {
                     let mut d_values = Vec::with_capacity(rows_per_block);
                     for i in 0..rows_per_block {
                         id_values.push(i as i32 + start * 3);
-                        c_values.push(format!("s-{}-{}", start, i).as_bytes().to_vec());
+                        c_values.push(format!("s-{}-{}", start, i));
                         d_values.push(i as i64 + (start * 10) as i64);
                     }
                     let column0 = Int32Type::from_data(id_values);
diff --git a/src/query/service/tests/it/api/http/status.rs b/src/query/service/tests/it/api/http/status.rs
index b5b5db2d3ff2..ac9b8ed229bc 100644
--- a/src/query/service/tests/it/api/http/status.rs
+++ b/src/query/service/tests/it/api/http/status.rs
@@ -25,7 +25,6 @@ use databend_query::interpreters::InterpreterFactory;
 use databend_query::sessions::QueryContext;
 use databend_query::sessions::SessionManager;
 use databend_query::sessions::SessionType;
-use databend_query::sessions::TableContext;
 use databend_query::sql::Planner;
 use databend_query::test_kits::*;
 use poem::get;
@@ -65,8 +64,7 @@ async fn run_query(query_ctx: &Arc<QueryContext>) -> Result<Arc<dyn Interpreter>
         .set_authed_user(user, None)
         .await?;
     let mut planner = Planner::new(query_ctx.clone());
-    let (plan, extras) = planner.plan_sql(sql).await?;
-    query_ctx.attach_query_str(plan.kind(), extras.statement.to_mask_sql());
+    let (plan, _) = planner.plan_sql(sql).await?;
     InterpreterFactory::get(query_ctx.clone(), &plan).await
 }
 
diff --git a/src/query/service/tests/it/pipelines/executor/executor_graph.rs b/src/query/service/tests/it/pipelines/executor/executor_graph.rs
index a247b7b4bad9..e7e436cf6e2d 100644
--- a/src/query/service/tests/it/pipelines/executor/executor_graph.rs
+++ b/src/query/service/tests/it/pipelines/executor/executor_graph.rs
@@ -492,7 +492,6 @@ async fn create_executor_with_simple_pipeline(
     pipeline.add_pipe(sink_pipe);
     pipeline.set_max_threads(size);
     let settings = ExecutorSettings {
-        enable_profiling: false,
         query_id: Arc::new("".to_string()),
         max_execute_time_in_seconds: Default::default(),
     };
diff --git a/src/query/service/tests/it/pipelines/executor/pipeline_executor.rs b/src/query/service/tests/it/pipelines/executor/pipeline_executor.rs
index d68fa1a16750..907c37840c03 100644
--- a/src/query/service/tests/it/pipelines/executor/pipeline_executor.rs
+++ b/src/query/service/tests/it/pipelines/executor/pipeline_executor.rs
@@ -41,7 +41,6 @@ async fn test_always_call_on_finished() -> Result<()> {
     let fixture = TestFixture::setup().await?;
 
     let settings = ExecutorSettings {
-        enable_profiling: false,
         query_id: Arc::new("".to_string()),
         max_execute_time_in_seconds: Default::default(),
     };
diff --git a/src/query/service/tests/it/sql/exec/get_table_bind_test.rs b/src/query/service/tests/it/sql/exec/get_table_bind_test.rs
index d5abe55eb88e..2b63f04dd248 100644
--- a/src/query/service/tests/it/sql/exec/get_table_bind_test.rs
+++ b/src/query/service/tests/it/sql/exec/get_table_bind_test.rs
@@ -492,9 +492,7 @@ impl TableContext for CtxDelegation {
         todo!()
     }
 
-    fn attach_query_str(&self, _kind: QueryKind, _query: String) {
-        todo!()
-    }
+    fn attach_query_str(&self, _kind: QueryKind, _query: String) {}
 
     fn get_query_str(&self) -> String {
         todo!()
@@ -812,6 +810,7 @@ async fn test_get_same_table_once() -> Result<()> {
 
     let mut planner = Planner::new(ctx.clone());
     let (_, _) = planner.plan_sql(query.as_str()).await?;
+
     assert_eq!(
         ctx.table_without_cache
             .load(std::sync::atomic::Ordering::SeqCst),
diff --git a/src/query/service/tests/it/sql/planner/builders/binder.rs b/src/query/service/tests/it/sql/planner/builders/binder.rs
new file mode 100644
index 000000000000..a3829436c27a
--- /dev/null
+++ b/src/query/service/tests/it/sql/planner/builders/binder.rs
@@ -0,0 +1,28 @@
+use databend_common_catalog::query_kind::QueryKind;
+use databend_common_exception::Result;
+use databend_common_sql::Planner;
+use databend_common_storages_fuse::TableContext;
+use databend_query::sessions::SessionType;
+use databend_query::test_kits::TestFixture;
+
+#[tokio::test(flavor = "multi_thread")]
+async fn test_query_kind() -> Result<()> {
+    let fixture = TestFixture::setup().await?;
+    fixture.create_default_database().await?;
+    fixture.create_default_table().await?;
+
+    let http_session = fixture
+        .new_session_with_type(SessionType::HTTPQuery)
+        .await?;
+    let ctx = http_session.create_query_context().await?;
+    let mut planner = Planner::new(ctx.clone());
+    let sql = format!(
+        "COPY INTO {}.{} from  @~/ pattern='.*' FILE_FORMAT = (TYPE = 'csv') PURGE=true FORCE=true max_files=10000;",
+        fixture.default_db_name(),
+        fixture.default_table_name()
+    );
+    let (_, _) = planner.plan_sql(&sql).await?;
+    let kind = ctx.get_query_kind();
+    assert_eq!(kind, QueryKind::CopyIntoTable);
+    Ok(())
+}
diff --git a/src/query/service/tests/it/sql/planner/builders/mod.rs b/src/query/service/tests/it/sql/planner/builders/mod.rs
index 2a94d6a6251a..f7a62d3c019d 100644
--- a/src/query/service/tests/it/sql/planner/builders/mod.rs
+++ b/src/query/service/tests/it/sql/planner/builders/mod.rs
@@ -12,4 +12,5 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+mod binder;
 mod select_builder;
diff --git a/src/query/service/tests/it/sql/planner/format/mod.rs b/src/query/service/tests/it/sql/planner/format/mod.rs
index 317fff56e603..bd3fade8aca2 100644
--- a/src/query/service/tests/it/sql/planner/format/mod.rs
+++ b/src/query/service/tests/it/sql/planner/format/mod.rs
@@ -153,7 +153,8 @@ fn test_format() {
                 marker_index: None,
                 from_correlated_subquery: false,
                 need_hold_hash_table: false,
-                broadcast: false,
+                is_lateral: false,
+                original_join_type: None,
             }
             .into(),
         ),
diff --git a/src/query/service/tests/it/storages/fuse/bloom_index_meta_size.rs b/src/query/service/tests/it/storages/fuse/bloom_index_meta_size.rs
index 4a6467113885..de5e306c01a9 100644
--- a/src/query/service/tests/it/storages/fuse/bloom_index_meta_size.rs
+++ b/src/query/service/tests/it/storages/fuse/bloom_index_meta_size.rs
@@ -294,8 +294,8 @@ fn build_test_segment_info(
     });
 
     let col_stat = ColumnStatistics::new(
-        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?.into_bytes()),
-        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?.into_bytes()),
+        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?),
+        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?),
         0,
         0,
         None,
diff --git a/src/query/service/tests/it/storages/fuse/operations/commit.rs b/src/query/service/tests/it/storages/fuse/operations/commit.rs
index b5db2eaa3563..f7a13c7f4526 100644
--- a/src/query/service/tests/it/storages/fuse/operations/commit.rs
+++ b/src/query/service/tests/it/storages/fuse/operations/commit.rs
@@ -468,9 +468,7 @@ impl TableContext for CtxDelegation {
         todo!()
     }
 
-    fn attach_query_str(&self, _kind: QueryKind, _query: String) {
-        todo!()
-    }
+    fn attach_query_str(&self, _kind: QueryKind, _query: String) {}
 
     fn get_query_str(&self) -> String {
         todo!()
diff --git a/src/query/service/tests/it/storages/fuse/operations/mutation/block_compact_mutator.rs b/src/query/service/tests/it/storages/fuse/operations/mutation/block_compact_mutator.rs
index a5d108589b67..591d2c879de8 100644
--- a/src/query/service/tests/it/storages/fuse/operations/mutation/block_compact_mutator.rs
+++ b/src/query/service/tests/it/storages/fuse/operations/mutation/block_compact_mutator.rs
@@ -122,7 +122,7 @@ async fn do_compact(ctx: Arc<QueryContext>, table: Arc<dyn Table>) -> Result<boo
         )?;
 
         let build_res =
-            build_query_pipeline_without_render_result_set(&ctx, &physical_plan, false).await?;
+            build_query_pipeline_without_render_result_set(&ctx, &physical_plan).await?;
         pipeline = build_res.main_pipeline;
     };
 
diff --git a/src/query/service/tests/it/storages/fuse/statistics.rs b/src/query/service/tests/it/storages/fuse/statistics.rs
index d22ea245490b..11f9d35a4c59 100644
--- a/src/query/service/tests/it/storages/fuse/statistics.rs
+++ b/src/query/service/tests/it/storages/fuse/statistics.rs
@@ -75,8 +75,8 @@ fn test_ft_stats_block_stats() -> databend_common_exception::Result<()> {
     assert_eq!(col_stats.max(), Scalar::Number(NumberScalar::Int32(3)));
     assert_eq!(col_stats.distinct_of_values, Some(3));
     let col_stats = r.get(&1).unwrap();
-    assert_eq!(col_stats.min(), Scalar::String(b"aa".to_vec()));
-    assert_eq!(col_stats.max(), Scalar::String(b"bb".to_vec()));
+    assert_eq!(col_stats.min(), Scalar::String("aa".to_string()));
+    assert_eq!(col_stats.max(), Scalar::String("bb".to_string()));
     assert_eq!(col_stats.distinct_of_values, Some(2));
     Ok(())
 }
@@ -102,8 +102,8 @@ fn test_ft_stats_block_stats_with_column_distinct_count() -> databend_common_exc
     assert_eq!(col_stats.max(), Scalar::Number(NumberScalar::Int32(3)));
     assert_eq!(col_stats.distinct_of_values, Some(3));
     let col_stats = r.get(&1).unwrap();
-    assert_eq!(col_stats.min(), Scalar::String(b"aa".to_vec()));
-    assert_eq!(col_stats.max(), Scalar::String(b"bb".to_vec()));
+    assert_eq!(col_stats.min(), Scalar::String("aa".to_string()));
+    assert_eq!(col_stats.max(), Scalar::String("bb".to_string()));
     assert_eq!(col_stats.distinct_of_values, Some(2));
     Ok(())
 }
@@ -447,8 +447,8 @@ fn test_ft_stats_block_stats_string_columns_trimming() -> databend_common_except
         let min_expr = rand_strings.iter().min().unwrap();
         let max_expr = rand_strings.iter().max().unwrap();
 
-        let data_value_min = Scalar::String(min_expr.clone().into_bytes());
-        let data_value_max = Scalar::String(max_expr.clone().into_bytes());
+        let data_value_min = Scalar::String(min_expr.clone());
+        let data_value_max = Scalar::String(max_expr.clone());
 
         let trimmed_min = data_value_min.clone().trim_min(STATS_STRING_PREFIX_LEN);
         let trimmed_max = data_value_max.clone().trim_max(STATS_STRING_PREFIX_LEN);
@@ -534,7 +534,7 @@ fn test_ft_stats_block_stats_string_columns_trimming_using_eval()
         // - the length of string value is larger or equal than STRING_PREFIX_LEN
         // - AND the string has a prefix of length STRING_PREFIX_LEN, for all the char C in prefix,
         //   C > REPLACEMENT_CHAR; which means we can not replace any of them.
-        let string_max_expr = String::from_utf8(max_expr.as_string().unwrap().to_vec()).unwrap();
+        let string_max_expr = max_expr.as_string().unwrap().to_string();
         let meaningless_to_collect_max = is_degenerated_case(string_max_expr.as_str());
 
         if meaningless_to_collect_max {
@@ -584,12 +584,8 @@ fn is_degenerated_case(value: &str) -> bool {
     larger_than_prefix_len && prefixed_with_irreplaceable_chars
 }
 
-fn char_len(value: &[u8]) -> usize {
-    String::from_utf8(value.to_vec())
-        .unwrap()
-        .as_str()
-        .chars()
-        .count()
+fn char_len(value: &str) -> usize {
+    value.chars().count()
 }
 
 #[test]
diff --git a/src/query/service/tests/it/storages/fuse/table_functions/clustering_information_table.rs b/src/query/service/tests/it/storages/fuse/table_functions/clustering_information_table.rs
index bd14c1c56e14..f5691e5ca303 100644
--- a/src/query/service/tests/it/storages/fuse/table_functions/clustering_information_table.rs
+++ b/src/query/service/tests/it/storages/fuse/table_functions/clustering_information_table.rs
@@ -42,8 +42,8 @@ async fn test_clustering_information_table_read() -> Result<()> {
     fixture.create_default_table().await?;
 
     // func args
-    let arg_db = Scalar::String(db.as_bytes().to_vec());
-    let arg_tbl = Scalar::String(tbl.as_bytes().to_vec());
+    let arg_db = Scalar::String(db.clone());
+    let arg_tbl = Scalar::String(tbl.clone());
 
     {
         let expected = vec![
diff --git a/src/query/service/tests/it/storages/testdata/columns_table.txt b/src/query/service/tests/it/storages/testdata/columns_table.txt
index 846aa4c00165..1b1056cd780b 100644
--- a/src/query/service/tests/it/storages/testdata/columns_table.txt
+++ b/src/query/service/tests/it/storages/testdata/columns_table.txt
@@ -66,7 +66,6 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'constraint_name'                 | 'information_schema' | 'key_column_usage'    | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'constraint_schema'               | 'information_schema' | 'key_column_usage'    | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'copy_options'                    | 'system'             | 'stages'              | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
-| 'cpu_time'                        | 'system'             | 'processor_profile'   | 'UInt64'              | 'BIGINT UNSIGNED'   | ''       | ''       | 'NO'     | ''       |
 | 'cpu_usage'                       | 'system'             | 'query_log'           | 'UInt32'              | 'INT UNSIGNED'      | ''       | ''       | 'NO'     | ''       |
 | 'create_time'                     | 'information_schema' | 'tables'              | 'Timestamp'           | 'TIMESTAMP'         | ''       | ''       | 'NO'     | ''       |
 | 'created_on'                      | 'system'             | 'background_jobs'     | 'Timestamp'           | 'TIMESTAMP'         | ''       | ''       | 'NO'     | ''       |
@@ -142,9 +141,6 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'exception_code'                  | 'system'             | 'task_history'        | 'Int64'               | 'BIGINT'            | ''       | ''       | 'NO'     | ''       |
 | 'exception_text'                  | 'system'             | 'query_log'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'exception_text'                  | 'system'             | 'task_history'        | 'Nullable(String)'    | 'VARCHAR'           | ''       | ''       | 'YES'    | ''       |
-| 'exchange_bytes'                  | 'system'             | 'processor_profile'   | 'UInt64'              | 'BIGINT UNSIGNED'   | ''       | ''       | 'NO'     | ''       |
-| 'exchange_rows'                   | 'system'             | 'processor_profile'   | 'UInt64'              | 'BIGINT UNSIGNED'   | ''       | ''       | 'NO'     | ''       |
-| 'execution_info'                  | 'system'             | 'query_profile'       | 'Variant'             | 'VARIANT'           | ''       | ''       | 'NO'     | ''       |
 | 'extra'                           | 'information_schema' | 'columns'             | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'extra'                           | 'system'             | 'query_log'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'extra_info'                      | 'system'             | 'locks'               | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
@@ -257,11 +253,6 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'numeric_precision'               | 'information_schema' | 'columns'             | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'numeric_precision_radix'         | 'information_schema' | 'columns'             | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'numeric_scale'                   | 'information_schema' | 'columns'             | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
-| 'operator_attribute'              | 'system'             | 'query_summary'       | 'Variant'             | 'VARIANT'           | ''       | ''       | 'NO'     | ''       |
-| 'operator_children'               | 'system'             | 'query_summary'       | 'Array(UInt32)'       | 'ARRAY(UINT32)'     | ''       | ''       | 'NO'     | ''       |
-| 'operator_id'                     | 'system'             | 'query_profile'       | 'UInt32'              | 'INT UNSIGNED'      | ''       | ''       | 'NO'     | ''       |
-| 'operator_id'                     | 'system'             | 'query_summary'       | 'UInt32'              | 'INT UNSIGNED'      | ''       | ''       | 'NO'     | ''       |
-| 'operator_type'                   | 'system'             | 'query_summary'       | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'options'                         | 'system'             | 'password_policies'   | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'ordinal_position'                | 'information_schema' | 'columns'             | 'UInt8'               | 'TINYINT UNSIGNED'  | ''       | ''       | 'NO'     | ''       |
 | 'ordinal_position'                | 'information_schema' | 'key_column_usage'    | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
@@ -290,8 +281,6 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'query_id'                        | 'system'             | 'processor_profile'   | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'query_id'                        | 'system'             | 'query_cache'         | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'query_id'                        | 'system'             | 'query_log'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
-| 'query_id'                        | 'system'             | 'query_profile'       | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
-| 'query_id'                        | 'system'             | 'query_summary'       | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'query_id'                        | 'system'             | 'task_history'        | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'query_kind'                      | 'system'             | 'query_log'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'query_start_time'                | 'system'             | 'query_log'           | 'Timestamp'           | 'TIMESTAMP'         | ''       | ''       | 'NO'     | ''       |
@@ -325,6 +314,8 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'schema_owner'                    | 'information_schema' | 'schemata'            | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'seq_in_index'                    | 'information_schema' | 'statistics'          | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'server_version'                  | 'system'             | 'query_log'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
+| 'session_parameters'              | 'system'             | 'task_history'        | 'Nullable(Variant)'   | 'VARIANT'           | ''       | ''       | 'YES'    | ''       |
+| 'session_parameters'              | 'system'             | 'tasks'               | 'Nullable(Variant)'   | 'VARIANT'           | ''       | ''       | 'YES'    | ''       |
 | 'session_settings'                | 'system'             | 'query_log'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'size'                            | 'system'             | 'caches'              | 'UInt64'              | 'BIGINT UNSIGNED'   | ''       | ''       | 'NO'     | ''       |
 | 'snapshot_location'               | 'system'             | 'streams'             | 'Nullable(String)'    | 'VARCHAR'           | ''       | ''       | 'YES'    | ''       |
@@ -342,6 +333,7 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'state'                           | 'system'             | 'task_history'        | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'state'                           | 'system'             | 'tasks'               | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'statistics'                      | 'system'             | 'malloc_stats'        | 'Variant'             | 'VARIANT'           | ''       | ''       | 'NO'     | ''       |
+| 'statistics'                      | 'system'             | 'processor_profile'   | 'Variant'             | 'VARIANT'           | ''       | ''       | 'NO'     | ''       |
 | 'status'                          | 'system'             | 'backtrace'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'status'                          | 'system'             | 'locks'               | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'status'                          | 'system'             | 'processes'           | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
@@ -411,7 +403,6 @@ DB.Table: 'system'.'columns', Table: columns-table_id:1, ver:0, Engine: SystemCo
 | 'version'                         | 'system'             | 'credits'             | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
 | 'view_definition'                 | 'information_schema' | 'views'               | 'NULL'                | 'NULL'              | ''       | ''       | 'NO'     | ''       |
 | 'virtual_columns'                 | 'system'             | 'virtual_columns'     | 'String'              | 'VARCHAR'           | ''       | ''       | 'NO'     | ''       |
-| 'wait_time'                       | 'system'             | 'processor_profile'   | 'UInt64'              | 'BIGINT UNSIGNED'   | ''       | ''       | 'NO'     | ''       |
 | 'warehouse'                       | 'system'             | 'task_history'        | 'Nullable(String)'    | 'VARCHAR'           | ''       | ''       | 'YES'    | ''       |
 | 'warehouse'                       | 'system'             | 'tasks'               | 'Nullable(String)'    | 'VARCHAR'           | ''       | ''       | 'YES'    | ''       |
 | 'written_bytes'                   | 'system'             | 'query_log'           | 'UInt64'              | 'BIGINT UNSIGNED'   | ''       | ''       | 'NO'     | ''       |
diff --git a/src/query/service/tests/it/storages/testdata/configs_table_basic.txt b/src/query/service/tests/it/storages/testdata/configs_table_basic.txt
index ea8c7d6f4cbf..00f6663075f0 100644
--- a/src/query/service/tests/it/storages/testdata/configs_table_basic.txt
+++ b/src/query/service/tests/it/storages/testdata/configs_table_basic.txt
@@ -123,7 +123,7 @@ DB.Table: 'system'.'configs', Table: configs-table_id:1, ver:0, Engine: SystemCo
 | 'query'   | 'udf_server_allow_list'                    | ''                                                             | ''       |
 | 'query'   | 'users'                                    | '{"name":"root","auth_type":"no_password","auth_string":null}' | ''       |
 | 'query'   | 'wait_timeout_mills'                       | '5000'                                                         | ''       |
-| 'storage' | 'allow_insecure'                           | 'false'                                                        | ''       |
+| 'storage' | 'allow_insecure'                           | 'true'                                                         | ''       |
 | 'storage' | 'azblob.account_key'                       | ''                                                             | ''       |
 | 'storage' | 'azblob.account_name'                      | ''                                                             | ''       |
 | 'storage' | 'azblob.container'                         | ''                                                             | ''       |
diff --git a/src/query/service/tests/it/table_functions/ai_to_sql.rs b/src/query/service/tests/it/table_functions/ai_to_sql.rs
index fbefaba8fa69..e02090ab99bb 100644
--- a/src/query/service/tests/it/table_functions/ai_to_sql.rs
+++ b/src/query/service/tests/it/table_functions/ai_to_sql.rs
@@ -21,16 +21,15 @@ use databend_query::table_functions::GPT2SQLTable;
 fn test_ai_to_sql_args() -> Result<()> {
     // 1 arg.
     {
-        let tbl_args =
-            TableArgs::new_positioned(vec![Scalar::String("prompt".to_string().into_bytes())]);
+        let tbl_args = TableArgs::new_positioned(vec![Scalar::String("prompt".to_string())]);
         let _ = GPT2SQLTable::create("system", "ai_to_sql", 1, tbl_args)?;
     }
 
     // 2 args.
     {
         let tbl_args = TableArgs::new_positioned(vec![
-            Scalar::String("prompt".to_string().into_bytes()),
-            Scalar::String("api-key".to_string().into_bytes()),
+            Scalar::String("prompt".to_string()),
+            Scalar::String("api-key".to_string()),
         ]);
         let result = GPT2SQLTable::create("system", "ai_to_sql", 1, tbl_args);
         assert!(result.is_err());
diff --git a/src/query/settings/src/settings_default.rs b/src/query/settings/src/settings_default.rs
index fcb35ca3fd34..74b15449d293 100644
--- a/src/query/settings/src/settings_default.rs
+++ b/src/query/settings/src/settings_default.rs
@@ -266,8 +266,8 @@ impl DefaultSettings {
                     mode: SettingMode::Both,
                     range: None,
                 }),
-                ("enable_runtime_filter", DefaultSettingValue {
-                    value: UserSettingValue::UInt64(0),
+                ("enable_bloom_runtime_filter", DefaultSettingValue {
+                    value: UserSettingValue::UInt64(1),
                     desc: "Enables runtime filter optimization for JOIN.",
                     mode: SettingMode::Both,
                     range: Some(SettingRange::Numeric(0..=1)),
diff --git a/src/query/settings/src/settings_getter_setter.rs b/src/query/settings/src/settings_getter_setter.rs
index 0ec4040cfb0d..3690d6e31708 100644
--- a/src/query/settings/src/settings_getter_setter.rs
+++ b/src/query/settings/src/settings_getter_setter.rs
@@ -278,8 +278,8 @@ impl Settings {
         Ok(self.try_get_u64("join_spilling_threshold")? as usize)
     }
 
-    pub fn get_runtime_filter(&self) -> Result<bool> {
-        Ok(self.try_get_u64("enable_runtime_filter")? != 0)
+    pub fn get_bloom_runtime_filter(&self) -> Result<bool> {
+        Ok(self.try_get_u64("enable_bloom_runtime_filter")? != 0)
     }
 
     pub fn get_prefer_broadcast_join(&self) -> Result<bool> {
diff --git a/src/query/sql/Cargo.toml b/src/query/sql/Cargo.toml
index 398039106d16..15f24f19f1c5 100644
--- a/src/query/sql/Cargo.toml
+++ b/src/query/sql/Cargo.toml
@@ -34,7 +34,6 @@ databend-common-pipeline-core = { path = "../pipeline/core" }
 databend-common-pipeline-sources = { path = "../pipeline/sources" }
 databend-common-pipeline-transforms = { path = "../pipeline/transforms" }
 
-databend-common-profile = { path = "../profile" }
 databend-common-settings = { path = "../settings" }
 databend-common-storage = { path = "../../common/storage" }
 databend-common-storages-delta = { path = "../storages/delta" }
diff --git a/src/query/sql/src/executor/format.rs b/src/query/sql/src/executor/format.rs
index 56bb49360ba9..8222f9a5c809 100644
--- a/src/query/sql/src/executor/format.rs
+++ b/src/query/sql/src/executor/format.rs
@@ -12,13 +12,17 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::HashMap;
+
 use databend_common_ast::ast::FormatTreeNode;
 use databend_common_catalog::plan::PartStatistics;
 use databend_common_exception::Result;
 use databend_common_expression::DataSchemaRef;
 use databend_common_functions::BUILTIN_FUNCTIONS;
-use databend_common_profile::SharedProcessorProfiles;
+use databend_common_pipeline_core::processors::profile::PlanProfile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use itertools::Itertools;
+use time::Duration;
 
 use crate::executor::explain::PlanStatsInfo;
 use crate::executor::physical_plans::AggregateExpand;
@@ -60,10 +64,10 @@ impl PhysicalPlan {
     pub fn format(
         &self,
         metadata: MetadataRef,
-        prof_span_set: SharedProcessorProfiles,
+        profs: HashMap<u32, PlanProfile>,
     ) -> Result<FormatTreeNode<String>> {
         let metadata = metadata.read().clone();
-        to_format_tree(self, &metadata, &prof_span_set)
+        to_format_tree(self, &metadata, &profs)
     }
 
     pub fn format_join(&self, metadata: &MetadataRef) -> Result<FormatTreeNode<String>> {
@@ -81,8 +85,8 @@ impl PhysicalPlan {
 
                 Ok(FormatTreeNode::with_children(
                     format!(
-                        "Scan: {} (read rows: {})",
-                        table_name, plan.source.statistics.read_rows
+                        "Scan: {} (#{}) (read rows: {})",
+                        table_name, plan.table_index, plan.source.statistics.read_rows
                     ),
                     vec![],
                 ))
@@ -176,7 +180,7 @@ impl PhysicalPlan {
 fn to_format_tree(
     plan: &PhysicalPlan,
     metadata: &Metadata,
-    profs: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     match plan {
         PhysicalPlan::TableScan(plan) => table_scan_to_format_tree(plan, metadata, profs),
@@ -238,25 +242,29 @@ fn to_format_tree(
 /// Helper function to add profile info to the format tree.
 fn append_profile_info(
     children: &mut Vec<FormatTreeNode<String>>,
-    profs: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
     plan_id: u32,
 ) {
-    if let Some(prof) = profs.lock().unwrap().get(&plan_id) {
+    if let Some(prof) = profs.get(&plan_id) {
         children.push(FormatTreeNode::new(format!(
             "output rows: {}",
-            prof.output_rows,
+            prof.statistics[ProfileStatisticsName::OutputRows as usize],
         )));
         children.push(FormatTreeNode::new(format!(
             "output bytes: {}",
-            prof.output_bytes,
+            prof.statistics[ProfileStatisticsName::OutputBytes as usize],
         )));
         children.push(FormatTreeNode::new(format!(
             "total cpu time: {:.3}ms",
-            prof.cpu_time.as_secs_f64() * 1000.0
+            Duration::nanoseconds(prof.statistics[ProfileStatisticsName::CpuTime as usize] as i64)
+                .as_seconds_f64()
+                * 1000.0,
         )));
         children.push(FormatTreeNode::new(format!(
             "total wait time: {:.3}ms",
-            prof.wait_time.as_secs_f64() * 1000.0
+            Duration::nanoseconds(prof.statistics[ProfileStatisticsName::WaitTime as usize] as i64)
+                .as_seconds_f64()
+                * 1000.0,
         )));
     }
 }
@@ -271,7 +279,7 @@ fn copy_into_table(plan: &CopyIntoTable) -> Result<FormatTreeNode<String>> {
 fn table_scan_to_format_tree(
     plan: &TableScan,
     metadata: &Metadata,
-    profs: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     if plan.table_index == DUMMY_TABLE_INDEX {
         return Ok(FormatTreeNode::new("DummyTableScan".to_string()));
@@ -415,7 +423,7 @@ fn constant_table_scan_to_format_tree(
 fn filter_to_format_tree(
     plan: &Filter,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let filter = plan
         .predicates
@@ -435,9 +443,9 @@ fn filter_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "Filter".to_string(),
@@ -448,7 +456,7 @@ fn filter_to_format_tree(
 fn project_to_format_tree(
     plan: &Project,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let mut children = vec![FormatTreeNode::new(format!(
         "output columns: [{}]",
@@ -460,9 +468,9 @@ fn project_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "Project".to_string(),
@@ -473,10 +481,10 @@ fn project_to_format_tree(
 fn eval_scalar_to_format_tree(
     plan: &EvalScalar,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     if plan.exprs.is_empty() {
-        return to_format_tree(&plan.input, metadata, prof_span_set);
+        return to_format_tree(&plan.input, metadata, profs);
     }
     let scalars = plan
         .exprs
@@ -497,9 +505,9 @@ fn eval_scalar_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "EvalScalar".to_string(),
@@ -522,7 +530,7 @@ pub fn pretty_display_agg_desc(desc: &AggregateFunctionDesc, metadata: &Metadata
 fn aggregate_expand_to_format_tree(
     plan: &AggregateExpand,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let sets = plan
         .grouping_sets
@@ -551,9 +559,9 @@ fn aggregate_expand_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "AggregateExpand".to_string(),
@@ -564,7 +572,7 @@ fn aggregate_expand_to_format_tree(
 fn aggregate_partial_to_format_tree(
     plan: &AggregatePartial,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let group_by = plan
         .group_by
@@ -592,9 +600,9 @@ fn aggregate_partial_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "AggregatePartial".to_string(),
@@ -605,7 +613,7 @@ fn aggregate_partial_to_format_tree(
 fn aggregate_final_to_format_tree(
     plan: &AggregateFinal,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let group_by = plan
         .group_by
@@ -643,9 +651,9 @@ fn aggregate_final_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "AggregateFinal".to_string(),
@@ -656,7 +664,7 @@ fn aggregate_final_to_format_tree(
 fn window_to_format_tree(
     plan: &Window,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let partition_by = plan
         .partition_by
@@ -696,12 +704,16 @@ fn window_to_format_tree(
         FormatTreeNode::new(format!("frame: [{frame}]")),
     ];
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    if let Some(limit) = plan.limit {
+        children.push(FormatTreeNode::new(format!("limit: [{limit}]")))
+    }
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    append_profile_info(&mut children, profs, plan.plan_id);
+
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
-        "Window".to_string(), // todo(ariesdevil): show full window expression.
+        "Window".to_string(),
         children,
     ))
 }
@@ -709,7 +721,7 @@ fn window_to_format_tree(
 fn sort_to_format_tree(
     plan: &Sort,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    prof_span_set: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let sort_keys = plan
         .order_by
@@ -753,7 +765,7 @@ fn sort_to_format_tree(
 fn limit_to_format_tree(
     plan: &Limit,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let mut children = vec![
         FormatTreeNode::new(format!(
@@ -773,9 +785,9 @@ fn limit_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children("Limit".to_string(), children))
 }
@@ -783,7 +795,7 @@ fn limit_to_format_tree(
 fn row_fetch_to_format_tree(
     plan: &RowFetch,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let table_schema = plan.source.source_info.schema();
     let projected_schema = plan.cols_to_fetch.project_schema(&table_schema);
@@ -805,9 +817,9 @@ fn row_fetch_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "RowFetch".to_string(),
@@ -818,7 +830,7 @@ fn row_fetch_to_format_tree(
 fn range_join_to_format_tree(
     plan: &RangeJoin,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let range_join_conditions = plan
         .conditions
@@ -843,8 +855,8 @@ fn range_join_to_format_tree(
         .collect::<Vec<_>>()
         .join(", ");
 
-    let mut left_child = to_format_tree(&plan.left, metadata, prof_span_set)?;
-    let mut right_child = to_format_tree(&plan.right, metadata, prof_span_set)?;
+    let mut left_child = to_format_tree(&plan.left, metadata, profs)?;
+    let mut right_child = to_format_tree(&plan.right, metadata, profs)?;
 
     left_child.payload = format!("{}(Left)", left_child.payload);
     right_child.payload = format!("{}(Right)", right_child.payload);
@@ -864,7 +876,7 @@ fn range_join_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
     children.push(left_child);
     children.push(right_child);
@@ -881,7 +893,7 @@ fn range_join_to_format_tree(
 fn hash_join_to_format_tree(
     plan: &HashJoin,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let build_keys = plan
         .build_keys
@@ -902,8 +914,8 @@ fn hash_join_to_format_tree(
         .collect::<Vec<_>>()
         .join(", ");
 
-    let mut build_child = to_format_tree(&plan.build, metadata, prof_span_set)?;
-    let mut probe_child = to_format_tree(&plan.probe, metadata, prof_span_set)?;
+    let mut build_child = to_format_tree(&plan.build, metadata, profs)?;
+    let mut probe_child = to_format_tree(&plan.probe, metadata, profs)?;
 
     build_child.payload = format!("{}(Build)", build_child.payload);
     probe_child.payload = format!("{}(Probe)", probe_child.payload);
@@ -924,7 +936,7 @@ fn hash_join_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
     children.push(build_child);
     children.push(probe_child);
@@ -938,7 +950,7 @@ fn hash_join_to_format_tree(
 fn exchange_to_format_tree(
     plan: &Exchange,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     Ok(FormatTreeNode::with_children("Exchange".to_string(), vec![
         FormatTreeNode::new(format!(
@@ -958,14 +970,14 @@ fn exchange_to_format_tree(
             FragmentKind::Expansive => "Broadcast".to_string(),
             FragmentKind::Merge => "Merge".to_string(),
         })),
-        to_format_tree(&plan.input, metadata, prof_span_set)?,
+        to_format_tree(&plan.input, metadata, profs)?,
     ]))
 }
 
 fn union_all_to_format_tree(
     plan: &UnionAll,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let mut children = vec![FormatTreeNode::new(format!(
         "output columns: [{}]",
@@ -977,11 +989,11 @@ fn union_all_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
     children.extend(vec![
-        to_format_tree(&plan.left, metadata, prof_span_set)?,
-        to_format_tree(&plan.right, metadata, prof_span_set)?,
+        to_format_tree(&plan.left, metadata, profs)?,
+        to_format_tree(&plan.right, metadata, profs)?,
     ]);
 
     Ok(FormatTreeNode::with_children(
@@ -1043,7 +1055,7 @@ fn exchange_source_to_format_tree(
 fn exchange_sink_to_format_tree(
     plan: &ExchangeSink,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let mut children = vec![FormatTreeNode::new(format!(
         "output columns: [{}]",
@@ -1055,7 +1067,7 @@ fn exchange_sink_to_format_tree(
         plan.destination_fragment_id
     )));
 
-    children.push(to_format_tree(&plan.input, metadata, prof_span_set)?);
+    children.push(to_format_tree(&plan.input, metadata, profs)?);
 
     Ok(FormatTreeNode::with_children(
         "ExchangeSink".to_string(),
@@ -1066,9 +1078,9 @@ fn exchange_sink_to_format_tree(
 fn distributed_insert_to_format_tree(
     plan: &DistributedInsertSelect,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
-    let children = vec![to_format_tree(&plan.input, metadata, prof_span_set)?];
+    let children = vec![to_format_tree(&plan.input, metadata, profs)?];
 
     Ok(FormatTreeNode::with_children(
         "DistributedInsertSelect".to_string(),
@@ -1079,9 +1091,9 @@ fn distributed_insert_to_format_tree(
 fn recluster_sink_to_format_tree(
     plan: &ReclusterSink,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
-    let children = vec![to_format_tree(&plan.input, metadata, prof_span_set)?];
+    let children = vec![to_format_tree(&plan.input, metadata, profs)?];
     Ok(FormatTreeNode::with_children(
         "ReclusterSink".to_string(),
         children,
@@ -1091,9 +1103,9 @@ fn recluster_sink_to_format_tree(
 fn commit_sink_to_format_tree(
     plan: &CommitSink,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
-    let children = vec![to_format_tree(&plan.input, metadata, prof_span_set)?];
+    let children = vec![to_format_tree(&plan.input, metadata, profs)?];
     Ok(FormatTreeNode::with_children(
         "CommitSink".to_string(),
         children,
@@ -1103,7 +1115,7 @@ fn commit_sink_to_format_tree(
 fn project_set_to_format_tree(
     plan: &ProjectSet,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let mut children = vec![FormatTreeNode::new(format!(
         "output columns: [{}]",
@@ -1115,7 +1127,7 @@ fn project_set_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
     children.extend(vec![FormatTreeNode::new(format!(
         "set returning functions: {}",
@@ -1126,7 +1138,7 @@ fn project_set_to_format_tree(
             .join(", ")
     ))]);
 
-    children.extend(vec![to_format_tree(&plan.input, metadata, prof_span_set)?]);
+    children.extend(vec![to_format_tree(&plan.input, metadata, profs)?]);
 
     Ok(FormatTreeNode::with_children(
         "ProjectSet".to_string(),
@@ -1137,7 +1149,7 @@ fn project_set_to_format_tree(
 fn udf_to_format_tree(
     plan: &Udf,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let mut children = vec![FormatTreeNode::new(format!(
         "output columns: [{}]",
@@ -1149,7 +1161,7 @@ fn udf_to_format_tree(
         children.extend(items);
     }
 
-    append_profile_info(&mut children, prof_span_set, plan.plan_id);
+    append_profile_info(&mut children, profs, plan.plan_id);
 
     children.extend(vec![FormatTreeNode::new(format!(
         "udf functions: {}",
@@ -1163,7 +1175,7 @@ fn udf_to_format_tree(
             .join(", ")
     ))]);
 
-    children.extend(vec![to_format_tree(&plan.input, metadata, prof_span_set)?]);
+    children.extend(vec![to_format_tree(&plan.input, metadata, profs)?]);
 
     Ok(FormatTreeNode::with_children("Udf".to_string(), children))
 }
@@ -1171,15 +1183,15 @@ fn udf_to_format_tree(
 fn materialized_cte_to_format_tree(
     plan: &MaterializedCte,
     metadata: &Metadata,
-    prof_span_set: &SharedProcessorProfiles,
+    profs: &HashMap<u32, PlanProfile>,
 ) -> Result<FormatTreeNode<String>> {
     let children = vec![
         FormatTreeNode::new(format!(
             "output columns: [{}]",
             format_output_columns(plan.output_schema()?, metadata, true)
         )),
-        to_format_tree(&plan.left, metadata, prof_span_set)?,
-        to_format_tree(&plan.right, metadata, prof_span_set)?,
+        to_format_tree(&plan.left, metadata, profs)?,
+        to_format_tree(&plan.right, metadata, profs)?,
     ];
     Ok(FormatTreeNode::with_children(
         "MaterializedCTE".to_string(),
diff --git a/src/query/sql/src/executor/mod.rs b/src/query/sql/src/executor/mod.rs
index 840393294fd9..f5db1f22246c 100644
--- a/src/query/sql/src/executor/mod.rs
+++ b/src/query/sql/src/executor/mod.rs
@@ -19,7 +19,6 @@ mod physical_plan_builder;
 mod physical_plan_display;
 mod physical_plan_visitor;
 pub mod physical_plans;
-mod profile;
 mod util;
 
 pub mod table_read_plan;
@@ -27,5 +26,4 @@ pub mod table_read_plan;
 pub use physical_plan::PhysicalPlan;
 pub use physical_plan_builder::PhysicalPlanBuilder;
 pub use physical_plan_visitor::PhysicalPlanReplacer;
-pub use profile::*;
 pub use util::*;
diff --git a/src/query/sql/src/executor/physical_plan.rs b/src/query/sql/src/executor/physical_plan.rs
index 429dce5cae5e..bfb5c1c7ad4d 100644
--- a/src/query/sql/src/executor/physical_plan.rs
+++ b/src/query/sql/src/executor/physical_plan.rs
@@ -12,10 +12,14 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::HashMap;
+
 use databend_common_catalog::plan::DataSourcePlan;
 use databend_common_exception::Result;
 use databend_common_expression::DataSchemaRef;
+use databend_common_functions::BUILTIN_FUNCTIONS;
 use enum_as_inner::EnumAsInner;
+use itertools::Itertools;
 
 use crate::executor::physical_plans::AggregateExpand;
 use crate::executor::physical_plans::AggregateFinal;
@@ -377,9 +381,9 @@ impl PhysicalPlan {
             PhysicalPlan::MaterializedCte(_) |
             // Todo: support union and range join return valid table index by join probe keys
             PhysicalPlan::UnionAll(_) |
-            PhysicalPlan::RangeJoin(_)|
+            PhysicalPlan::RangeJoin(_) |
             PhysicalPlan::ConstantTableScan(_)
-            |PhysicalPlan::CteScan(_)
+            | PhysicalPlan::CteScan(_)
             | PhysicalPlan::Udf(_)
             | PhysicalPlan::DeleteSource(_)
             | PhysicalPlan::CopyIntoTable(_)
@@ -397,4 +401,303 @@ impl PhysicalPlan {
             | PhysicalPlan::UpdateSource(_) => usize::MAX,
         }
     }
+
+    pub fn get_desc(&self) -> Result<String> {
+        Ok(match self {
+            PhysicalPlan::TableScan(v) => format!(
+                "{}.{}",
+                v.source.catalog_info.name_ident.catalog_name,
+                v.source.source_info.desc()
+            ),
+            PhysicalPlan::Filter(v) => match v.predicates.is_empty() {
+                true => String::new(),
+                false => v.predicates[0].as_expr(&BUILTIN_FUNCTIONS).sql_display(),
+            },
+            PhysicalPlan::AggregatePartial(v) => v
+                .agg_funcs
+                .iter()
+                .map(|x| {
+                    format!(
+                        "{}({})",
+                        x.sig.name,
+                        x.arg_indices
+                            .iter()
+                            .map(|x| x.to_string())
+                            .collect::<Vec<_>>()
+                            .join(", ")
+                    )
+                })
+                .join(", "),
+            PhysicalPlan::AggregateFinal(v) => v
+                .agg_funcs
+                .iter()
+                .map(|x| {
+                    format!(
+                        "{}({})",
+                        x.sig.name,
+                        x.arg_indices
+                            .iter()
+                            .map(|x| x.to_string())
+                            .collect::<Vec<_>>()
+                            .join(", ")
+                    )
+                })
+                .join(", "),
+            PhysicalPlan::Sort(v) => v
+                .order_by
+                .iter()
+                .map(|x| {
+                    format!(
+                        "{}{}{}",
+                        x.order_by,
+                        if x.asc { "" } else { " DESC" },
+                        if x.nulls_first { " NULLS FIRST" } else { "" },
+                    )
+                })
+                .join(", "),
+            PhysicalPlan::Limit(v) => match v.limit {
+                Some(limit) => format!("LIMIT {} OFFSET {}", limit, v.offset),
+                None => format!("OFFSET {}", v.offset),
+            },
+            PhysicalPlan::Project(v) => v
+                .output_schema()?
+                .fields
+                .iter()
+                .map(|x| x.name())
+                .join(", "),
+            PhysicalPlan::EvalScalar(v) => v
+                .exprs
+                .iter()
+                .map(|(x, _)| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                .join(", "),
+            PhysicalPlan::HashJoin(v) => {
+                format!(
+                    "{} AND {}",
+                    v.build_keys
+                        .iter()
+                        .zip(v.probe_keys.iter())
+                        .map(|(l, r)| format!(
+                            "({} = {})",
+                            l.as_expr(&BUILTIN_FUNCTIONS).sql_display(),
+                            r.as_expr(&BUILTIN_FUNCTIONS).sql_display()
+                        ))
+                        .join(" AND "),
+                    v.non_equi_conditions
+                        .iter()
+                        .map(|x| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                        .join(" AND ")
+                )
+            }
+            PhysicalPlan::ProjectSet(v) => v
+                .srf_exprs
+                .iter()
+                .map(|(x, _)| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                .join(", "),
+            PhysicalPlan::AggregateExpand(v) => v
+                .grouping_sets
+                .sets
+                .iter()
+                .map(|set| {
+                    set.iter()
+                        .map(|x| x.to_string())
+                        .collect::<Vec<_>>()
+                        .join(", ")
+                })
+                .map(|s| format!("({})", s))
+                .collect::<Vec<_>>()
+                .join(", "),
+            PhysicalPlan::Window(v) => {
+                let partition_by = v
+                    .partition_by
+                    .iter()
+                    .map(|x| x.to_string())
+                    .collect::<Vec<_>>()
+                    .join(", ");
+
+                let order_by = v
+                    .order_by
+                    .iter()
+                    .map(|x| {
+                        format!(
+                            "{}{}{}",
+                            x.order_by,
+                            if x.asc { "" } else { " DESC" },
+                            if x.nulls_first { " NULLS FIRST" } else { "" },
+                        )
+                    })
+                    .collect::<Vec<_>>()
+                    .join(", ");
+
+                format!("partition by {}, order by {}", partition_by, order_by)
+            }
+            PhysicalPlan::RowFetch(v) => {
+                let table_schema = v.source.source_info.schema();
+                let projected_schema = v.cols_to_fetch.project_schema(&table_schema);
+                projected_schema.fields.iter().map(|f| f.name()).join(", ")
+            }
+            PhysicalPlan::RangeJoin(v) => {
+                format!(
+                    "{} AND {}",
+                    v.conditions
+                        .iter()
+                        .map(|condition| {
+                            let left = condition
+                                .left_expr
+                                .as_expr(&BUILTIN_FUNCTIONS)
+                                .sql_display();
+                            let right = condition
+                                .right_expr
+                                .as_expr(&BUILTIN_FUNCTIONS)
+                                .sql_display();
+                            format!("{left} {:?} {right}", condition.operator)
+                        })
+                        .join(" AND "),
+                    v.other_conditions
+                        .iter()
+                        .map(|x| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                        .join(" AND ")
+                )
+            }
+            PhysicalPlan::Udf(v) => v
+                .udf_funcs
+                .iter()
+                .map(|x| format!("{}({})", x.func_name, x.arg_exprs.join(", ")))
+                .join(", "),
+            PhysicalPlan::CteScan(v) => {
+                format!("CTE index: {}, sub index: {}", v.cte_idx.0, v.cte_idx.1)
+            }
+            PhysicalPlan::UnionAll(v) => v
+                .pairs
+                .iter()
+                .map(|(l, r)| format!("{} <- {}", l, r))
+                .join(", "),
+            _ => String::new(),
+        })
+    }
+
+    pub fn get_labels(&self) -> Result<HashMap<String, Vec<String>>> {
+        Ok(match self {
+            PhysicalPlan::TableScan(v) => {
+                let output_schema = v.output_schema()?;
+                let source_schema = v.source.source_info.schema();
+                let columns_name = format!(
+                    "Columns ({} / {})",
+                    output_schema.num_fields(),
+                    source_schema.num_fields()
+                );
+                HashMap::from([
+                    (String::from("Full table name"), vec![format!(
+                        "{}.{}",
+                        v.source.catalog_info.name_ident.catalog_name,
+                        v.source.source_info.desc()
+                    )]),
+                    (columns_name, v.name_mapping.keys().cloned().collect()),
+                    (String::from("Total partitions"), vec![
+                        v.source.statistics.partitions_total.to_string(),
+                    ]),
+                ])
+            }
+            PhysicalPlan::Filter(v) => HashMap::from([(
+                String::from("Filter condition"),
+                v.predicates
+                    .iter()
+                    .map(|x| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                    .collect(),
+            )]),
+            PhysicalPlan::Limit(v) => match v.limit {
+                Some(limit) => HashMap::from([
+                    (String::from("Number of rows"), vec![limit.to_string()]),
+                    (String::from("Offset"), vec![v.offset.to_string()]),
+                ]),
+                None => HashMap::from([(String::from("Offset"), vec![v.offset.to_string()])]),
+            },
+            PhysicalPlan::EvalScalar(v) => HashMap::from([(
+                String::from("List of Expressions"),
+                v.exprs
+                    .iter()
+                    .map(|(x, _)| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                    .collect(),
+            )]),
+            PhysicalPlan::Project(v) => HashMap::from([(
+                String::from("List of Expressions"),
+                v.output_schema()?
+                    .fields
+                    .iter()
+                    .map(|x| x.name())
+                    .cloned()
+                    .collect(),
+            )]),
+            PhysicalPlan::AggregatePartial(v) => HashMap::from([
+                (
+                    String::from("Grouping keys"),
+                    v.group_by.iter().map(|x| x.to_string()).collect(),
+                ),
+                (
+                    String::from("Aggregate Functions"),
+                    v.agg_funcs
+                        .iter()
+                        .map(|x| {
+                            format!(
+                                "{}({})",
+                                x.sig.name,
+                                x.arg_indices
+                                    .iter()
+                                    .map(|x| x.to_string())
+                                    .collect::<Vec<_>>()
+                                    .join(", ")
+                            )
+                        })
+                        .collect(),
+                ),
+            ]),
+            PhysicalPlan::AggregateFinal(v) => HashMap::from([
+                (
+                    String::from("Grouping keys"),
+                    v.group_by.iter().map(|x| x.to_string()).collect(),
+                ),
+                (
+                    String::from("Aggregate Functions"),
+                    v.agg_funcs
+                        .iter()
+                        .map(|x| {
+                            format!(
+                                "{}({})",
+                                x.sig.name,
+                                x.arg_indices
+                                    .iter()
+                                    .map(|x| x.to_string())
+                                    .collect::<Vec<_>>()
+                                    .join(", ")
+                            )
+                        })
+                        .collect(),
+                ),
+            ]),
+            PhysicalPlan::HashJoin(v) => HashMap::from([
+                (String::from("Join Type"), vec![v.join_type.to_string()]),
+                (
+                    String::from("Join Build Side Keys"),
+                    v.build_keys
+                        .iter()
+                        .map(|x| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                        .collect(),
+                ),
+                (
+                    String::from("Join Probe Side Keys"),
+                    v.probe_keys
+                        .iter()
+                        .map(|x| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                        .collect(),
+                ),
+                (
+                    String::from("Join Conditions"),
+                    v.non_equi_conditions
+                        .iter()
+                        .map(|x| x.as_expr(&BUILTIN_FUNCTIONS).sql_display())
+                        .collect(),
+                ),
+            ]),
+            _ => HashMap::new(),
+        })
+    }
 }
diff --git a/src/query/sql/src/executor/physical_plan_builder.rs b/src/query/sql/src/executor/physical_plan_builder.rs
index 753f8fecc074..b44c0096b73e 100644
--- a/src/query/sql/src/executor/physical_plan_builder.rs
+++ b/src/query/sql/src/executor/physical_plan_builder.rs
@@ -16,7 +16,6 @@ use std::collections::HashMap;
 use std::sync::Arc;
 
 use databend_common_catalog::table_context::TableContext;
-use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_expression::FunctionContext;
 
@@ -112,10 +111,6 @@ impl PhysicalPlanBuilder {
             }
             RelOperator::AddRowNumber(_) => self.build_add_row_number(s_expr, required).await,
             RelOperator::Udf(udf) => self.build_udf(s_expr, udf, required, stat_info).await,
-            _ => Err(ErrorCode::Internal(format!(
-                "Unsupported physical plan: {:?}",
-                s_expr.plan()
-            ))),
         }
     }
 }
diff --git a/src/query/sql/src/executor/physical_plan_visitor.rs b/src/query/sql/src/executor/physical_plan_visitor.rs
index c79d9305b2c8..089712cbbe0d 100644
--- a/src/query/sql/src/executor/physical_plan_visitor.rs
+++ b/src/query/sql/src/executor/physical_plan_visitor.rs
@@ -209,6 +209,7 @@ pub trait PhysicalPlanReplacer {
             partition_by: plan.partition_by.clone(),
             order_by: plan.order_by.clone(),
             window_frame: plan.window_frame.clone(),
+            limit: plan.limit,
         }))
     }
 
@@ -234,7 +235,9 @@ pub trait PhysicalPlanReplacer {
             need_hold_hash_table: plan.need_hold_hash_table,
             stat_info: plan.stat_info.clone(),
             probe_keys_rt: plan.probe_keys_rt.clone(),
+            enable_bloom_runtime_filter: plan.enable_bloom_runtime_filter,
             broadcast: plan.broadcast,
+            single_to_inner: plan.single_to_inner.clone(),
         }))
     }
 
diff --git a/src/query/sql/src/executor/physical_plans/physical_hash_join.rs b/src/query/sql/src/executor/physical_plans/physical_hash_join.rs
index 31be0ba434c7..33f65539ba5c 100644
--- a/src/query/sql/src/executor/physical_plans/physical_hash_join.rs
+++ b/src/query/sql/src/executor/physical_plans/physical_hash_join.rs
@@ -12,6 +12,9 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::sync::Arc;
+
+use databend_common_catalog::table_context::TableContext;
 use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_expression::type_check::check_cast;
@@ -27,14 +30,17 @@ use databend_common_functions::BUILTIN_FUNCTIONS;
 
 use crate::executor::explain::PlanStatsInfo;
 use crate::executor::physical_plans::Exchange;
+use crate::executor::physical_plans::FragmentKind;
 use crate::executor::PhysicalPlan;
 use crate::executor::PhysicalPlanBuilder;
 use crate::optimizer::ColumnSet;
+use crate::optimizer::RelExpr;
 use crate::optimizer::SExpr;
 use crate::plans::Join;
 use crate::plans::JoinType;
 use crate::ColumnEntry;
 use crate::IndexType;
+use crate::MetadataRef;
 use crate::ScalarExpr;
 use crate::TypeCheck;
 
@@ -69,10 +75,15 @@ pub struct HashJoin {
     // Only used for explain
     pub stat_info: Option<PlanStatsInfo>,
 
-    // probe keys for runtime filter
-    pub probe_keys_rt: Vec<Option<RemoteExpr<String>>>,
+    // probe keys for runtime filter, and record the index of table that used in probe keys.
+    pub probe_keys_rt: Vec<Option<(RemoteExpr<String>, IndexType)>>,
+    // If enable bloom runtime filter
+    pub enable_bloom_runtime_filter: bool,
     // Under cluster, mark if the join is broadcast join.
     pub broadcast: bool,
+    // When left/right single join converted to inner join, record the original join type
+    // and do some special processing during runtime.
+    pub single_to_inner: Option<JoinType>,
 }
 
 impl HashJoin {
@@ -93,6 +104,16 @@ impl PhysicalPlanBuilder {
     ) -> Result<PhysicalPlan> {
         let mut probe_side = Box::new(self.build(s_expr.child(0)?, required.0).await?);
         let mut build_side = Box::new(self.build(s_expr.child(1)?, required.1).await?);
+
+        let mut is_broadcast = false;
+        // Check if join is broadcast join
+        if let PhysicalPlan::Exchange(Exchange {
+            kind: FragmentKind::Expansive,
+            ..
+        }) = build_side.as_ref()
+        {
+            is_broadcast = true;
+        }
         // Unify the data types of the left and right exchange keys.
         if let (
             PhysicalPlan::Exchange(Exchange {
@@ -177,6 +198,7 @@ impl PhysicalPlanBuilder {
         let mut right_join_conditions = Vec::new();
         let mut left_join_conditions_rt = Vec::new();
         let mut probe_to_build_index = Vec::new();
+        let mut table_index = None;
         for (left_condition, right_condition) in join
             .left_conditions
             .iter()
@@ -189,25 +211,38 @@ impl PhysicalPlanBuilder {
                 .type_check(probe_schema.as_ref())?
                 .project_column_ref(|index| probe_schema.index_of(&index.to_string()).unwrap());
 
-            let left_expr_for_runtime_filter =
-                if left_condition.used_columns().iter().all(|idx| {
-                    // Runtime filter only support column in base table. It's possible to use a wrong derived column with
-                    // the same name as a base table column, so we need to check if the column is a base table column.
-                    matches!(
-                        self.metadata.read().column(*idx),
-                        ColumnEntry::BaseTableColumn(_)
-                    )
-                }) && matches!(probe_side, box PhysicalPlan::TableScan(_))
-                {
-                    Some(
+            let left_expr_for_runtime_filter = if left_condition.used_columns().iter().all(|idx| {
+                // Runtime filter only support column in base table. It's possible to use a wrong derived column with
+                // the same name as a base table column, so we need to check if the column is a base table column.
+                matches!(
+                    self.metadata.read().column(*idx),
+                    ColumnEntry::BaseTableColumn(_)
+                )
+            }) {
+                if let Some(column_idx) = left_condition.used_columns().iter().next() {
+                    // Safe to unwrap because we have checked the column is a base table column.
+                    if table_index.is_none() {
+                        table_index = Some(
+                            self.metadata
+                                .read()
+                                .column(*column_idx)
+                                .table_index()
+                                .unwrap(),
+                        );
+                    }
+                    Some((
                         left_condition
                             .as_raw_expr()
                             .type_check(&*self.metadata.read())?
                             .project_column_ref(|col| col.column_name.clone()),
-                    )
+                        table_index.unwrap(),
+                    ))
                 } else {
                     None
-                };
+                }
+            } else {
+                None
+            };
 
             if join.join_type == JoinType::Inner {
                 if let (ScalarExpr::BoundColumnRef(left), ScalarExpr::BoundColumnRef(right)) =
@@ -266,7 +301,10 @@ impl PhysicalPlanBuilder {
             )?;
 
             let left_expr_for_runtime_filter = left_expr_for_runtime_filter
-                .map(|expr| check_cast(expr.span(), false, expr, &common_ty, &BUILTIN_FUNCTIONS))
+                .map(|(expr, idx)| {
+                    check_cast(expr.span(), false, expr, &common_ty, &BUILTIN_FUNCTIONS)
+                        .map(|casted_expr| (casted_expr, idx))
+                })
                 .transpose()?;
 
             let (left_expr, _) =
@@ -274,13 +312,17 @@ impl PhysicalPlanBuilder {
             let (right_expr, _) =
                 ConstantFolder::fold(&right_expr, &self.func_ctx, &BUILTIN_FUNCTIONS);
 
-            let left_expr_for_runtime_filter = left_expr_for_runtime_filter
-                .map(|expr| ConstantFolder::fold(&expr, &self.func_ctx, &BUILTIN_FUNCTIONS).0);
+            let left_expr_for_runtime_filter = left_expr_for_runtime_filter.map(|(expr, idx)| {
+                (
+                    ConstantFolder::fold(&expr, &self.func_ctx, &BUILTIN_FUNCTIONS).0,
+                    idx,
+                )
+            });
 
             left_join_conditions.push(left_expr.as_remote_expr());
             right_join_conditions.push(right_expr.as_remote_expr());
             left_join_conditions_rt
-                .push(left_expr_for_runtime_filter.map(|expr| expr.as_remote_expr()));
+                .push(left_expr_for_runtime_filter.map(|(expr, idx)| (expr.as_remote_expr(), idx)));
         }
 
         let mut probe_projections = ColumnSet::new();
@@ -461,7 +503,37 @@ impl PhysicalPlanBuilder {
             output_schema,
             need_hold_hash_table: join.need_hold_hash_table,
             stat_info: Some(stat_info),
-            broadcast: join.broadcast,
+            broadcast: is_broadcast,
+            single_to_inner: join.single_to_inner.clone(),
+            enable_bloom_runtime_filter: adjust_bloom_runtime_filter(
+                self.ctx.clone(),
+                &self.metadata,
+                table_index,
+                s_expr,
+            )
+            .await?,
         }))
     }
 }
+
+// Check if enable bloom runtime filter
+async fn adjust_bloom_runtime_filter(
+    ctx: Arc<dyn TableContext>,
+    metadata: &MetadataRef,
+    table_index: Option<IndexType>,
+    s_expr: &SExpr,
+) -> Result<bool> {
+    if let Some(table_index) = table_index {
+        let table = metadata.read().table(table_index).table();
+        if let Some(stats) = table.table_statistics(ctx.clone()).await? {
+            if let Some(num_rows) = stats.num_rows {
+                let join_cardinality = RelExpr::with_s_expr(s_expr)
+                    .derive_cardinality()?
+                    .cardinality;
+                // If the filtered data reduces to less than 1/1000 of the original dataset, we will enable bloom runtime filter.
+                return Ok(join_cardinality <= (num_rows / 1000) as f64);
+            }
+        }
+    }
+    Ok(false)
+}
diff --git a/src/query/sql/src/executor/physical_plans/physical_window.rs b/src/query/sql/src/executor/physical_plans/physical_window.rs
index b7639998d4c8..cf2bf05bcec1 100644
--- a/src/query/sql/src/executor/physical_plans/physical_window.rs
+++ b/src/query/sql/src/executor/physical_plans/physical_window.rs
@@ -55,6 +55,7 @@ pub struct Window {
     pub partition_by: Vec<IndexType>,
     pub order_by: Vec<SortDesc>,
     pub window_frame: WindowFuncFrame,
+    pub limit: Option<usize>,
 }
 
 impl Window {
@@ -357,6 +358,7 @@ impl PhysicalPlanBuilder {
             partition_by: partition_items,
             order_by: order_by_items,
             window_frame: w.frame.clone(),
+            limit: w.limit,
         }))
     }
 }
diff --git a/src/query/sql/src/executor/profile.rs b/src/query/sql/src/executor/profile.rs
deleted file mode 100644
index 189ac0e68534..000000000000
--- a/src/query/sql/src/executor/profile.rs
+++ /dev/null
@@ -1,508 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use databend_common_exception::Result;
-use databend_common_functions::BUILTIN_FUNCTIONS;
-use databend_common_profile::AggregateAttribute;
-use databend_common_profile::AggregateExpandAttribute;
-use databend_common_profile::CteScanAttribute;
-use databend_common_profile::EvalScalarAttribute;
-use databend_common_profile::ExchangeAttribute;
-use databend_common_profile::FilterAttribute;
-use databend_common_profile::JoinAttribute;
-use databend_common_profile::LimitAttribute;
-use databend_common_profile::OperatorAttribute;
-use databend_common_profile::OperatorProfile;
-use databend_common_profile::OperatorType;
-use databend_common_profile::ProcessorProfiles;
-use databend_common_profile::ProjectSetAttribute;
-use databend_common_profile::QueryProfile;
-use databend_common_profile::SortAttribute;
-use databend_common_profile::TableScanAttribute;
-use databend_common_profile::UdfAttribute;
-use databend_common_profile::WindowAttribute;
-use itertools::Itertools;
-
-use crate::executor::format::pretty_display_agg_desc;
-use crate::executor::physical_plans::FragmentKind;
-use crate::executor::physical_plans::WindowFunction;
-use crate::executor::PhysicalPlan;
-use crate::planner::Metadata;
-use crate::MetadataRef;
-
-pub struct ProfileHelper;
-
-impl ProfileHelper {
-    pub fn build_query_profile(
-        query_id: &str,
-        metadata: &MetadataRef,
-        plan: &PhysicalPlan,
-        profs: &ProcessorProfiles,
-    ) -> Result<QueryProfile> {
-        let mut plan_node_profs = vec![];
-        let metadata = metadata.read().clone();
-        flatten_plan_node_profile(&metadata, plan, profs, &mut plan_node_profs)?;
-
-        Ok(QueryProfile::new(query_id.to_string(), plan_node_profs))
-    }
-}
-
-fn flatten_plan_node_profile(
-    metadata: &Metadata,
-    plan: &PhysicalPlan,
-    profs: &ProcessorProfiles,
-    plan_node_profs: &mut Vec<OperatorProfile>,
-) -> Result<()> {
-    match plan {
-        PhysicalPlan::TableScan(scan) => {
-            let table = metadata.table(scan.table_index).clone();
-            let qualified_name = format!("{}.{}", table.database(), table.name());
-            let proc_prof = profs.get(&scan.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: scan.plan_id,
-                operator_type: OperatorType::TableScan,
-                children: vec![],
-                execution_info: proc_prof.into(),
-                attribute: OperatorAttribute::TableScan(TableScanAttribute { qualified_name }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::CteScan(scan) => {
-            let prof = OperatorProfile {
-                id: scan.plan_id,
-                operator_type: OperatorType::CteScan,
-                children: vec![],
-                execution_info: Default::default(),
-                attribute: OperatorAttribute::CteScan(CteScanAttribute {
-                    cte_idx: scan.cte_idx.0,
-                }),
-            };
-            plan_node_profs.push(prof)
-        }
-        PhysicalPlan::ConstantTableScan(scan) => {
-            let proc_prof = profs.get(&scan.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: scan.plan_id,
-                operator_type: OperatorType::ConstantTableScan,
-                children: vec![],
-                execution_info: proc_prof.into(),
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Filter(filter) => {
-            flatten_plan_node_profile(metadata, &filter.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&filter.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: filter.plan_id,
-                operator_type: OperatorType::Filter,
-                children: vec![filter.input.get_id()],
-                execution_info: proc_prof.into(),
-                attribute: OperatorAttribute::Filter(FilterAttribute {
-                    predicate: filter
-                        .predicates
-                        .iter()
-                        .map(|pred| pred.as_expr(&BUILTIN_FUNCTIONS).sql_display())
-                        .join(" AND "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Project(project) => {
-            flatten_plan_node_profile(metadata, &project.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&project.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: project.plan_id,
-                operator_type: OperatorType::Project,
-                children: vec![project.input.get_id()],
-                execution_info: proc_prof.into(),
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::EvalScalar(eval) => {
-            flatten_plan_node_profile(metadata, &eval.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&eval.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: eval.plan_id,
-                operator_type: OperatorType::EvalScalar,
-                execution_info: proc_prof.into(),
-                children: vec![eval.input.get_id()],
-                attribute: OperatorAttribute::EvalScalar(EvalScalarAttribute {
-                    scalars: eval
-                        .exprs
-                        .iter()
-                        .map(|(expr, _)| expr.as_expr(&BUILTIN_FUNCTIONS).sql_display())
-                        .join(", "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::ProjectSet(project_set) => {
-            flatten_plan_node_profile(metadata, &project_set.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&project_set.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: project_set.plan_id,
-                operator_type: OperatorType::ProjectSet,
-                execution_info: proc_prof.into(),
-                children: vec![project_set.input.get_id()],
-                attribute: OperatorAttribute::ProjectSet(ProjectSetAttribute {
-                    functions: project_set
-                        .srf_exprs
-                        .iter()
-                        .map(|(expr, _)| expr.as_expr(&BUILTIN_FUNCTIONS).sql_display())
-                        .join(", "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::AggregateExpand(expand) => {
-            flatten_plan_node_profile(metadata, &expand.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&expand.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: expand.plan_id,
-                operator_type: OperatorType::AggregateExpand,
-                execution_info: proc_prof.into(),
-                children: vec![expand.input.get_id()],
-                attribute: OperatorAttribute::AggregateExpand(AggregateExpandAttribute {
-                    group_keys: expand
-                        .grouping_sets
-                        .sets
-                        .iter()
-                        .map(|columns| {
-                            format!(
-                                "[{}]",
-                                columns
-                                    .iter()
-                                    .map(|column| metadata.column(*column).name())
-                                    .join(", ")
-                            )
-                        })
-                        .join(", "),
-                    aggr_exprs: "".to_string(),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::AggregatePartial(agg_partial) => {
-            flatten_plan_node_profile(metadata, &agg_partial.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&agg_partial.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: agg_partial.plan_id,
-                operator_type: OperatorType::Aggregate,
-                execution_info: proc_prof.into(),
-                children: vec![agg_partial.input.get_id()],
-                attribute: OperatorAttribute::Aggregate(AggregateAttribute {
-                    group_keys: agg_partial
-                        .group_by
-                        .iter()
-                        .map(|column| metadata.column(*column).name())
-                        .join(", "),
-                    functions: agg_partial
-                        .agg_funcs
-                        .iter()
-                        .map(|desc| pretty_display_agg_desc(desc, metadata))
-                        .join(", "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::AggregateFinal(agg_final) => {
-            flatten_plan_node_profile(metadata, &agg_final.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&agg_final.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: agg_final.plan_id,
-                operator_type: OperatorType::Aggregate,
-                execution_info: proc_prof.into(),
-                children: vec![agg_final.input.get_id()],
-                attribute: OperatorAttribute::Aggregate(AggregateAttribute {
-                    group_keys: agg_final
-                        .group_by
-                        .iter()
-                        .map(|column| metadata.column(*column).name())
-                        .join(", "),
-                    functions: agg_final
-                        .agg_funcs
-                        .iter()
-                        .map(|desc| pretty_display_agg_desc(desc, metadata))
-                        .join(", "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Window(window) => {
-            flatten_plan_node_profile(metadata, &window.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&window.plan_id).copied().unwrap_or_default();
-            let partition_by = window
-                .partition_by
-                .iter()
-                .map(|&index| {
-                    let name = metadata.column(index).name();
-                    Ok(name)
-                })
-                .collect::<Result<Vec<_>>>()?
-                .join(", ");
-
-            let order_by = window
-                .order_by
-                .iter()
-                .map(|v| {
-                    let name = metadata.column(v.order_by).name();
-                    Ok(name)
-                })
-                .collect::<Result<Vec<_>>>()?
-                .join(", ");
-
-            let frame = window.window_frame.to_string();
-
-            let func = match &window.func {
-                WindowFunction::Aggregate(agg) => pretty_display_agg_desc(agg, metadata),
-                func => format!("{}", func),
-            };
-            let prof = OperatorProfile {
-                id: window.plan_id,
-                operator_type: OperatorType::Window,
-                children: vec![window.input.get_id()],
-                execution_info: proc_prof.into(),
-                attribute: OperatorAttribute::Window(WindowAttribute {
-                    functions: format!(
-                        "{} OVER (PARTITION BY {} ORDER BY {} {})",
-                        func, partition_by, order_by, frame
-                    ),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Sort(sort) => {
-            flatten_plan_node_profile(metadata, &sort.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&sort.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: sort.plan_id,
-                operator_type: OperatorType::Sort,
-                execution_info: proc_prof.into(),
-                children: vec![sort.input.get_id()],
-                attribute: OperatorAttribute::Sort(SortAttribute {
-                    sort_keys: sort
-                        .order_by
-                        .iter()
-                        .map(|desc| {
-                            format!(
-                                "{} {}",
-                                metadata.column(desc.order_by).name(),
-                                if desc.asc { "ASC" } else { "DESC" }
-                            )
-                        })
-                        .join(", "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Limit(limit) => {
-            flatten_plan_node_profile(metadata, &limit.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&limit.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: limit.plan_id,
-                operator_type: OperatorType::Limit,
-                execution_info: proc_prof.into(),
-                children: vec![limit.input.get_id()],
-                attribute: OperatorAttribute::Limit(LimitAttribute {
-                    limit: limit.limit.unwrap_or_default(),
-                    offset: limit.offset,
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::RowFetch(fetch) => {
-            flatten_plan_node_profile(metadata, &fetch.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&fetch.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: fetch.plan_id,
-                operator_type: OperatorType::RowFetch,
-                execution_info: proc_prof.into(),
-                children: vec![fetch.input.get_id()],
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::HashJoin(hash_join) => {
-            flatten_plan_node_profile(metadata, &hash_join.probe, profs, plan_node_profs)?;
-            flatten_plan_node_profile(metadata, &hash_join.build, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&hash_join.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: hash_join.plan_id,
-                operator_type: OperatorType::Join,
-                execution_info: proc_prof.into(),
-                children: vec![hash_join.probe.get_id(), hash_join.build.get_id()],
-                attribute: OperatorAttribute::Join(JoinAttribute {
-                    join_type: hash_join.join_type.to_string(),
-                    equi_conditions: hash_join
-                        .probe_keys
-                        .iter()
-                        .zip(hash_join.build_keys.iter())
-                        .map(|(l, r)| {
-                            format!(
-                                "{} = {}",
-                                l.as_expr(&BUILTIN_FUNCTIONS).sql_display(),
-                                r.as_expr(&BUILTIN_FUNCTIONS).sql_display(),
-                            )
-                        })
-                        .join(" AND "),
-                    non_equi_conditions: hash_join
-                        .non_equi_conditions
-                        .iter()
-                        .map(|expr| expr.as_expr(&BUILTIN_FUNCTIONS).sql_display())
-                        .join(" AND "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::RangeJoin(range_join) => {
-            flatten_plan_node_profile(metadata, &range_join.left, profs, plan_node_profs)?;
-            flatten_plan_node_profile(metadata, &range_join.right, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&range_join.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: range_join.plan_id,
-                operator_type: OperatorType::Join,
-                children: vec![range_join.left.get_id(), range_join.right.get_id()],
-                execution_info: proc_prof.into(),
-                attribute: OperatorAttribute::Join(JoinAttribute {
-                    join_type: range_join.join_type.to_string(),
-                    equi_conditions: range_join
-                        .conditions
-                        .iter()
-                        .map(|expr| {
-                            format!(
-                                "{} {} {}",
-                                expr.left_expr.as_expr(&BUILTIN_FUNCTIONS).sql_display(),
-                                expr.operator,
-                                expr.right_expr.as_expr(&BUILTIN_FUNCTIONS).sql_display(),
-                            )
-                        })
-                        .join(" AND "),
-                    non_equi_conditions: range_join
-                        .other_conditions
-                        .iter()
-                        .map(|expr| expr.as_expr(&BUILTIN_FUNCTIONS).sql_display())
-                        .join(" AND "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Exchange(exchange) => {
-            flatten_plan_node_profile(metadata, &exchange.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&exchange.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: exchange.plan_id,
-                operator_type: OperatorType::Exchange,
-                execution_info: proc_prof.into(),
-                children: vec![exchange.input.get_id()],
-                attribute: OperatorAttribute::Exchange(ExchangeAttribute {
-                    exchange_mode: match exchange.kind {
-                        FragmentKind::Init => "Init".to_string(),
-                        FragmentKind::Normal => "Hash".to_string(),
-                        FragmentKind::Expansive => "Broadcast".to_string(),
-                        FragmentKind::Merge => "Merge".to_string(),
-                    },
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::UnionAll(union) => {
-            flatten_plan_node_profile(metadata, &union.left, profs, plan_node_profs)?;
-            flatten_plan_node_profile(metadata, &union.right, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&union.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: union.plan_id,
-                operator_type: OperatorType::UnionAll,
-                execution_info: proc_prof.into(),
-                children: vec![union.left.get_id(), union.right.get_id()],
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::DistributedInsertSelect(select) => {
-            flatten_plan_node_profile(metadata, &select.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&select.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: select.plan_id,
-                operator_type: OperatorType::Insert,
-                execution_info: proc_prof.into(),
-                children: vec![],
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::ExchangeSource(source) => {
-            let proc_prof = profs.get(&source.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: source.plan_id,
-                operator_type: OperatorType::Exchange,
-                execution_info: proc_prof.into(),
-                children: vec![],
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::ExchangeSink(sink) => {
-            flatten_plan_node_profile(metadata, &sink.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&sink.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: sink.plan_id,
-                operator_type: OperatorType::Exchange,
-                execution_info: proc_prof.into(),
-                children: vec![],
-                attribute: OperatorAttribute::Empty,
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::Udf(udf) => {
-            flatten_plan_node_profile(metadata, &udf.input, profs, plan_node_profs)?;
-            let proc_prof = profs.get(&udf.plan_id).copied().unwrap_or_default();
-            let prof = OperatorProfile {
-                id: udf.plan_id,
-                operator_type: OperatorType::Udf,
-                execution_info: proc_prof.into(),
-                children: vec![udf.input.get_id()],
-                attribute: OperatorAttribute::Udf(UdfAttribute {
-                    scalars: udf
-                        .udf_funcs
-                        .iter()
-                        .map(|func| {
-                            let arg_exprs = func.arg_exprs.join(", ");
-                            format!("{}({})", func.func_name, arg_exprs)
-                        })
-                        .join(", "),
-                }),
-            };
-            plan_node_profs.push(prof);
-        }
-        PhysicalPlan::MaterializedCte(_) => todo!(),
-        PhysicalPlan::DeleteSource(_)
-        | PhysicalPlan::CommitSink(_)
-        | PhysicalPlan::CopyIntoTable(_)
-        | PhysicalPlan::ReplaceAsyncSourcer(_)
-        | PhysicalPlan::MergeInto(_)
-        | PhysicalPlan::MergeIntoAddRowNumber(_)
-        | PhysicalPlan::MergeIntoAppendNotMatched(_)
-        | PhysicalPlan::MergeIntoSource(_)
-        | PhysicalPlan::ReplaceDeduplicate(_)
-        | PhysicalPlan::ReplaceInto(_)
-        | PhysicalPlan::CompactSource(_)
-        | PhysicalPlan::ReclusterSource(_)
-        | PhysicalPlan::ReclusterSink(_)
-        | PhysicalPlan::UpdateSource(_) => unreachable!(),
-    }
-
-    Ok(())
-}
diff --git a/src/query/sql/src/lib.rs b/src/query/sql/src/lib.rs
index b9ce4e6ed454..d66b138eb680 100644
--- a/src/query/sql/src/lib.rs
+++ b/src/query/sql/src/lib.rs
@@ -19,6 +19,7 @@
 #![feature(try_blocks)]
 #![feature(extend_one)]
 #![feature(lazy_cell)]
+#![feature(if_let_guard)]
 
 pub mod evaluator;
 pub mod executor;
diff --git a/src/query/sql/src/planner/binder/binder.rs b/src/query/sql/src/planner/binder/binder.rs
index 452bcf0e9084..d1869a138c57 100644
--- a/src/query/sql/src/planner/binder/binder.rs
+++ b/src/query/sql/src/planner/binder/binder.rs
@@ -161,7 +161,7 @@ impl<'a> Binder {
                 ConstantFolder::fold(&expr, &self.ctx.get_function_context()?, &BUILTIN_FUNCTIONS);
             match new_expr {
                 Expr::Constant { scalar, .. } => {
-                    let value = String::from_utf8(scalar.into_string().unwrap())?;
+                    let value = scalar.into_string().unwrap();
                     if variable.to_lowercase().as_str() == "timezone" {
                         let tz = value.trim_matches(|c| c == '\'' || c == '\"');
                         tz.parse::<Tz>().map_err(|_| {
diff --git a/src/query/sql/src/planner/binder/ddl/task.rs b/src/query/sql/src/planner/binder/ddl/task.rs
index f84dc9a8f114..ec748a625b06 100644
--- a/src/query/sql/src/planner/binder/ddl/task.rs
+++ b/src/query/sql/src/planner/binder/ddl/task.rs
@@ -93,6 +93,7 @@ impl Binder {
             after,
             when_condition,
             sql,
+            session_parameters,
         } = stmt;
         if (schedule_opts.is_none() && after.is_empty())
             || (schedule_opts.is_some() && !after.is_empty())
@@ -114,6 +115,7 @@ impl Binder {
             after: after.clone(),
             when_condition: when_condition.clone(),
             comment: comments.clone(),
+            session_parameters: session_parameters.clone(),
             sql: sql.clone(),
         };
         Ok(Plan::CreateTask(Box::new(plan)))
@@ -135,12 +137,14 @@ impl Binder {
             schedule,
             suspend_task_after_num_failures,
             comments,
+            session_parameters,
         } = options
         {
             if warehouse.is_none()
                 && schedule.is_none()
                 && suspend_task_after_num_failures.is_none()
                 && comments.is_none()
+                && session_parameters.is_none()
             {
                 return Err(ErrorCode::SyntaxException(
                     "alter task must set at least one option".to_string(),
diff --git a/src/query/sql/src/planner/binder/join.rs b/src/query/sql/src/planner/binder/join.rs
index 336187445b43..be9a8ff7c0a8 100644
--- a/src/query/sql/src/planner/binder/join.rs
+++ b/src/query/sql/src/planner/binder/join.rs
@@ -209,6 +209,7 @@ impl Binder {
         }
 
         let right_prop = RelExpr::with_s_expr(&right_child).derive_relational_prop()?;
+        let mut is_lateral = false;
         if !right_prop.outer_columns.is_empty() {
             // If there are outer columns in right child, then the join is a correlated lateral join
             let mut decorrelator = SubqueryRewriter::new(self.metadata.clone());
@@ -229,6 +230,7 @@ impl Binder {
             if join_type == JoinType::Cross {
                 join_type = JoinType::Inner;
             }
+            is_lateral = true;
         }
 
         let logical_join = Join {
@@ -239,7 +241,8 @@ impl Binder {
             marker_index: None,
             from_correlated_subquery: false,
             need_hold_hash_table: false,
-            broadcast: false,
+            is_lateral,
+            single_to_inner: None,
         };
         Ok(SExpr::create_binary(
             Arc::new(logical_join.into()),
diff --git a/src/query/sql/src/planner/binder/project.rs b/src/query/sql/src/planner/binder/project.rs
index e5a8c12b7239..efa59140cffd 100644
--- a/src/query/sql/src/planner/binder/project.rs
+++ b/src/query/sql/src/planner/binder/project.rs
@@ -130,6 +130,7 @@ impl Binder {
                         projection_index: Some(column_binding.index),
                         data_type,
                         outer_columns,
+                        contain_agg: None,
                     })
                 } else {
                     item.scalar.clone()
diff --git a/src/query/sql/src/planner/binder/scalar_common.rs b/src/query/sql/src/planner/binder/scalar_common.rs
index 695dc75f5c6b..c7da9957fba6 100644
--- a/src/query/sql/src/planner/binder/scalar_common.rs
+++ b/src/query/sql/src/planner/binder/scalar_common.rs
@@ -22,7 +22,6 @@ use crate::optimizer::RelationalProperty;
 use crate::plans::walk_expr;
 use crate::plans::BoundColumnRef;
 use crate::plans::CastExpr;
-use crate::plans::ComparisonOp;
 use crate::plans::FunctionCall;
 use crate::plans::ScalarExpr;
 use crate::plans::Visitor;
@@ -116,7 +115,7 @@ pub enum JoinPredicate<'a> {
     Both {
         left: &'a ScalarExpr,
         right: &'a ScalarExpr,
-        op: ComparisonOp,
+        is_equal_op: bool,
     },
     Other(&'a ScalarExpr),
 }
@@ -127,10 +126,6 @@ impl<'a> JoinPredicate<'a> {
         left_prop: &RelationalProperty,
         right_prop: &RelationalProperty,
     ) -> Self {
-        if contain_subquery(scalar) {
-            return Self::Other(scalar);
-        }
-
         if scalar.used_columns().is_empty() {
             return Self::ALL(scalar);
         }
@@ -144,20 +139,28 @@ impl<'a> JoinPredicate<'a> {
         }
 
         if let ScalarExpr::FunctionCall(func) = scalar {
-            if let Some(op) = ComparisonOp::try_from_func_name(func.func_name.as_str()) {
-                let left = &func.arguments[0];
-                let right = &func.arguments[1];
-                if satisfied_by(left, left_prop) && satisfied_by(right, right_prop) {
-                    return Self::Both { left, right, op };
-                }
+            if func.arguments.len() != 2 {
+                return Self::Other(scalar);
+            }
 
-                if satisfied_by(right, left_prop) && satisfied_by(left, right_prop) {
-                    return Self::Both {
-                        left: right,
-                        right: left,
-                        op,
-                    };
-                }
+            let is_equal_op = func.func_name.as_str() == "eq";
+            let left = &func.arguments[0];
+            let right = &func.arguments[1];
+
+            if satisfied_by(left, left_prop) && satisfied_by(right, right_prop) {
+                return Self::Both {
+                    left,
+                    right,
+                    is_equal_op,
+                };
+            }
+
+            if satisfied_by(right, left_prop) && satisfied_by(left, right_prop) {
+                return Self::Both {
+                    left: right,
+                    right: left,
+                    is_equal_op,
+                };
             }
         }
 
diff --git a/src/query/sql/src/planner/binder/setting.rs b/src/query/sql/src/planner/binder/setting.rs
index d2a505b05944..51923f9d025a 100644
--- a/src/query/sql/src/planner/binder/setting.rs
+++ b/src/query/sql/src/planner/binder/setting.rs
@@ -58,7 +58,7 @@ impl Binder {
             ConstantFolder::fold(&expr, &self.ctx.get_function_context()?, &BUILTIN_FUNCTIONS);
         match new_expr {
             databend_common_expression::Expr::Constant { scalar, .. } => {
-                let value = String::from_utf8(scalar.into_string().unwrap())?;
+                let value = scalar.into_string().unwrap();
                 let vars = vec![VarValue {
                     is_global,
                     variable,
diff --git a/src/query/sql/src/planner/binder/table.rs b/src/query/sql/src/planner/binder/table.rs
index abcd91f14244..6fcc6634d2e7 100644
--- a/src/query/sql/src/planner/binder/table.rs
+++ b/src/query/sql/src/planner/binder/table.rs
@@ -100,6 +100,7 @@ use crate::optimizer::SExpr;
 use crate::planner::semantic::normalize_identifier;
 use crate::planner::semantic::TypeChecker;
 use crate::plans::CteScan;
+use crate::plans::DummyTableScan;
 use crate::plans::EvalScalar;
 use crate::plans::FunctionCall;
 use crate::plans::RelOperator;
@@ -135,24 +136,11 @@ impl Binder {
                 }
             }
         }
-        let catalog = CATALOG_DEFAULT;
-        let database = "system";
-        let tenant = self.ctx.get_tenant();
-        let table_meta = self
-            .resolve_data_source(tenant.as_str(), catalog, database, "one", &None)
-            .await?;
-        let table_index = self.metadata.write().add_table(
-            CATALOG_DEFAULT.to_owned(),
-            database.to_string(),
-            table_meta,
-            None,
-            false,
-            false,
-            false,
-        );
-
-        self.bind_base_table(bind_context, database, table_index, None)
-            .await
+        let bind_context = BindContext::with_parent(Box::new(bind_context.clone()));
+        Ok((
+            SExpr::create_leaf(Arc::new(DummyTableScan.into())),
+            bind_context,
+        ))
     }
 
     fn check_view_dep(bind_context: &BindContext, database: &str, view_name: &str) -> Result<()> {
@@ -1494,8 +1482,7 @@ impl Binder {
 // copy from common-storages-fuse to avoid cyclic dependency.
 fn string_value(value: &Scalar) -> Result<String> {
     match value {
-        Scalar::String(val) => String::from_utf8(val.clone())
-            .map_err(|e| ErrorCode::BadArguments(format!("invalid string. {}", e))),
+        Scalar::String(val) => Ok(val.clone()),
         other => Err(ErrorCode::BadArguments(format!(
             "Expected a string value, but found a '{}'.",
             other
diff --git a/src/query/sql/src/planner/binder/window.rs b/src/query/sql/src/planner/binder/window.rs
index 304de507fc27..d44cfb8bb988 100644
--- a/src/query/sql/src/planner/binder/window.rs
+++ b/src/query/sql/src/planner/binder/window.rs
@@ -60,6 +60,7 @@ impl Binder {
             partition_by: window_info.partition_by_items.clone(),
             order_by: window_info.order_by_items.clone(),
             frame: window_info.frame.clone(),
+            limit: None,
         };
 
         Ok(SExpr::create_unary(
diff --git a/src/query/sql/src/planner/format/display_rel_operator.rs b/src/query/sql/src/planner/format/display_rel_operator.rs
index 13c7835f9ce5..73dfb39382ec 100644
--- a/src/query/sql/src/planner/format/display_rel_operator.rs
+++ b/src/query/sql/src/planner/format/display_rel_operator.rs
@@ -71,7 +71,6 @@ impl Display for FormatContext {
                 RelOperator::Limit(_) => write!(f, "Limit"),
                 RelOperator::Exchange(op) => format_exchange(f, metadata, op),
                 RelOperator::UnionAll(_) => write!(f, "Union"),
-                RelOperator::Pattern(_) => write!(f, "Pattern"),
                 RelOperator::DummyTableScan(_) => write!(f, "DummyTableScan"),
                 RelOperator::Window(_) => write!(f, "WindowFunc"),
                 RelOperator::ProjectSet(_) => write!(f, "ProjectSet"),
diff --git a/src/query/sql/src/planner/mod.rs b/src/query/sql/src/planner/mod.rs
index 1e6ff878462f..3380ecc260bb 100644
--- a/src/query/sql/src/planner/mod.rs
+++ b/src/query/sql/src/planner/mod.rs
@@ -39,6 +39,7 @@ pub use bloom_index::BloomIndexColumns;
 pub use expression_parser::*;
 pub use format::format_scalar;
 pub use metadata::*;
+pub use planner::get_query_kind;
 pub use planner::PlanExtras;
 pub use planner::Planner;
 pub use plans::insert::InsertInputSource;
diff --git a/src/query/sql/src/planner/optimizer/cascades/tasks/optimize_expr.rs b/src/query/sql/src/planner/optimizer/cascades/tasks/optimize_expr.rs
index b9750498bac7..4cb8645be82e 100644
--- a/src/query/sql/src/planner/optimizer/cascades/tasks/optimize_expr.rs
+++ b/src/query/sql/src/planner/optimizer/cascades/tasks/optimize_expr.rs
@@ -26,6 +26,7 @@ use crate::optimizer::cascades::tasks::SharedCounter;
 use crate::optimizer::cascades::CascadesOptimizer;
 use crate::optimizer::cost::Cost;
 use crate::optimizer::cost::CostContext;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::Distribution;
 use crate::optimizer::DistributionEnforcer;
 use crate::optimizer::Enforcer;
@@ -33,8 +34,6 @@ use crate::optimizer::PatternExtractor;
 use crate::optimizer::RelExpr;
 use crate::optimizer::RequiredProperty;
 use crate::optimizer::SExpr;
-use crate::plans::PatternPlan;
-use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::IndexType;
 
@@ -401,19 +400,10 @@ impl OptimizeExprTask {
             return Ok(OptimizeExprEvent::OptimizingSelf);
         }
 
-        let mut extractor = PatternExtractor::create();
+        let mut extractor = PatternExtractor::new();
         let enforcer_child = Arc::new(
             extractor
-                .extract(
-                    &optimizer.memo,
-                    m_expr,
-                    &SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    )),
-                )?
+                .extract(&optimizer.memo, m_expr, &Matcher::Leaf)?
                 .pop()
                 .ok_or_else(|| {
                     ErrorCode::Internal(format!("Cannot find child of m_expr: {:?}", m_expr.plan))
diff --git a/src/query/sql/src/planner/optimizer/decorrelate/decorrelate.rs b/src/query/sql/src/planner/optimizer/decorrelate/decorrelate.rs
index bf5d319d4db3..b55045ce8675 100644
--- a/src/query/sql/src/planner/optimizer/decorrelate/decorrelate.rs
+++ b/src/query/sql/src/planner/optimizer/decorrelate/decorrelate.rs
@@ -25,16 +25,15 @@ use crate::binder::Visibility;
 use crate::optimizer::decorrelate::subquery_rewriter::FlattenInfo;
 use crate::optimizer::decorrelate::subquery_rewriter::SubqueryRewriter;
 use crate::optimizer::decorrelate::subquery_rewriter::UnnestResult;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::ColumnSet;
 use crate::optimizer::RelExpr;
 use crate::optimizer::SExpr;
 use crate::plans::BoundColumnRef;
-use crate::plans::ComparisonOp;
 use crate::plans::Filter;
 use crate::plans::FunctionCall;
 use crate::plans::Join;
 use crate::plans::JoinType;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 use crate::plans::SubqueryExpr;
@@ -87,63 +86,34 @@ impl SubqueryRewriter {
         //         EvalScalar
         //          \
         //           Get
-        let patterns = vec![
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Filter,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Scan,
-                        }
-                        .into(),
-                    ))),
-                )),
-            ),
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Filter,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::EvalScalar,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_leaf(Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Scan,
-                            }
-                            .into(),
-                        ))),
-                    )),
-                )),
-            ),
+        let matchers = vec![
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Filter,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Scan,
+                        children: vec![],
+                    }],
+                }],
+            },
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Filter,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::EvalScalar,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Scan,
+                            children: vec![],
+                        }],
+                    }],
+                }],
+            },
         ];
         let mut matched = false;
-        for pattern in patterns {
-            if subquery.subquery.match_pattern(&pattern) {
+        for matcher in matchers {
+            if matcher.matches(&subquery.subquery) {
                 matched = true;
                 break;
             }
@@ -195,9 +165,12 @@ impl SubqueryRewriter {
                 }
 
                 JoinPredicate::Both {
-                    left, right, op, ..
+                    left,
+                    right,
+                    is_equal_op,
+                    ..
                 } => {
-                    if op == ComparisonOp::Equal {
+                    if is_equal_op {
                         left_conditions.push(left.clone());
                         right_conditions.push(right.clone());
                     } else {
@@ -221,7 +194,8 @@ impl SubqueryRewriter {
             marker_index: None,
             from_correlated_subquery: true,
             need_hold_hash_table: false,
-            broadcast: false,
+            is_lateral: false,
+            single_to_inner: None,
         };
 
         // Rewrite plan to semi-join.
@@ -288,15 +262,30 @@ impl SubqueryRewriter {
                     &mut right_conditions,
                     &mut left_conditions,
                 )?;
+
+                let mut join_type = JoinType::LeftSingle;
+                if subquery.contain_agg.unwrap() {
+                    let rel_expr = RelExpr::with_s_expr(&subquery.subquery);
+                    let has_precise_cardinality = rel_expr
+                        .derive_cardinality()?
+                        .statistics
+                        .precise_cardinality
+                        .is_some();
+                    if has_precise_cardinality {
+                        join_type = JoinType::Left;
+                    }
+                }
+
                 let join_plan = Join {
                     left_conditions,
                     right_conditions,
                     non_equi_conditions: vec![],
-                    join_type: JoinType::LeftSingle,
+                    join_type,
                     marker_index: None,
                     from_correlated_subquery: true,
                     need_hold_hash_table: false,
-                    broadcast: false,
+                    is_lateral: false,
+                    single_to_inner: None,
                 };
                 let s_expr = SExpr::create_binary(
                     Arc::new(join_plan.into()),
@@ -344,7 +333,8 @@ impl SubqueryRewriter {
                     marker_index: Some(marker_index),
                     from_correlated_subquery: true,
                     need_hold_hash_table: false,
-                    broadcast: false,
+                    is_lateral: false,
+                    single_to_inner: None,
                 };
                 let s_expr = SExpr::create_binary(
                     Arc::new(join_plan.into()),
@@ -407,7 +397,8 @@ impl SubqueryRewriter {
                     marker_index: Some(marker_index),
                     from_correlated_subquery: true,
                     need_hold_hash_table: false,
-                    broadcast: false,
+                    is_lateral: false,
+                    single_to_inner: None,
                 }
                 .into();
                 Ok((
@@ -441,6 +432,7 @@ impl SubqueryRewriter {
                     Box::from(column_entry.data_type()),
                     Visibility::Visible,
                 )
+                .table_index(column_entry.table_index())
                 .build(),
             });
             let derive_column = self.derived_columns.get(correlated_column).unwrap();
@@ -453,6 +445,7 @@ impl SubqueryRewriter {
                     Box::from(column_entry.data_type()),
                     Visibility::Visible,
                 )
+                .table_index(column_entry.table_index())
                 .build(),
             });
             left_conditions.push(left_column);
diff --git a/src/query/sql/src/planner/optimizer/decorrelate/flatten_plan.rs b/src/query/sql/src/planner/optimizer/decorrelate/flatten_plan.rs
index 9a3058ff7105..ee98642ae48f 100644
--- a/src/query/sql/src/planner/optimizer/decorrelate/flatten_plan.rs
+++ b/src/query/sql/src/planner/optimizer/decorrelate/flatten_plan.rs
@@ -132,7 +132,8 @@ impl SubqueryRewriter {
                 marker_index: None,
                 from_correlated_subquery: false,
                 need_hold_hash_table: false,
-                broadcast: false,
+                is_lateral: false,
+                single_to_inner: None,
             }
             .into();
 
@@ -460,7 +461,8 @@ impl SubqueryRewriter {
                     marker_index: join.marker_index,
                     from_correlated_subquery: false,
                     need_hold_hash_table: false,
-                    broadcast: false,
+                    is_lateral: false,
+                    single_to_inner: None,
                 }
                 .into(),
             ),
diff --git a/src/query/sql/src/planner/optimizer/decorrelate/subquery_rewriter.rs b/src/query/sql/src/planner/optimizer/decorrelate/subquery_rewriter.rs
index eb26f4d551f7..b2f4eeddcf73 100644
--- a/src/query/sql/src/planner/optimizer/decorrelate/subquery_rewriter.rs
+++ b/src/query/sql/src/planner/optimizer/decorrelate/subquery_rewriter.rs
@@ -478,7 +478,8 @@ impl SubqueryRewriter {
                     marker_index: None,
                     from_correlated_subquery: false,
                     need_hold_hash_table: false,
-                    broadcast: false,
+                    is_lateral: false,
+                    single_to_inner: None,
                 }
                 .into();
                 Ok((
@@ -547,7 +548,8 @@ impl SubqueryRewriter {
                     marker_index: Some(marker_index),
                     from_correlated_subquery: false,
                     need_hold_hash_table: false,
-                    broadcast: false,
+                    is_lateral: false,
+                    single_to_inner: None,
                 }
                 .into();
                 let s_expr = SExpr::create_binary(
@@ -577,7 +579,8 @@ impl SubqueryRewriter {
             marker_index: None,
             from_correlated_subquery: false,
             need_hold_hash_table: false,
-            broadcast: false,
+            is_lateral: false,
+            single_to_inner: None,
         }
         .into();
 
diff --git a/src/query/sql/src/planner/optimizer/distributed/distributed_merge.rs b/src/query/sql/src/planner/optimizer/distributed/distributed_merge.rs
index 54bf1bd8b1c4..3ad4c230d99e 100644
--- a/src/query/sql/src/planner/optimizer/distributed/distributed_merge.rs
+++ b/src/query/sql/src/planner/optimizer/distributed/distributed_merge.rs
@@ -16,21 +16,21 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::SExpr;
 use crate::plans::AddRowNumber;
 use crate::plans::Exchange::Broadcast;
 use crate::plans::Join;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 pub struct MergeSourceOptimizer {
-    pub merge_source_pattern: SExpr,
+    pub merge_source_matcher: Matcher,
 }
 
 impl MergeSourceOptimizer {
     pub fn create() -> Self {
         Self {
-            merge_source_pattern: Self::merge_source_pattern(),
+            merge_source_matcher: Self::merge_source_matcher(),
         }
     }
 
@@ -68,14 +68,13 @@ impl MergeSourceOptimizer {
 
         let mut join: Join = join_s_expr.plan().clone().try_into()?;
         join.need_hold_hash_table = true;
-        join.broadcast = true;
         let mut join_s_expr = join_s_expr.replace_plan(Arc::new(RelOperator::Join(join)));
         join_s_expr = join_s_expr.replace_children(new_join_children);
         Ok(s_expr.replace_children(vec![Arc::new(join_s_expr)]))
     }
 
     // for right outer join (source as build)
-    fn merge_source_pattern() -> SExpr {
+    fn merge_source_matcher() -> Matcher {
         // Input:
         //       Exchange(Merge)
         //          |
@@ -97,49 +96,21 @@ impl MergeSourceOptimizer {
         //    |          AddRowNumber
         //    |               |
         //    *               *
-        SExpr::create_unary(
-            Arc::new(
-                PatternPlan {
-                    plan_type: RelOp::Exchange,
-                }
-                .into(),
-            ),
-            Arc::new(SExpr::create_binary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Join,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Exchange,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Exchange,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )),
-        )
+        Matcher::MatchOp {
+            op_type: RelOp::Exchange,
+            children: vec![Matcher::MatchOp {
+                op_type: RelOp::Join,
+                children: vec![
+                    Matcher::MatchOp {
+                        op_type: RelOp::Exchange,
+                        children: vec![Matcher::Leaf],
+                    },
+                    Matcher::MatchOp {
+                        op_type: RelOp::Exchange,
+                        children: vec![Matcher::Leaf],
+                    },
+                ],
+            }],
+        }
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/distributed/sort_and_limit.rs b/src/query/sql/src/planner/optimizer/distributed/sort_and_limit.rs
index 526253e9f68c..2dc99aac4034 100644
--- a/src/query/sql/src/planner/optimizer/distributed/sort_and_limit.rs
+++ b/src/query/sql/src/planner/optimizer/distributed/sort_and_limit.rs
@@ -16,30 +16,30 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::SExpr;
 use crate::plans::Exchange;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::plans::Sort;
 
 pub struct SortAndLimitPushDownOptimizer {
-    sort_pattern: SExpr,
-    limit_pattern: SExpr,
+    sort_matcher: Matcher,
+    limit_matcher: Matcher,
 }
 
 impl SortAndLimitPushDownOptimizer {
     pub fn create() -> Self {
         Self {
-            sort_pattern: Self::sort_pattern(),
-            limit_pattern: Self::limit_pattern(),
+            sort_matcher: Self::sort_matcher(),
+            limit_matcher: Self::limit_matcher(),
         }
     }
 
     /// `limit` is already pushed down to `Sort`,
     /// so the TopN scenario is already contained in this pattern.
-    fn sort_pattern() -> SExpr {
+    fn sort_matcher() -> Matcher {
         // Input:
         //   Sort
         //    \
@@ -54,31 +54,16 @@ impl SortAndLimitPushDownOptimizer {
         //       Sort (after_exchange = false)
         //        \
         //         *
-        SExpr::create_unary(
-            Arc::new(
-                PatternPlan {
-                    plan_type: RelOp::Sort,
-                }
-                .into(),
-            ),
-            Arc::new(SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Exchange,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )),
-        )
+        Matcher::MatchOp {
+            op_type: RelOp::Sort,
+            children: vec![Matcher::MatchOp {
+                op_type: RelOp::Exchange,
+                children: vec![Matcher::Leaf],
+            }],
+        }
     }
 
-    fn limit_pattern() -> SExpr {
+    fn limit_matcher() -> Matcher {
         // Input:
         // Limit
         //  \
@@ -93,28 +78,13 @@ impl SortAndLimitPushDownOptimizer {
         //     Limit
         //      \
         //       *
-        SExpr::create_unary(
-            Arc::new(
-                PatternPlan {
-                    plan_type: RelOp::Limit,
-                }
-                .into(),
-            ),
-            Arc::new(SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Exchange,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )),
-        )
+        Matcher::MatchOp {
+            op_type: RelOp::Limit,
+            children: vec![Matcher::MatchOp {
+                op_type: RelOp::Exchange,
+                children: vec![Matcher::Leaf],
+            }],
+        }
     }
 
     pub fn optimize(&self, s_expr: &SExpr) -> Result<SExpr> {
@@ -129,7 +99,7 @@ impl SortAndLimitPushDownOptimizer {
     }
 
     fn apply_sort(&self, s_expr: &SExpr) -> Result<SExpr> {
-        if !s_expr.match_pattern(&self.sort_pattern) {
+        if !self.sort_matcher.matches(s_expr) {
             return Ok(s_expr.clone());
         }
 
@@ -153,7 +123,7 @@ impl SortAndLimitPushDownOptimizer {
     }
 
     fn apply_limit(&self, s_expr: &SExpr) -> Result<SExpr> {
-        if !s_expr.match_pattern(&self.limit_pattern) {
+        if !self.limit_matcher.matches(s_expr) {
             return Ok(s_expr.clone());
         }
 
diff --git a/src/query/sql/src/planner/optimizer/extract.rs b/src/query/sql/src/planner/optimizer/extract.rs
new file mode 100644
index 000000000000..b6996076489c
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/extract.rs
@@ -0,0 +1,276 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+
+use databend_common_exception::ErrorCode;
+use databend_common_exception::Result;
+
+use crate::optimizer::group::Group;
+use crate::optimizer::m_expr::MExpr;
+use crate::optimizer::memo::Memo;
+use crate::optimizer::SExpr;
+use crate::plans::Operator;
+use crate::plans::RelOp;
+use crate::plans::RelOperator;
+
+/// A matcher used to describe a pattern to be matched.
+pub enum Matcher {
+    /// A pattern to match an operator with children.
+    MatchOp {
+        /// The type of the operator to match.
+        op_type: RelOp,
+        children: Vec<Self>,
+    },
+    /// A pattern to match an operator with a predicate.
+    MatchFn {
+        /// The function to match the operator.
+        predicate: Box<dyn Fn(&RelOperator) -> bool + 'static>,
+        children: Vec<Self>,
+    },
+    /// A leaf pattern to match any node.
+    Leaf,
+}
+
+impl Matcher {
+    /// Check if the `SExpr` can be matched by the `Matcher`.
+    pub fn matches(&self, s_expr: &SExpr) -> bool {
+        match self {
+            Matcher::MatchOp { op_type, children } => {
+                if s_expr.plan().rel_op() != *op_type {
+                    return false;
+                }
+
+                if s_expr.plan().arity() != children.len() {
+                    return false;
+                }
+
+                for (child_expr, child_matcher) in s_expr.children.iter().zip(children.iter()) {
+                    if !child_matcher.matches(child_expr) {
+                        return false;
+                    }
+                }
+
+                true
+            }
+
+            Matcher::MatchFn {
+                predicate,
+                children,
+            } => {
+                if !predicate(s_expr.plan()) {
+                    return false;
+                }
+
+                if s_expr.plan().arity() != children.len() {
+                    return false;
+                }
+
+                for (child_expr, child_matcher) in s_expr.children.iter().zip(children.iter()) {
+                    if !child_matcher.matches(child_expr) {
+                        return false;
+                    }
+                }
+
+                true
+            }
+
+            Matcher::Leaf => true,
+        }
+    }
+
+    /// Check if the `RelOperator` can be matched by the `Matcher`.
+    pub fn matches_op(&self, op: &RelOperator) -> bool {
+        match self {
+            Matcher::MatchOp { op_type, children } => {
+                op.rel_op() == *op_type && op.arity() == children.len()
+            }
+
+            Matcher::MatchFn {
+                predicate,
+                children,
+            } => predicate(op) && op.arity() == children.len(),
+
+            Matcher::Leaf => true,
+        }
+    }
+}
+
+/// A helper to extract `SExpr`s from `Memo` that match the given pattern.
+pub struct PatternExtractor {}
+
+impl Default for PatternExtractor {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl PatternExtractor {
+    pub fn new() -> Self {
+        PatternExtractor {}
+    }
+
+    pub fn extract(
+        &mut self,
+        memo: &Memo,
+        m_expr: &MExpr,
+        matcher: &Matcher,
+    ) -> Result<Vec<SExpr>> {
+        match matcher {
+            Matcher::MatchOp { children, .. } => {
+                if !matcher.matches_op(&m_expr.plan) {
+                    return Ok(vec![]);
+                }
+
+                let mut children_results = Vec::with_capacity(m_expr.arity());
+                for (child_index, child_matcher) in m_expr.children.iter().zip(children.iter()) {
+                    let child_group = memo.group(*child_index)?;
+                    let result = self.extract_group(memo, child_group, child_matcher)?;
+                    children_results.push(result);
+                }
+
+                Self::generate_expression_with_children(memo, m_expr, children_results)
+            }
+
+            Matcher::MatchFn {
+                predicate,
+                children,
+            } => {
+                if !predicate(&m_expr.plan) {
+                    return Ok(vec![]);
+                }
+
+                let mut children_results = Vec::with_capacity(m_expr.arity());
+                for (child_index, child_matcher) in m_expr.children.iter().zip(children.iter()) {
+                    let child_group = memo.group(*child_index)?;
+                    let result = self.extract_group(memo, child_group, child_matcher)?;
+                    children_results.push(result);
+                }
+
+                Self::generate_expression_with_children(memo, m_expr, children_results)
+            }
+
+            Matcher::Leaf => {
+                // Expand the pattern node to a complete `SExpr` if it's a leaf node.
+                let child = Self::expand_pattern(memo, m_expr)?;
+                Ok(vec![child])
+            }
+        }
+    }
+
+    fn extract_group(
+        &mut self,
+        memo: &Memo,
+        group: &Group,
+        matcher: &Matcher,
+    ) -> Result<Vec<SExpr>> {
+        let mut results = vec![];
+        for m_expr in group.m_exprs.iter() {
+            let result = self.extract(memo, m_expr, matcher)?;
+            results.extend(result.into_iter());
+        }
+
+        Ok(results)
+    }
+
+    fn generate_expression_with_children(
+        memo: &Memo,
+        m_expr: &MExpr,
+        candidates: Vec<Vec<SExpr>>,
+    ) -> Result<Vec<SExpr>> {
+        let mut results = vec![];
+
+        // Initialize cursors
+        let mut cursors: Vec<usize> = vec![];
+        for candidate in candidates.iter() {
+            if candidate.is_empty() {
+                // Every child should have at least one candidate
+                return Ok(results);
+            }
+            cursors.push(0);
+        }
+
+        if cursors.is_empty() {
+            results.push(SExpr::create(
+                m_expr.plan.clone(),
+                vec![],
+                Some(m_expr.group_index),
+                Some(memo.group(m_expr.group_index)?.relational_prop.clone()),
+                Some(memo.group(m_expr.group_index)?.stat_info.clone()),
+            ));
+            return Ok(results);
+        }
+
+        'LOOP: loop {
+            let mut children = vec![];
+            for (index, cursor) in cursors.iter().enumerate() {
+                children.push(Arc::new(candidates[index][*cursor].clone()));
+            }
+            results.push(SExpr::create(
+                m_expr.plan.clone(),
+                children,
+                Some(m_expr.group_index),
+                Some(memo.group(m_expr.group_index)?.relational_prop.clone()),
+                Some(memo.group(m_expr.group_index)?.stat_info.clone()),
+            ));
+
+            let mut shifted = false;
+            // Shift cursor
+            for i in (0..cursors.len()).rev() {
+                if !shifted {
+                    // Shift cursor
+                    cursors[i] += 1;
+                    shifted = true;
+                }
+
+                if i == 0 && cursors[0] > candidates[0].len() - 1 {
+                    // Candidates are exhausted
+                    break 'LOOP;
+                } else if i > 0 && cursors[i] > candidates[i].len() - 1 {
+                    // Shift previous children
+                    cursors[i] = 0;
+                    cursors[i - 1] += 1;
+                    continue;
+                } else {
+                    break;
+                }
+            }
+        }
+
+        Ok(results)
+    }
+
+    /// Expand a `Pattern` node to an arbitrary `SExpr` with `m_expr` as the root.
+    /// Since we don't care about the actual content of the `Pattern` node, we will
+    /// choose the first `MExpr` in each group to construct the `SExpr`.
+    fn expand_pattern(memo: &Memo, m_expr: &MExpr) -> Result<SExpr> {
+        let mut children = Vec::with_capacity(m_expr.arity());
+        for child in m_expr.children.iter() {
+            let child_group = memo.group(*child)?;
+            let child_m_expr = child_group
+                .m_exprs
+                .first()
+                .ok_or_else(|| ErrorCode::Internal(format!("No MExpr in group {child}")))?;
+            children.push(Arc::new(Self::expand_pattern(memo, child_m_expr)?));
+        }
+
+        Ok(SExpr::create(
+            m_expr.plan.clone(),
+            children,
+            Some(m_expr.group_index),
+            Some(memo.group(m_expr.group_index)?.relational_prop.clone()),
+            Some(memo.group(m_expr.group_index)?.stat_info.clone()),
+        ))
+    }
+}
diff --git a/src/query/sql/src/planner/optimizer/filter/deduplicate_join_condition.rs b/src/query/sql/src/planner/optimizer/filter/deduplicate_join_condition.rs
new file mode 100644
index 000000000000..45ef4d9bb247
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/filter/deduplicate_join_condition.rs
@@ -0,0 +1,134 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::collections::HashMap;
+use std::sync::Arc;
+
+use databend_common_exception::Result;
+
+use crate::optimizer::SExpr;
+use crate::plans::Join;
+use crate::plans::JoinType;
+use crate::plans::RelOperator;
+use crate::ScalarExpr;
+
+// The DeduplicateJoinConditionOptimizer will use the Union-Find algorithm to remove duplicate join conditions,
+// for example: select * from t1, t2, t3 where t1.id = t2.id and t1.id = t3.id and t2.id = t3.id, its join tree is:
+//
+//    Join [t1.id = t3.id, t2.id = t3.id(removed)]
+//    /  \
+//   t3   \
+//       Join: [t1.id = t2.id]
+//       /  \
+//      t1  t2
+//
+// Join condition `t2.id = t3.id` is duplicate, so we can remove it.
+pub struct DeduplicateJoinConditionOptimizer {
+    pub scalar_expr_index: HashMap<ScalarExpr, usize>,
+    pub parent: HashMap<usize, usize>,
+    pub num_scalar_expr: usize,
+}
+
+impl DeduplicateJoinConditionOptimizer {
+    pub fn new() -> Self {
+        DeduplicateJoinConditionOptimizer {
+            scalar_expr_index: HashMap::new(),
+            parent: HashMap::new(),
+            num_scalar_expr: 0,
+        }
+    }
+
+    pub fn run(mut self, s_expr: &SExpr) -> Result<SExpr> {
+        self.deduplicate(s_expr)
+    }
+
+    pub fn deduplicate(&mut self, s_expr: &SExpr) -> Result<SExpr> {
+        match s_expr.plan.as_ref() {
+            RelOperator::Join(join) if join.join_type == JoinType::Inner => {
+                self.deduplicate_join_conditions(s_expr, join)
+            }
+            _ => self.deduplicate_children(s_expr),
+        }
+    }
+
+    fn deduplicate_join_conditions(&mut self, s_expr: &SExpr, join: &Join) -> Result<SExpr> {
+        debug_assert!(join.join_type == JoinType::Inner);
+
+        let left = self.deduplicate(s_expr.child(0)?)?;
+        let right = self.deduplicate(s_expr.child(1)?)?;
+        let mut join = join.clone();
+        let mut new_left_conditions = Vec::new();
+        let mut new_right_conditions = Vec::new();
+        for (left_condition, right_condition) in join
+            .left_conditions
+            .iter()
+            .zip(join.right_conditions.iter())
+        {
+            let left_index = self.get_scalar_expr_index(left_condition);
+            let right_index = self.get_scalar_expr_index(right_condition);
+            let left_parent_index = self.find(left_index);
+            let right_parent_index = self.find(right_index);
+            if left_parent_index != right_parent_index {
+                *self.parent.get_mut(&right_parent_index).unwrap() = left_parent_index;
+                new_left_conditions.push(left_condition.clone());
+                new_right_conditions.push(right_condition.clone());
+            }
+        }
+        if new_left_conditions.len() != join.left_conditions.len() {
+            join.left_conditions = new_left_conditions;
+            join.right_conditions = new_right_conditions;
+        }
+        let s_expr = s_expr.replace_plan(Arc::new(RelOperator::Join(join)));
+        Ok(s_expr.replace_children(vec![Arc::new(left), Arc::new(right)]))
+    }
+
+    pub fn deduplicate_children(&mut self, s_expr: &SExpr) -> Result<SExpr> {
+        let mut children = Vec::with_capacity(s_expr.children().len());
+        for child in s_expr.children() {
+            let child = self.deduplicate(child)?;
+            children.push(Arc::new(child));
+        }
+        Ok(s_expr.replace_children(children))
+    }
+
+    fn get_scalar_expr_index(&mut self, scalar_expr: &ScalarExpr) -> usize {
+        match self.scalar_expr_index.get(scalar_expr) {
+            Some(index) => *index,
+            None => {
+                let index = self.num_scalar_expr;
+                self.scalar_expr_index.insert(scalar_expr.clone(), index);
+                self.num_scalar_expr += 1;
+                index
+            }
+        }
+    }
+
+    fn find(&mut self, index: usize) -> usize {
+        match self.parent.get(&index) {
+            Some(parent_index) => {
+                if index != *parent_index {
+                    let new_parent_index = self.find(*parent_index);
+                    self.parent.insert(index, new_parent_index);
+                    new_parent_index
+                } else {
+                    index
+                }
+            }
+            None => {
+                self.parent.insert(index, index);
+                index
+            }
+        }
+    }
+}
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_infer_filter.rs b/src/query/sql/src/planner/optimizer/filter/infer_filter.rs
similarity index 58%
rename from src/query/sql/src/planner/optimizer/rule/rewrite/rule_infer_filter.rs
rename to src/query/sql/src/planner/optimizer/filter/infer_filter.rs
index de60d5c79e9f..8df451b523b3 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_infer_filter.rs
+++ b/src/query/sql/src/planner/optimizer/filter/infer_filter.rs
@@ -13,7 +13,7 @@
 // limitations under the License.
 
 use std::collections::HashMap;
-use std::sync::Arc;
+use std::collections::HashSet;
 
 use databend_common_exception::Result;
 use databend_common_expression::types::DataType;
@@ -26,144 +26,197 @@ use crate::optimizer::rule::constant::check_float_range;
 use crate::optimizer::rule::constant::check_int_range;
 use crate::optimizer::rule::constant::check_uint_range;
 use crate::optimizer::rule::constant::remove_trivial_type_cast;
-use crate::optimizer::rule::Rule;
-use crate::optimizer::rule::TransformResult;
-use crate::optimizer::RuleID;
-use crate::optimizer::SExpr;
 use crate::plans::ComparisonOp;
 use crate::plans::ConstantExpr;
-use crate::plans::Filter;
 use crate::plans::FunctionCall;
-use crate::plans::PatternPlan;
-use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
+use crate::plans::VisitorMut;
+use crate::ColumnSet;
 
-// The rule tries to infer new predicates from existing predicates, for example:
+// The InferFilterOptimizer tries to infer new predicates from existing predicates, for example:
 // 1. [A > 1 and A > 5] => [A > 5], [A > 1 and A <= 1 => false], [A = 1 and A < 10] => [A = 1]
 // 2. [A = 10 and A = B] => [B = 10]
-// TODO(Dousir9): [A = B and A = C] => [B = C]
-pub struct RuleInferFilter {
-    id: RuleID,
-    patterns: Vec<SExpr>,
+// 3. [A = B and A = C] => [B = C]
+pub struct InferFilterOptimizer<'a> {
+    // All ScalarExprs.
+    exprs: Vec<ScalarExpr>,
+    // The index of ScalarExpr in `exprs`.
+    expr_index: HashMap<ScalarExpr, usize>,
+    // The equal ScalarExprs of each ScalarExpr.
+    expr_equal_to: Vec<Vec<ScalarExpr>>,
+    // The predicates of each ScalarExpr.
+    expr_predicates: Vec<Vec<Predicate>>,
+    // If the whole predicates is false.
+    is_falsy: bool,
+    // The `join_prop` is used for filter push down join.
+    join_prop: Option<JoinProperty<'a>>,
 }
 
-impl RuleInferFilter {
-    pub fn new() -> Self {
+impl<'a> InferFilterOptimizer<'a> {
+    pub fn new(join_prop: Option<JoinProperty<'a>>) -> Self {
         Self {
-            id: RuleID::InferFilter,
-            // Filter
-            //  \
-            //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            exprs: vec![],
+            expr_index: HashMap::new(),
+            expr_equal_to: vec![],
+            expr_predicates: vec![],
+            is_falsy: false,
+            join_prop,
         }
     }
-}
 
-#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Debug)]
-pub struct Predicate {
-    op: ComparisonOp,
-    constant: ConstantExpr,
-}
+    pub fn run(mut self, mut predicates: Vec<ScalarExpr>) -> Result<Vec<ScalarExpr>> {
+        // Remove trivial type cast.
+        for predicate in predicates.iter_mut() {
+            if let ScalarExpr::FunctionCall(func) = predicate {
+                if ComparisonOp::try_from_func_name(&func.func_name).is_some() {
+                    let (left, right) = remove_trivial_type_cast(
+                        func.arguments[0].clone(),
+                        func.arguments[1].clone(),
+                    );
+                    if left != func.arguments[0] {
+                        func.arguments[0] = left;
+                    }
+                    if right != func.arguments[1] {
+                        func.arguments[1] = right;
+                    }
+                }
+            }
+        }
 
-pub struct PredicateSet {
-    exprs: Vec<ScalarExpr>,
-    num_exprs: usize,
-    expr_to_idx: HashMap<ScalarExpr, usize>,
-    equal_exprs: Vec<Vec<ScalarExpr>>,
-    predicates: Vec<Vec<Predicate>>,
-    is_merged: bool,
-    is_falsy: bool,
-}
+        // Process each predicate, add it to the optimizer if it can be used to infer new predicates,
+        // otherwise, add it to the remaining predicates.
+        let mut remaining_predicates = vec![];
+        for predicate in predicates.into_iter() {
+            if let ScalarExpr::FunctionCall(func) = &predicate {
+                if let Some(op) = ComparisonOp::try_from_func_name(&func.func_name) {
+                    match (
+                        func.arguments[0].has_one_column_ref(),
+                        func.arguments[1].has_one_column_ref(),
+                    ) {
+                        (true, true) => {
+                            if op == ComparisonOp::Equal {
+                                self.add_equal_expr(&func.arguments[0], &func.arguments[1]);
+                            } else {
+                                remaining_predicates.push(predicate);
+                            }
+                        }
+                        (true, false)
+                            if let ScalarExpr::ConstantExpr(constant) = &func.arguments[1] =>
+                        {
+                            let (is_adjusted, constant) = adjust_scalar(
+                                constant.value.clone(),
+                                func.arguments[0].data_type()?,
+                            );
+                            if is_adjusted {
+                                self.add_expr_predicate(&func.arguments[0], Predicate {
+                                    op,
+                                    constant,
+                                });
+                            } else {
+                                remaining_predicates.push(predicate);
+                            }
+                        }
+                        (false, true)
+                            if let ScalarExpr::ConstantExpr(constant) = &func.arguments[0] =>
+                        {
+                            let (is_adjusted, constant) = adjust_scalar(
+                                constant.value.clone(),
+                                func.arguments[1].data_type()?,
+                            );
+                            if is_adjusted {
+                                self.add_expr_predicate(&func.arguments[1], Predicate {
+                                    op: op.reverse(),
+                                    constant,
+                                });
+                            } else {
+                                remaining_predicates.push(predicate);
+                            }
+                        }
+                        _ => remaining_predicates.push(predicate),
+                    }
+                } else {
+                    remaining_predicates.push(predicate);
+                }
+            } else {
+                remaining_predicates.push(predicate);
+            }
+        }
 
-enum MergeResult {
-    All,
-    Left,
-    Right,
-    None,
-}
+        let mut new_predicates = vec![];
+        if !self.is_falsy {
+            // Derive new predicates from existing predicates, `derive_predicates` may change is_falsy to true.
+            new_predicates = self.derive_predicates();
+        }
 
-impl PredicateSet {
-    fn new() -> Self {
-        Self {
-            exprs: vec![],
-            num_exprs: 0,
-            expr_to_idx: HashMap::new(),
-            equal_exprs: vec![],
-            predicates: vec![],
-            is_merged: false,
-            is_falsy: false,
+        if self.is_falsy {
+            new_predicates = vec![
+                ConstantExpr {
+                    span: None,
+                    value: Scalar::Boolean(false),
+                }
+                .into(),
+            ];
+        } else {
+            // Derive new predicates from remaining predicates.
+            new_predicates.extend(self.derive_remaining_predicates(remaining_predicates));
         }
+
+        Ok(new_predicates)
     }
 
     fn add_expr(
         &mut self,
         expr: &ScalarExpr,
-        predicates: Vec<Predicate>,
-        equal_exprs: Vec<ScalarExpr>,
+        expr_predicates: Vec<Predicate>,
+        expr_equal_to: Vec<ScalarExpr>,
     ) {
+        self.expr_index.insert(expr.clone(), self.exprs.len());
         self.exprs.push(expr.clone());
-        self.expr_to_idx.insert(expr.clone(), self.num_exprs);
-        self.predicates.push(predicates);
-        self.equal_exprs.push(equal_exprs);
-        self.num_exprs += 1;
+        self.expr_predicates.push(expr_predicates);
+        self.expr_equal_to.push(expr_equal_to);
     }
 
-    fn add_equal(&mut self, left: &ScalarExpr, right: &ScalarExpr) {
-        match self.expr_to_idx.get(left) {
-            Some(idx) => {
-                let equal_exprs = &mut self.equal_exprs[*idx];
-                equal_exprs.push(right.clone());
-            }
+    pub fn add_equal_expr(&mut self, left: &ScalarExpr, right: &ScalarExpr) {
+        match self.expr_index.get(left) {
+            Some(index) => self.expr_equal_to[*index].push(right.clone()),
             None => self.add_expr(left, vec![], vec![right.clone()]),
         };
-        if self.expr_to_idx.get(right).is_none() {
-            self.add_expr(right, vec![], vec![]);
-        }
+
+        match self.expr_index.get(right) {
+            Some(index) => self.expr_equal_to[*index].push(left.clone()),
+            None => self.add_expr(right, vec![], vec![left.clone()]),
+        };
     }
 
-    fn add_predicate(&mut self, left: &ScalarExpr, right: Predicate) {
-        match self.expr_to_idx.get(left) {
-            Some(idx) => {
-                let predicates = &mut self.predicates[*idx];
+    fn add_expr_predicate(&mut self, expr: &ScalarExpr, new_predicate: Predicate) {
+        match self.expr_index.get(expr) {
+            Some(index) => {
+                let predicates = &mut self.expr_predicates[*index];
                 for predicate in predicates.iter_mut() {
-                    match Self::merge(predicate, &right) {
+                    match Self::merge_predicate(predicate, &new_predicate) {
                         MergeResult::None => {
                             self.is_falsy = true;
-                            self.is_merged = true;
                             return;
                         }
                         MergeResult::Left => {
-                            self.is_merged = true;
                             return;
                         }
                         MergeResult::Right => {
-                            *predicate = right;
-                            self.is_merged = true;
+                            *predicate = new_predicate;
                             return;
                         }
                         MergeResult::All => (),
                     }
                 }
-                predicates.push(right);
+                predicates.push(new_predicate);
+            }
+            None => {
+                self.add_expr(expr, vec![new_predicate], vec![]);
             }
-            None => self.add_expr(left, vec![right], vec![]),
         };
     }
 
-    fn merge(left: &Predicate, right: &Predicate) -> MergeResult {
+    fn merge_predicate(left: &Predicate, right: &Predicate) -> MergeResult {
         match left.op {
             ComparisonOp::Equal => match right.op {
                 ComparisonOp::Equal => match left.constant == right.constant {
@@ -323,60 +376,216 @@ impl PredicateSet {
         }
     }
 
-    fn derive_predicates(&mut self) -> (bool, Vec<ScalarExpr>) {
-        let mut is_updated = self.is_merged;
+    fn derive_predicates(&mut self) -> Vec<ScalarExpr> {
         let mut result = vec![];
-        let num_exprs = self.num_exprs;
+        let num_exprs = self.exprs.len();
+
+        // Using the Union-Find algorithm to construct the equal ScalarExpr index sets.
+        let mut equal_index_sets: HashMap<usize, HashSet<usize>> = HashMap::new();
         let mut parents = vec![0; num_exprs];
         for (i, parent) in parents.iter_mut().enumerate().take(num_exprs) {
             *parent = i;
         }
-        for (left_idx, equal_exprs) in self.equal_exprs.iter().enumerate() {
-            for expr in equal_exprs.iter() {
-                let right_idx = self.expr_to_idx.get(expr).unwrap();
-                Self::union(&mut parents, left_idx, *right_idx);
+        for (left_index, expr_equal_to) in self.expr_equal_to.iter().enumerate() {
+            for expr in expr_equal_to.iter() {
+                let right_index = self.expr_index.get(expr).unwrap();
+                Self::union(&mut parents, left_index, *right_index);
             }
         }
-        let mut old_predicates_set = self.predicates.clone();
-        for predicates in old_predicates_set.iter_mut() {
-            predicates.sort();
-        }
-        for idx in 0..num_exprs {
-            let parent_idx = Self::find(&mut parents, idx);
-            if idx != parent_idx {
-                let expr = self.exprs[parent_idx].clone();
-                let predicates = self.predicates[idx].clone();
+        for index in 0..num_exprs {
+            let parent_index = Self::find(&mut parents, index);
+            match equal_index_sets.get_mut(&parent_index) {
+                Some(equal_index_set) => {
+                    equal_index_set.insert(index);
+                }
+                None => {
+                    equal_index_sets.insert(parent_index, HashSet::from([index]));
+                }
+            }
+            if index != parent_index {
+                // Add the predicates to the parent ScalarExpr.
+                let expr = self.exprs[parent_index].clone();
+                let predicates = self.expr_predicates[index].clone();
                 for predicate in predicates {
-                    self.add_predicate(&expr, predicate);
+                    self.add_expr_predicate(&expr, predicate);
                 }
             }
         }
-        for predicates in self.predicates.iter_mut() {
-            predicates.sort();
-        }
-        for (scalar, idx) in self.expr_to_idx.iter() {
-            let parent_idx = Self::find(&mut parents, *idx);
-            let old_predicates = &old_predicates_set[*idx];
-            let parent_predicates = &self.predicates[parent_idx];
-            if old_predicates.len() != parent_predicates.len() {
-                is_updated = true;
-            }
-            for (i, predicate) in parent_predicates.iter().enumerate() {
-                if i < old_predicates.len() && &old_predicates[i] != predicate {
-                    is_updated = true;
-                }
+
+        // Construct predicates for each ScalarExpr.
+        for expr in self.exprs.iter() {
+            let index = self.expr_index.get(expr).unwrap();
+            let parent_index = Self::find(&mut parents, *index);
+            let parent_predicates = &self.expr_predicates[parent_index];
+            for predicate in parent_predicates.iter() {
                 result.push(ScalarExpr::FunctionCall(FunctionCall {
                     span: None,
                     func_name: String::from(predicate.op.to_func_name()),
                     params: vec![],
                     arguments: vec![
-                        scalar.clone(),
+                        expr.clone(),
                         ScalarExpr::ConstantExpr(predicate.constant.clone()),
                     ],
                 }));
             }
         }
-        (is_updated | self.is_falsy, result)
+
+        // Construct equal condition predicates for each equal ScalarExpr index set.
+        for index in 0..num_exprs {
+            let parent_index = Self::find(&mut parents, index);
+            if index == parent_index {
+                if let Some(equal_index_set) = equal_index_sets.get(&parent_index) {
+                    let mut equal_indexes = equal_index_set.iter().copied().collect::<Vec<_>>();
+                    equal_indexes.sort();
+                    let equal_indexes_len = equal_indexes.len();
+                    for i in 0..equal_indexes_len {
+                        for j in i + 1..equal_indexes_len {
+                            result.push(ScalarExpr::FunctionCall(FunctionCall {
+                                span: None,
+                                func_name: String::from(ComparisonOp::Equal.to_func_name()),
+                                params: vec![],
+                                arguments: vec![
+                                    self.exprs[equal_indexes[i]].clone(),
+                                    self.exprs[equal_indexes[j]].clone(),
+                                ],
+                            }));
+                        }
+                    }
+                }
+            }
+        }
+
+        result
+    }
+
+    fn derive_remaining_predicates(&self, predicates: Vec<ScalarExpr>) -> Vec<ScalarExpr> {
+        // The ReplaceScalarExpr is used to replace the ScalarExpr of a predicate.
+        struct ReplaceScalarExpr<'a> {
+            // The index of ScalarExpr in `exprs`.
+            expr_index: &'a HashMap<ScalarExpr, usize>,
+            // The equal ScalarExprs of each ScalarExpr.
+            expr_equal_to: &'a Vec<Vec<ScalarExpr>>,
+            // The columns used by the predicate.
+            column_set: HashSet<usize>,
+            // If the predicate can be replaced to generate a new predicate.
+            can_replace: bool,
+        }
+
+        impl<'a> ReplaceScalarExpr<'a> {
+            fn reset(&mut self) {
+                self.column_set.clear();
+                self.can_replace = true;
+            }
+        }
+
+        impl<'a> VisitorMut<'_> for ReplaceScalarExpr<'a> {
+            fn visit(&mut self, expr: &mut ScalarExpr) -> Result<()> {
+                if let Some(index) = self.expr_index.get(expr) {
+                    let equal_to = &self.expr_equal_to[*index];
+                    if !equal_to.is_empty() {
+                        let used_columns = expr.used_columns();
+                        for column in used_columns {
+                            self.column_set.insert(column);
+                        }
+                        *expr = equal_to[0].clone();
+                        return Ok(());
+                    }
+                }
+                match expr {
+                    ScalarExpr::FunctionCall(expr) => self.visit_function_call(expr),
+                    ScalarExpr::CastExpr(expr) => self.visit_cast_expr(expr),
+                    ScalarExpr::ConstantExpr(_) => Ok(()),
+                    ScalarExpr::BoundColumnRef(_)
+                    | ScalarExpr::WindowFunction(_)
+                    | ScalarExpr::AggregateFunction(_)
+                    | ScalarExpr::LambdaFunction(_)
+                    | ScalarExpr::SubqueryExpr(_)
+                    | ScalarExpr::UDFServerCall(_)
+                    | ScalarExpr::UDFLambdaCall(_) => {
+                        // Can not replace `BoundColumnRef` or can not replace unsupported ScalarExpr.
+                        self.can_replace = false;
+                        Ok(())
+                    }
+                }
+            }
+        }
+
+        let mut replace = ReplaceScalarExpr {
+            expr_index: &self.expr_index,
+            expr_equal_to: &self.expr_equal_to,
+            column_set: HashSet::new(),
+            can_replace: true,
+        };
+
+        let mut result_predicates = Vec::with_capacity(predicates.len());
+        for predicate in predicates {
+            replace.reset();
+            let mut new_predicate = predicate.clone();
+            replace.visit(&mut new_predicate).unwrap();
+            if !replace.can_replace {
+                result_predicates.push(predicate);
+                continue;
+            }
+
+            let mut can_replace = false;
+            if let Some(join_prop) = &self.join_prop {
+                let mut has_left = false;
+                let mut has_right = false;
+                for column in replace.column_set.iter() {
+                    if join_prop.left_columns.contains(column) {
+                        has_left = true;
+                    } else if join_prop.right_columns.contains(column) {
+                        has_right = true;
+                    }
+                }
+                // We only derive new predicates when the predicate contains columns only from one side of the join.
+                if has_left && !has_right || !has_left && has_right {
+                    can_replace = true;
+                }
+            } else if replace.column_set.len() == 1 {
+                can_replace = true;
+            }
+
+            if !can_replace {
+                result_predicates.push(predicate);
+                continue;
+            }
+
+            if new_predicate != predicate {
+                result_predicates.push(new_predicate);
+            }
+
+            result_predicates.push(predicate);
+        }
+
+        result_predicates
+    }
+}
+
+#[derive(Eq, PartialEq, Ord, PartialOrd, Clone, Debug)]
+struct Predicate {
+    op: ComparisonOp,
+    constant: ConstantExpr,
+}
+
+enum MergeResult {
+    All,
+    Left,
+    Right,
+    None,
+}
+
+pub struct JoinProperty<'a> {
+    left_columns: &'a ColumnSet,
+    right_columns: &'a ColumnSet,
+}
+
+impl<'a> JoinProperty<'a> {
+    pub fn new(left_columns: &'a ColumnSet, right_columns: &'a ColumnSet) -> Self {
+        Self {
+            left_columns,
+            right_columns,
+        }
     }
 }
 
@@ -489,127 +698,3 @@ pub fn adjust_scalar(scalar: Scalar, data_type: DataType) -> (bool, ConstantExpr
         value: scalar,
     })
 }
-
-impl Rule for RuleInferFilter {
-    fn id(&self) -> RuleID {
-        self.id
-    }
-
-    fn apply(&self, s_expr: &SExpr, state: &mut TransformResult) -> Result<()> {
-        let filter: Filter = s_expr.plan().clone().try_into()?;
-        let mut predicates = filter.predicates;
-        let mut new_predicates = vec![];
-        let mut is_rewritten = false;
-        let mut predicate_set = PredicateSet::new();
-        for predicate in predicates.iter_mut() {
-            if let ScalarExpr::FunctionCall(func) = predicate {
-                if ComparisonOp::try_from_func_name(&func.func_name).is_some() {
-                    let (left, right) = remove_trivial_type_cast(
-                        func.arguments[0].clone(),
-                        func.arguments[1].clone(),
-                    );
-                    if left != func.arguments[0] {
-                        is_rewritten = true;
-                        func.arguments[0] = left;
-                    }
-                    if right != func.arguments[1] {
-                        is_rewritten = true;
-                        func.arguments[1] = right;
-                    }
-                }
-            }
-        }
-        for predicate in predicates.into_iter() {
-            if let ScalarExpr::FunctionCall(func) = &predicate {
-                if let Some(op) = ComparisonOp::try_from_func_name(&func.func_name) {
-                    match (
-                        func.arguments[0].is_column_ref(),
-                        func.arguments[1].is_column_ref(),
-                    ) {
-                        (true, true) => {
-                            if op == ComparisonOp::Equal {
-                                predicate_set.add_equal(&func.arguments[0], &func.arguments[1]);
-                            }
-                            new_predicates.push(predicate);
-                        }
-                        (true, false) => {
-                            if let ScalarExpr::ConstantExpr(constant) = &func.arguments[1] {
-                                let (is_adjusted, constant) = adjust_scalar(
-                                    constant.value.clone(),
-                                    func.arguments[0].data_type()?,
-                                );
-                                if is_adjusted {
-                                    predicate_set.add_predicate(&func.arguments[0], Predicate {
-                                        op,
-                                        constant,
-                                    });
-                                } else {
-                                    new_predicates.push(predicate);
-                                }
-                            } else {
-                                new_predicates.push(predicate);
-                            }
-                        }
-                        (false, true) => {
-                            if let ScalarExpr::ConstantExpr(constant) = &func.arguments[0] {
-                                let (is_adjusted, constant) = adjust_scalar(
-                                    constant.value.clone(),
-                                    func.arguments[1].data_type()?,
-                                );
-                                if is_adjusted {
-                                    predicate_set.add_predicate(&func.arguments[1], Predicate {
-                                        op: op.reverse(),
-                                        constant,
-                                    });
-                                } else {
-                                    new_predicates.push(predicate);
-                                }
-                            } else {
-                                new_predicates.push(predicate);
-                            }
-                        }
-                        (false, false) => {
-                            new_predicates.push(predicate);
-                        }
-                    }
-                } else {
-                    new_predicates.push(predicate);
-                }
-            } else {
-                new_predicates.push(predicate);
-            }
-        }
-        is_rewritten |= predicate_set.is_merged;
-        if !predicate_set.is_falsy {
-            // `derive_predicates` may change is_falsy to true.
-            let (is_merged, infer_predicates) = predicate_set.derive_predicates();
-            is_rewritten |= is_merged;
-            new_predicates.extend(infer_predicates);
-        }
-        if predicate_set.is_falsy {
-            new_predicates = vec![
-                ConstantExpr {
-                    span: None,
-                    value: Scalar::Boolean(false),
-                }
-                .into(),
-            ];
-        }
-        if is_rewritten {
-            state.add_result(SExpr::create_unary(
-                Arc::new(
-                    Filter {
-                        predicates: new_predicates,
-                    }
-                    .into(),
-                ),
-                Arc::new(s_expr.child(0)?.clone()),
-            ));
-        }
-        Ok(())
-    }
-
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
-    }
-}
diff --git a/src/query/sql/src/planner/optimizer/filter/mod.rs b/src/query/sql/src/planner/optimizer/filter/mod.rs
new file mode 100644
index 000000000000..012cd32ead4f
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/filter/mod.rs
@@ -0,0 +1,24 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+mod deduplicate_join_condition;
+mod infer_filter;
+mod normalize_disjunctive_filter;
+mod pull_up_filter;
+
+pub use deduplicate_join_condition::DeduplicateJoinConditionOptimizer;
+pub use infer_filter::InferFilterOptimizer;
+pub use infer_filter::JoinProperty;
+pub use normalize_disjunctive_filter::NormalizeDisjunctiveFilterOptimizer;
+pub use pull_up_filter::PullUpFilterOptimizer;
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_disjunctive_filter.rs b/src/query/sql/src/planner/optimizer/filter/normalize_disjunctive_filter.rs
similarity index 66%
rename from src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_disjunctive_filter.rs
rename to src/query/sql/src/planner/optimizer/filter/normalize_disjunctive_filter.rs
index cddcf2c0a43c..062740d7d20d 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_disjunctive_filter.rs
+++ b/src/query/sql/src/planner/optimizer/filter/normalize_disjunctive_filter.rs
@@ -12,24 +12,43 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::Arc;
-
 use databend_common_exception::Result;
 use databend_common_expression::Scalar;
 use itertools::Itertools;
 
 use crate::binder::split_conjunctions;
-use crate::optimizer::rule::Rule;
-use crate::optimizer::rule::TransformResult;
-use crate::optimizer::RuleID;
-use crate::optimizer::SExpr;
 use crate::plans::ConstantExpr;
-use crate::plans::Filter;
 use crate::plans::FunctionCall;
-use crate::plans::PatternPlan;
-use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 
+// The NormalizeDisjunctiveFilterOptimizer tries to apply the inverse OR distributive law to the predicate.
+// (A AND B) OR (A AND C) => A AND (B OR C)
+// It'll find all OR expressions and extract the common terms.
+#[derive(Default)]
+pub struct NormalizeDisjunctiveFilterOptimizer {}
+
+impl NormalizeDisjunctiveFilterOptimizer {
+    pub fn new() -> Self {
+        NormalizeDisjunctiveFilterOptimizer::default()
+    }
+}
+
+impl NormalizeDisjunctiveFilterOptimizer {
+    pub fn run(self, predicates: Vec<ScalarExpr>) -> Result<Vec<ScalarExpr>> {
+        let mut rewritten_predicates = Vec::with_capacity(predicates.len());
+        for predicate in predicates.iter() {
+            let predicate_scalar = predicate_scalar(predicate);
+            let rewritten_predicate_scalar = rewrite_predicate_ors(predicate_scalar);
+            rewritten_predicates.push(normalize_predicate_scalar(rewritten_predicate_scalar));
+        }
+        let mut split_predicates: Vec<ScalarExpr> = Vec::with_capacity(rewritten_predicates.len());
+        for predicate in rewritten_predicates.iter() {
+            split_predicates.extend_from_slice(&split_conjunctions(predicate));
+        }
+        Ok(split_predicates)
+    }
+}
+
 #[derive(Clone, PartialEq, Eq, Hash, Debug)]
 enum PredicateScalar {
     And(Vec<PredicateScalar>),
@@ -37,51 +56,43 @@ enum PredicateScalar {
     Other(Box<ScalarExpr>),
 }
 
-fn predicate_scalar(scalar: &ScalarExpr) -> (bool, PredicateScalar) {
+fn predicate_scalar(scalar: &ScalarExpr) -> PredicateScalar {
     match scalar {
         ScalarExpr::FunctionCall(func) if func.func_name == "and" => {
             let mut and_args = vec![];
-            let mut is_rewritten = false;
             for argument in func.arguments.iter() {
                 // Recursively flatten the AND expressions.
-                let (rewritten, predicate) = predicate_scalar(argument);
-                is_rewritten |= rewritten;
+                let predicate = predicate_scalar(argument);
                 if let PredicateScalar::And(args) = predicate {
                     and_args.extend(args);
                 } else {
                     and_args.push(predicate);
                 }
             }
-            let original_len = and_args.len();
             and_args = and_args
                 .into_iter()
                 .unique()
                 .collect::<Vec<PredicateScalar>>();
-            is_rewritten |= original_len != and_args.len();
-            (is_rewritten, PredicateScalar::And(and_args))
+            PredicateScalar::And(and_args)
         }
         ScalarExpr::FunctionCall(func) if func.func_name == "or" => {
             let mut or_args = vec![];
-            let mut is_rewritten = false;
             for argument in func.arguments.iter() {
                 // Recursively flatten the OR expressions.
-                let (rewritten, predicate) = predicate_scalar(argument);
-                is_rewritten |= rewritten;
+                let predicate = predicate_scalar(argument);
                 if let PredicateScalar::Or(args) = predicate {
                     or_args.extend(args);
                 } else {
                     or_args.push(predicate);
                 }
             }
-            let original_len = or_args.len();
             or_args = or_args
                 .into_iter()
                 .unique()
                 .collect::<Vec<PredicateScalar>>();
-            is_rewritten |= original_len != or_args.len();
-            (is_rewritten, PredicateScalar::Or(or_args))
+            PredicateScalar::Or(or_args)
         }
-        _ => (false, PredicateScalar::Other(Box::from(scalar.clone()))),
+        _ => PredicateScalar::Other(Box::from(scalar.clone())),
     }
 }
 
@@ -119,112 +130,36 @@ fn normalize_predicate_scalar(predicate_scalar: PredicateScalar) -> ScalarExpr {
     }
 }
 
-// The rule tries to apply the inverse OR distributive law to the predicate.
-// (A AND B) OR (A AND C) => A AND (B OR C)
-// It'll find all OR expressions and extract the common terms.
-pub struct RuleNormalizeDisjunctiveFilter {
-    id: RuleID,
-    patterns: Vec<SExpr>,
-}
-
-impl RuleNormalizeDisjunctiveFilter {
-    pub fn new() -> Self {
-        Self {
-            id: RuleID::NormalizeDisjunctiveFilter,
-            // Filter
-            //  \
-            //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
-        }
-    }
-}
-
-impl Rule for RuleNormalizeDisjunctiveFilter {
-    fn id(&self) -> RuleID {
-        self.id
-    }
-
-    fn apply(&self, s_expr: &SExpr, state: &mut TransformResult) -> Result<()> {
-        let filter: Filter = s_expr.plan().clone().try_into()?;
-        let predicates = filter.predicates;
-        let mut rewritten_predicates = Vec::with_capacity(predicates.len());
-        let mut is_rewritten = false;
-        for predicate in predicates.iter() {
-            let (rewritten, predicate_scalar) = predicate_scalar(predicate);
-            is_rewritten |= rewritten;
-            let (rewritten_predicate_scalar, rewritten) = rewrite_predicate_ors(predicate_scalar);
-            is_rewritten |= rewritten;
-            rewritten_predicates.push(normalize_predicate_scalar(rewritten_predicate_scalar));
-        }
-        let mut split_predicates: Vec<ScalarExpr> = Vec::with_capacity(rewritten_predicates.len());
-        for predicate in rewritten_predicates.iter() {
-            split_predicates.extend_from_slice(&split_conjunctions(predicate));
-        }
-        if is_rewritten {
-            state.add_result(SExpr::create_unary(
-                Arc::new(
-                    Filter {
-                        predicates: split_predicates,
-                    }
-                    .into(),
-                ),
-                Arc::new(s_expr.child(0)?.clone()),
-            ));
-        }
-        Ok(())
-    }
-
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
-    }
-}
-
-fn rewrite_predicate_ors(predicate: PredicateScalar) -> (PredicateScalar, bool) {
+fn rewrite_predicate_ors(predicate: PredicateScalar) -> PredicateScalar {
     match predicate {
         PredicateScalar::Or(args) => {
             let mut or_args = Vec::with_capacity(args.len());
             for arg in args.iter() {
-                or_args.push(rewrite_predicate_ors(arg.clone()).0);
+                or_args.push(rewrite_predicate_ors(arg.clone()));
             }
             process_duplicate_or_exprs(or_args)
         }
         PredicateScalar::And(args) => {
             let mut and_args = Vec::with_capacity(args.len());
             for arg in args.iter() {
-                and_args.push(rewrite_predicate_ors(arg.clone()).0);
+                and_args.push(rewrite_predicate_ors(arg.clone()));
             }
-            (PredicateScalar::And(and_args), false)
+            PredicateScalar::And(and_args)
         }
-        PredicateScalar::Other(_) => (predicate, false),
+        PredicateScalar::Other(_) => predicate,
     }
 }
 
 // Apply the inverse OR distributive law.
-fn process_duplicate_or_exprs(mut or_args: Vec<PredicateScalar>) -> (PredicateScalar, bool) {
+fn process_duplicate_or_exprs(mut or_args: Vec<PredicateScalar>) -> PredicateScalar {
     if or_args.is_empty() {
-        return (
-            PredicateScalar::Other(Box::from(ScalarExpr::ConstantExpr(ConstantExpr {
-                span: None,
-                value: Scalar::Boolean(false),
-            }))),
-            false,
-        );
+        return PredicateScalar::Other(Box::from(ScalarExpr::ConstantExpr(ConstantExpr {
+            span: None,
+            value: Scalar::Boolean(false),
+        })));
     }
     if or_args.len() == 1 {
-        return (or_args[0].clone(), false);
+        return or_args[0].clone();
     }
     // choose the shortest AND expression
     let mut shortest_exprs: Vec<PredicateScalar> = vec![];
@@ -259,7 +194,7 @@ fn process_duplicate_or_exprs(mut or_args: Vec<PredicateScalar>) -> (PredicateSc
     }
 
     if exist_exprs.is_empty() {
-        return (PredicateScalar::Or(or_args), false);
+        return PredicateScalar::Or(or_args);
     }
 
     // Rebuild the OR predicate.
@@ -297,8 +232,8 @@ fn process_duplicate_or_exprs(mut or_args: Vec<PredicateScalar>) -> (PredicateSc
     }
 
     if exist_exprs.len() == 1 {
-        (exist_exprs[0].clone(), true)
+        exist_exprs[0].clone()
     } else {
-        (PredicateScalar::And(exist_exprs), true)
+        PredicateScalar::And(exist_exprs)
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/filter/pull_up_filter.rs b/src/query/sql/src/planner/optimizer/filter/pull_up_filter.rs
new file mode 100644
index 000000000000..9b07d034a3f0
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/filter/pull_up_filter.rs
@@ -0,0 +1,248 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+
+use databend_common_exception::Result;
+
+use crate::binder::split_conjunctions;
+use crate::optimizer::filter::InferFilterOptimizer;
+use crate::optimizer::filter::NormalizeDisjunctiveFilterOptimizer;
+use crate::optimizer::SExpr;
+use crate::plans::EvalScalar;
+use crate::plans::Filter;
+use crate::plans::FunctionCall;
+use crate::plans::Join;
+use crate::plans::JoinType;
+use crate::plans::RelOperator;
+use crate::plans::ScalarItem;
+use crate::plans::WindowFuncType;
+use crate::MetadataRef;
+use crate::ScalarExpr;
+
+// The PullUpFilterOptimizer will pull up filters to the top of the plan tree and infer new filters.
+pub struct PullUpFilterOptimizer {
+    pub predicates: Vec<ScalarExpr>,
+    pub metadata: MetadataRef,
+}
+
+impl PullUpFilterOptimizer {
+    pub fn new(metadata: MetadataRef) -> Self {
+        PullUpFilterOptimizer {
+            predicates: vec![],
+            metadata,
+        }
+    }
+
+    pub fn run(mut self, s_expr: &SExpr) -> Result<SExpr> {
+        let mut s_expr = self.pull_up(s_expr)?;
+        s_expr = self.finish(s_expr)?;
+        Ok(s_expr)
+    }
+
+    pub fn finish(self, s_expr: SExpr) -> Result<SExpr> {
+        if self.predicates.is_empty() {
+            Ok(s_expr)
+        } else {
+            let predicates = InferFilterOptimizer::new(None).run(self.predicates)?;
+            let predicates = NormalizeDisjunctiveFilterOptimizer::new().run(predicates)?;
+            let filter = Filter { predicates };
+            Ok(SExpr::create_unary(
+                Arc::new(filter.into()),
+                Arc::new(s_expr),
+            ))
+        }
+    }
+
+    pub fn pull_up(&mut self, s_expr: &SExpr) -> Result<SExpr> {
+        match s_expr.plan.as_ref() {
+            RelOperator::Filter(filter) => self.pull_up_filter(s_expr, filter),
+            RelOperator::Join(join) if !join.is_lateral => self.pull_up_join(s_expr, join),
+            RelOperator::EvalScalar(eval_scalar) => self.pull_up_eval_scalar(s_expr, eval_scalar),
+            RelOperator::MaterializedCte(_) => Ok(s_expr.clone()),
+            _ => self.pull_up_others(s_expr),
+        }
+    }
+
+    fn pull_up_filter(&mut self, s_expr: &SExpr, filter: &Filter) -> Result<SExpr> {
+        let child = self.pull_up(s_expr.child(0)?)?;
+        for predicate in filter.predicates.iter() {
+            self.predicates.extend(split_conjunctions(predicate));
+        }
+        Ok(child)
+    }
+
+    fn pull_up_join(&mut self, s_expr: &SExpr, join: &Join) -> Result<SExpr> {
+        let (left_need_pull_up, right_need_pull_up) = match join.join_type {
+            JoinType::Inner | JoinType::Cross => (true, true),
+            JoinType::Left | JoinType::LeftSingle | JoinType::LeftSemi | JoinType::LeftAnti => {
+                (true, false)
+            }
+            JoinType::Right | JoinType::RightSingle | JoinType::RightSemi | JoinType::RightAnti => {
+                (false, true)
+            }
+            _ => (false, false),
+        };
+        let mut left_pull_up = PullUpFilterOptimizer::new(self.metadata.clone());
+        let mut right_pull_up = PullUpFilterOptimizer::new(self.metadata.clone());
+        let mut left = left_pull_up.pull_up(s_expr.child(0)?)?;
+        let mut right = right_pull_up.pull_up(s_expr.child(1)?)?;
+        if left_need_pull_up {
+            for predicate in left_pull_up.predicates {
+                self.predicates.extend(split_conjunctions(&predicate));
+            }
+        } else {
+            left = left_pull_up.finish(left)?;
+        }
+        if right_need_pull_up {
+            for predicate in right_pull_up.predicates {
+                self.predicates.extend(split_conjunctions(&predicate));
+            }
+        } else {
+            right = right_pull_up.finish(right)?;
+        }
+        let mut join = join.clone();
+        if left_need_pull_up && right_need_pull_up {
+            for (left_condition, right_condition) in join
+                .left_conditions
+                .iter()
+                .zip(join.right_conditions.iter())
+            {
+                let predicate = ScalarExpr::FunctionCall(FunctionCall {
+                    span: None,
+                    func_name: "eq".to_string(),
+                    params: vec![],
+                    arguments: vec![left_condition.clone(), right_condition.clone()],
+                });
+                self.predicates.push(predicate);
+            }
+            for predicate in join.non_equi_conditions.iter() {
+                self.predicates.extend(split_conjunctions(predicate));
+            }
+            join.left_conditions.clear();
+            join.right_conditions.clear();
+            join.non_equi_conditions.clear();
+            join.join_type = JoinType::Cross;
+        }
+        let s_expr = s_expr.replace_plan(Arc::new(RelOperator::Join(join)));
+        Ok(s_expr.replace_children(vec![Arc::new(left), Arc::new(right)]))
+    }
+
+    fn pull_up_eval_scalar(&mut self, s_expr: &SExpr, eval_scalar: &EvalScalar) -> Result<SExpr> {
+        let child = self.pull_up(s_expr.child(0)?)?;
+        let mut eval_scalar = eval_scalar.clone();
+        for predicate in self.predicates.iter_mut() {
+            Self::replace_predicate(predicate, &mut eval_scalar.items, &self.metadata)?;
+        }
+        let s_expr = s_expr.replace_plan(Arc::new(RelOperator::EvalScalar(eval_scalar)));
+        Ok(s_expr.replace_children(vec![Arc::new(child)]))
+    }
+
+    pub fn pull_up_others(&mut self, s_expr: &SExpr) -> Result<SExpr> {
+        let mut children = Vec::with_capacity(s_expr.children().len());
+        for child in s_expr.children() {
+            let child = PullUpFilterOptimizer::new(self.metadata.clone()).run(child)?;
+            children.push(Arc::new(child));
+        }
+        Ok(s_expr.replace_children(children))
+    }
+
+    fn replace_predicate(
+        predicate: &mut ScalarExpr,
+        items: &mut Vec<ScalarItem>,
+        metadata: &MetadataRef,
+    ) -> Result<()> {
+        match predicate {
+            ScalarExpr::BoundColumnRef(column) => {
+                for item in items.iter() {
+                    if item.index == column.column.index {
+                        return Ok(());
+                    }
+                    if let ScalarExpr::BoundColumnRef(item_column) = &item.scalar {
+                        if item_column.column.index == column.column.index {
+                            column.column.index = item.index;
+                            return Ok(());
+                        }
+                    }
+                }
+
+                let new_index = metadata.write().add_derived_column(
+                    column.column.column_name.clone(),
+                    *column.column.data_type.clone(),
+                );
+                let new_column = column.clone();
+                items.push(ScalarItem {
+                    scalar: ScalarExpr::BoundColumnRef(new_column),
+                    index: new_index,
+                });
+                column.column.index = new_index;
+            }
+            ScalarExpr::WindowFunction(window) => {
+                match &mut window.func {
+                    WindowFuncType::Aggregate(agg) => {
+                        for arg in agg.args.iter_mut() {
+                            Self::replace_predicate(arg, items, metadata)?;
+                        }
+                    }
+                    WindowFuncType::LagLead(ll) => {
+                        Self::replace_predicate(&mut ll.arg, items, metadata)?;
+                        if let Some(default) = ll.default.as_mut() {
+                            Self::replace_predicate(default, items, metadata)?;
+                        }
+                    }
+                    WindowFuncType::NthValue(func) => {
+                        Self::replace_predicate(&mut func.arg, items, metadata)?;
+                    }
+                    _ => (),
+                };
+
+                for window_partition_by in window.partition_by.iter_mut() {
+                    Self::replace_predicate(window_partition_by, items, metadata)?;
+                }
+
+                for window_order_by in window.order_by.iter_mut() {
+                    Self::replace_predicate(&mut window_order_by.expr, items, metadata)?;
+                }
+            }
+            ScalarExpr::AggregateFunction(agg_func) => {
+                for arg in agg_func.args.iter_mut() {
+                    Self::replace_predicate(arg, items, metadata)?;
+                }
+            }
+            ScalarExpr::FunctionCall(func) => {
+                for arg in func.arguments.iter_mut() {
+                    Self::replace_predicate(arg, items, metadata)?;
+                }
+            }
+            ScalarExpr::LambdaFunction(lambda_func) => {
+                for arg in lambda_func.args.iter_mut() {
+                    Self::replace_predicate(arg, items, metadata)?;
+                }
+            }
+            ScalarExpr::CastExpr(cast) => {
+                Self::replace_predicate(&mut cast.argument, items, metadata)?;
+            }
+            ScalarExpr::UDFServerCall(udf) => {
+                for arg in udf.arguments.iter_mut() {
+                    Self::replace_predicate(arg, items, metadata)?;
+                }
+            }
+            ScalarExpr::UDFLambdaCall(udf) => {
+                Self::replace_predicate(&mut udf.scalar, items, metadata)?;
+            }
+            _ => (),
+        }
+        Ok(())
+    }
+}
diff --git a/src/query/sql/src/planner/optimizer/format.rs b/src/query/sql/src/planner/optimizer/format.rs
index 662a8d61b50b..5a3c0bb71766 100644
--- a/src/query/sql/src/planner/optimizer/format.rs
+++ b/src/query/sql/src/planner/optimizer/format.rs
@@ -60,7 +60,6 @@ pub fn display_rel_op(rel_op: &RelOperator) -> String {
                 Exchange::MergeSort => "MergeSort".to_string(),
             })
         }
-        RelOperator::Pattern(_) => "Pattern".to_string(),
         RelOperator::DummyTableScan(_) => "DummyTableScan".to_string(),
         RelOperator::ProjectSet(_) => "ProjectSet".to_string(),
         RelOperator::Window(_) => "WindowFunc".to_string(),
diff --git a/src/query/sql/src/planner/optimizer/hyper_dp/dphyp.rs b/src/query/sql/src/planner/optimizer/hyper_dp/dphyp.rs
index dd84bc13cb42..599f6e335ed8 100644
--- a/src/query/sql/src/planner/optimizer/hyper_dp/dphyp.rs
+++ b/src/query/sql/src/planner/optimizer/hyper_dp/dphyp.rs
@@ -178,6 +178,12 @@ impl DPhpy {
                 }
                 // Add join conditions
                 for condition_pair in op.left_conditions.iter().zip(op.right_conditions.iter()) {
+                    let left_used_tables = condition_pair.0.used_tables()?;
+                    let right_used_tables = condition_pair.1.used_tables()?;
+                    if left_used_tables.is_empty() || right_used_tables.is_empty() {
+                        is_inner_join = false;
+                        break;
+                    }
                     join_conditions.push((condition_pair.0.clone(), condition_pair.1.clone()));
                 }
                 if !op.non_equi_conditions.is_empty() {
@@ -249,7 +255,7 @@ impl DPhpy {
                 self.join_relations.push(JoinRelation::new(&new_s_expr));
                 Ok((new_s_expr, true))
             }
-            RelOperator::Exchange(_) | RelOperator::AddRowNumber(_) | RelOperator::Pattern(_) => {
+            RelOperator::Exchange(_) | RelOperator::AddRowNumber(_) => {
                 unreachable!()
             }
             RelOperator::DummyTableScan(_)
@@ -709,6 +715,9 @@ impl DPhpy {
             RelOperator::Join(_) => {
                 new_s_expr.plan = join_expr.plan.clone();
                 new_s_expr.children = join_expr.children.clone();
+                if self.filters.is_empty() {
+                    return Ok(new_s_expr);
+                }
                 // Add filters to `new_s_expr`, then push down filters if possible
                 let mut predicates = vec![];
                 for filter in self.filters.iter() {
@@ -760,9 +769,9 @@ impl DPhpy {
         let rule = RuleFactory::create_rule(RuleID::PushDownFilterJoin, self.metadata.clone())?;
         let mut state = TransformResult::new();
         if rule
-            .patterns()
+            .matchers()
             .iter()
-            .any(|pattern| s_expr.match_pattern(pattern))
+            .any(|matcher| matcher.matches(&s_expr))
             && !s_expr.applied_rule(&rule.id())
         {
             s_expr.set_applied_rule(&rule.id());
diff --git a/src/query/sql/src/planner/optimizer/hyper_dp/join_node.rs b/src/query/sql/src/planner/optimizer/hyper_dp/join_node.rs
index 191910e23275..b42cea78ff6b 100644
--- a/src/query/sql/src/planner/optimizer/hyper_dp/join_node.rs
+++ b/src/query/sql/src/planner/optimizer/hyper_dp/join_node.rs
@@ -87,7 +87,8 @@ impl JoinNode {
             marker_index: None,
             from_correlated_subquery: false,
             need_hold_hash_table: false,
-            broadcast: false,
+            is_lateral: false,
+            single_to_inner: None,
         });
         let children = self
             .children
diff --git a/src/query/profile/src/lib.rs b/src/query/sql/src/planner/optimizer/join/mod.rs
similarity index 85%
rename from src/query/profile/src/lib.rs
rename to src/query/sql/src/planner/optimizer/join/mod.rs
index 5c3fb77006e6..d7ea095997ba 100644
--- a/src/query/profile/src/lib.rs
+++ b/src/query/sql/src/planner/optimizer/join/mod.rs
@@ -12,10 +12,6 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-mod mgr;
-mod proc;
-mod prof;
+mod single_to_inner;
 
-pub use mgr::QueryProfileManager;
-pub use proc::*;
-pub use prof::*;
+pub use single_to_inner::SingleToInnerOptimizer;
diff --git a/src/query/sql/src/planner/optimizer/join/single_to_inner.rs b/src/query/sql/src/planner/optimizer/join/single_to_inner.rs
new file mode 100644
index 000000000000..85545c119952
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/join/single_to_inner.rs
@@ -0,0 +1,52 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+
+use databend_common_exception::Result;
+
+use crate::optimizer::SExpr;
+use crate::plans::JoinType;
+use crate::plans::RelOperator;
+
+// The SingleToInnerOptimizer will convert some single join to inner join.
+pub struct SingleToInnerOptimizer {}
+
+impl SingleToInnerOptimizer {
+    pub fn new() -> Self {
+        SingleToInnerOptimizer {}
+    }
+
+    pub fn run(self, s_expr: &SExpr) -> Result<SExpr> {
+        Self::single_to_inner(s_expr)
+    }
+
+    fn single_to_inner(s_expr: &SExpr) -> Result<SExpr> {
+        let s_expr = if let RelOperator::Join(join) = s_expr.plan.as_ref() {
+            let mut join = join.clone();
+            if join.single_to_inner.is_some() {
+                join.join_type = JoinType::Inner;
+            }
+            s_expr.replace_plan(Arc::new(RelOperator::Join(join)))
+        } else {
+            s_expr.clone()
+        };
+        let mut children = Vec::with_capacity(s_expr.arity());
+        for child in s_expr.children() {
+            let child = Self::single_to_inner(child)?;
+            children.push(Arc::new(child));
+        }
+        Ok(s_expr.replace_children(children))
+    }
+}
diff --git a/src/query/sql/src/planner/optimizer/m_expr.rs b/src/query/sql/src/planner/optimizer/m_expr.rs
index 498ab8691710..630bfd40fa64 100644
--- a/src/query/sql/src/planner/optimizer/m_expr.rs
+++ b/src/query/sql/src/planner/optimizer/m_expr.rs
@@ -18,13 +18,11 @@ use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 
 use super::group::Group;
+use crate::optimizer::extract::PatternExtractor;
 use crate::optimizer::memo::Memo;
-use crate::optimizer::pattern_extractor::PatternExtractor;
 use crate::optimizer::rule::AppliedRules;
 use crate::optimizer::rule::RulePtr;
 use crate::optimizer::rule::TransformResult;
-use crate::optimizer::SExpr;
-use crate::plans::Operator;
 use crate::plans::RelOperator;
 use crate::IndexType;
 
@@ -76,19 +74,6 @@ impl MExpr {
         memo.group(*group_index)
     }
 
-    /// Doesn't check if children are matched
-    pub fn match_pattern(&self, _memo: &Memo, pattern: &SExpr) -> bool {
-        if pattern.is_pattern() {
-            return true;
-        }
-
-        if self.arity() != pattern.arity() {
-            return false;
-        }
-
-        self.plan.rel_op() == pattern.plan().rel_op()
-    }
-
     pub fn apply_rule(
         &self,
         memo: &Memo,
@@ -99,8 +84,8 @@ impl MExpr {
             return Ok(());
         }
 
-        let mut extractor = PatternExtractor::create();
-        for pattern in rule.patterns() {
+        let mut extractor = PatternExtractor::new();
+        for pattern in rule.matchers() {
             let exprs = extractor.extract(memo, self, pattern)?;
             for expr in exprs.iter() {
                 rule.apply(expr, transform_state)?;
diff --git a/src/query/sql/src/planner/optimizer/mod.rs b/src/query/sql/src/planner/optimizer/mod.rs
index e2fad7131b44..5885fb62f0ad 100644
--- a/src/query/sql/src/planner/optimizer/mod.rs
+++ b/src/query/sql/src/planner/optimizer/mod.rs
@@ -16,14 +16,16 @@ mod cascades;
 mod cost;
 mod decorrelate;
 mod distributed;
+mod extract;
+mod filter;
 mod format;
 mod group;
 mod hyper_dp;
+mod join;
 mod m_expr;
 mod memo;
 #[allow(clippy::module_inception)]
 mod optimizer;
-mod pattern_extractor;
 mod property;
 mod rule;
 pub mod s_expr;
@@ -32,6 +34,7 @@ mod util;
 pub use cascades::CascadesOptimizer;
 pub use decorrelate::FlattenInfo;
 pub use decorrelate::SubqueryRewriter;
+pub use extract::PatternExtractor;
 pub use hyper_dp::DPhpy;
 pub use m_expr::MExpr;
 pub use memo::Memo;
@@ -39,7 +42,6 @@ pub use optimizer::optimize;
 pub use optimizer::optimize_query;
 pub use optimizer::OptimizerContext;
 pub use optimizer::RecursiveOptimizer;
-pub use pattern_extractor::PatternExtractor;
 pub use property::*;
 pub use rule::agg_index;
 pub use rule::try_push_down_filter_join;
diff --git a/src/query/sql/src/planner/optimizer/optimizer.rs b/src/query/sql/src/planner/optimizer/optimizer.rs
index f5f02c7c3dec..f4e89dbf3884 100644
--- a/src/query/sql/src/planner/optimizer/optimizer.rs
+++ b/src/query/sql/src/planner/optimizer/optimizer.rs
@@ -32,7 +32,10 @@ use crate::optimizer::cascades::CascadesOptimizer;
 use crate::optimizer::decorrelate::decorrelate_subquery;
 use crate::optimizer::distributed::optimize_distributed_query;
 use crate::optimizer::distributed::SortAndLimitPushDownOptimizer;
+use crate::optimizer::filter::DeduplicateJoinConditionOptimizer;
+use crate::optimizer::filter::PullUpFilterOptimizer;
 use crate::optimizer::hyper_dp::DPhpy;
+use crate::optimizer::join::SingleToInnerOptimizer;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::util::contains_local_table_scan;
 use crate::optimizer::RuleFactory;
@@ -119,14 +122,13 @@ impl<'a> RecursiveOptimizer<'a> {
 
     fn apply_transform_rules(&self, s_expr: &SExpr, rules: &[RuleID]) -> Result<SExpr> {
         let mut s_expr = s_expr.clone();
-
         for rule_id in rules {
             let rule = RuleFactory::create_rule(*rule_id, self.ctx.metadata.clone())?;
             let mut state = TransformResult::new();
             if rule
-                .patterns()
+                .matchers()
                 .iter()
-                .any(|pattern| s_expr.match_pattern(pattern))
+                .any(|matcher| matcher.matches(&s_expr))
                 && !s_expr.applied_rule(&rule.id())
             {
                 s_expr.set_applied_rule(&rule.id());
@@ -222,6 +224,9 @@ pub fn optimize_query(opt_ctx: OptimizerContext, mut s_expr: SExpr) -> Result<SE
         s_expr = decorrelate_subquery(opt_ctx.metadata.clone(), s_expr.clone())?;
     }
 
+    // Pull up and infer filter.
+    s_expr = PullUpFilterOptimizer::new(opt_ctx.metadata.clone()).run(&s_expr)?;
+
     // Run default rewrite rules
     s_expr = RecursiveOptimizer::new(&DEFAULT_REWRITE_RULES, &opt_ctx).run(&s_expr)?;
 
@@ -232,10 +237,16 @@ pub fn optimize_query(opt_ctx: OptimizerContext, mut s_expr: SExpr) -> Result<SE
             .optimize(Arc::new(s_expr.clone()))?;
         if optimized {
             s_expr = (*dp_res).clone();
-            s_expr = RecursiveOptimizer::new(&[RuleID::CommuteJoin], &opt_ctx).run(&s_expr)?;
             dphyp_optimized = true;
         }
     }
+
+    // After join reorder, Convert some single join to inner join.
+    s_expr = SingleToInnerOptimizer::new().run(&s_expr)?;
+
+    // Deduplicate join conditions.
+    s_expr = DeduplicateJoinConditionOptimizer::new().run(&s_expr)?;
+
     let mut cascades = CascadesOptimizer::new(
         opt_ctx.table_ctx.clone(),
         opt_ctx.metadata.clone(),
@@ -246,8 +257,13 @@ pub fn optimize_query(opt_ctx: OptimizerContext, mut s_expr: SExpr) -> Result<SE
     // Cascades optimizer may fail due to timeout, fallback to heuristic optimizer in this case.
     s_expr = match cascades.optimize(s_expr.clone()) {
         Ok(mut s_expr) => {
-            s_expr =
-                RecursiveOptimizer::new(&[RuleID::EliminateEvalScalar], &opt_ctx).run(&s_expr)?;
+            let rules = if opt_ctx.enable_join_reorder {
+                [RuleID::EliminateEvalScalar, RuleID::CommuteJoin].as_slice()
+            } else {
+                [RuleID::EliminateEvalScalar].as_slice()
+            };
+
+            s_expr = RecursiveOptimizer::new(rules, &opt_ctx).run(&s_expr)?;
 
             // Push down sort and limit
             // TODO(leiysky): do this optimization in cascades optimizer
@@ -340,6 +356,8 @@ fn optimize_merge_into(opt_ctx: OptimizerContext, plan: Box<MergeInto>) -> Resul
         Arc::new(right_source),
     ]));
 
+    let join_op = Join::try_from(join_sexpr.plan().clone())?;
+    let non_equal_join = join_op.right_conditions.is_empty() && join_op.left_conditions.is_empty();
     // before, we think source table is always the small table.
     // 1. for matched only, we use inner join
     // 2. for insert only, we use right anti join
@@ -400,7 +418,7 @@ fn optimize_merge_into(opt_ctx: OptimizerContext, plan: Box<MergeInto>) -> Resul
     {
         // distributed execution stargeties:
         // I. change join order is true, we use the `optimize_distributed_query`'s result.
-        // II. change join order is false and match_pattern and not enable spill, we use right outer join with rownumber distributed strategies.
+        // II. change join order is false and match_pattern and not enable spill and not non-equal-join, we use right outer join with rownumber distributed strategies.
         // III otherwise, use `merge_into_join_sexpr` as standalone execution(so if change join order is false,but doesn't match_pattern, we don't support distributed,in fact. case I
         // can take this at most time, if that's a hash shuffle, the I can take it. We think source is always very small).
         // input is a Join_SExpr
@@ -415,7 +433,10 @@ fn optimize_merge_into(opt_ctx: OptimizerContext, plan: Box<MergeInto>) -> Resul
             .get_join_spilling_threshold()?
             == 0
             && !change_join_order
-            && merge_into_join_sexpr.match_pattern(&merge_source_optimizer.merge_source_pattern)
+            && merge_source_optimizer
+                .merge_source_matcher
+                .matches(&merge_into_join_sexpr)
+            && !non_equal_join
         {
             (
                 merge_source_optimizer.optimize(&merge_into_join_sexpr)?,
@@ -472,8 +493,7 @@ fn try_to_change_as_broadcast_join(
     if let RelOperator::Exchange(Exchange::Merge) = merge_into_join_sexpr.plan.as_ref() {
         let right_exchange = merge_into_join_sexpr.child(0)?.child(1)?;
         if let RelOperator::Exchange(Exchange::Broadcast) = right_exchange.plan.as_ref() {
-            let mut join: Join = merge_into_join_sexpr.child(0)?.plan().clone().try_into()?;
-            join.broadcast = true;
+            let join: Join = merge_into_join_sexpr.child(0)?.plan().clone().try_into()?;
             let join_s_expr = merge_into_join_sexpr
                 .child(0)?
                 .replace_plan(Arc::new(RelOperator::Join(join)));
diff --git a/src/query/sql/src/planner/optimizer/pattern_extractor.rs b/src/query/sql/src/planner/optimizer/pattern_extractor.rs
deleted file mode 100644
index ea3d4b8b418e..000000000000
--- a/src/query/sql/src/planner/optimizer/pattern_extractor.rs
+++ /dev/null
@@ -1,160 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::sync::Arc;
-
-use databend_common_exception::ErrorCode;
-use databend_common_exception::Result;
-
-use crate::optimizer::group::Group;
-use crate::optimizer::m_expr::MExpr;
-use crate::optimizer::memo::Memo;
-use crate::optimizer::SExpr;
-
-/// A helper to extract `SExpr`s from `Memo` that match the given pattern.
-pub struct PatternExtractor {}
-
-impl PatternExtractor {
-    pub fn create() -> Self {
-        PatternExtractor {}
-    }
-
-    pub fn extract(&mut self, memo: &Memo, m_expr: &MExpr, pattern: &SExpr) -> Result<Vec<SExpr>> {
-        if !m_expr.match_pattern(memo, pattern) {
-            return Ok(vec![]);
-        }
-
-        if pattern.is_pattern() {
-            // Expand the pattern node to a complete `SExpr`.
-            let child = Self::expand_pattern(memo, m_expr)?;
-            return Ok(vec![child]);
-        }
-
-        let pattern_children = pattern.children();
-
-        if m_expr.arity() != pattern_children.len() {
-            return Ok(vec![]);
-        }
-
-        let mut children_results = vec![];
-        for (i, child) in m_expr.children.iter().enumerate().take(m_expr.arity()) {
-            let pattern = &pattern_children[i];
-            let child_group = memo.group(*child)?;
-            let result = self.extract_group(memo, child_group, pattern)?;
-            children_results.push(result);
-        }
-
-        Self::generate_expression_with_children(memo, m_expr, children_results)
-    }
-
-    fn extract_group(&mut self, memo: &Memo, group: &Group, pattern: &SExpr) -> Result<Vec<SExpr>> {
-        let mut results = vec![];
-        for m_expr in group.m_exprs.iter() {
-            let result = self.extract(memo, m_expr, pattern)?;
-            results.extend(result.into_iter());
-        }
-
-        Ok(results)
-    }
-
-    fn generate_expression_with_children(
-        memo: &Memo,
-        m_expr: &MExpr,
-        candidates: Vec<Vec<SExpr>>,
-    ) -> Result<Vec<SExpr>> {
-        let mut results = vec![];
-
-        // Initialize cursors
-        let mut cursors: Vec<usize> = vec![];
-        for candidate in candidates.iter() {
-            if candidate.is_empty() {
-                // Every child should have at least one candidate
-                return Ok(results);
-            }
-            cursors.push(0);
-        }
-
-        if cursors.is_empty() {
-            results.push(SExpr::create(
-                m_expr.plan.clone(),
-                vec![],
-                Some(m_expr.group_index),
-                Some(memo.group(m_expr.group_index)?.relational_prop.clone()),
-                Some(memo.group(m_expr.group_index)?.stat_info.clone()),
-            ));
-            return Ok(results);
-        }
-
-        'LOOP: loop {
-            let mut children = vec![];
-            for (index, cursor) in cursors.iter().enumerate() {
-                children.push(Arc::new(candidates[index][*cursor].clone()));
-            }
-            results.push(SExpr::create(
-                m_expr.plan.clone(),
-                children,
-                Some(m_expr.group_index),
-                Some(memo.group(m_expr.group_index)?.relational_prop.clone()),
-                Some(memo.group(m_expr.group_index)?.stat_info.clone()),
-            ));
-
-            let mut shifted = false;
-            // Shift cursor
-            for i in (0..cursors.len()).rev() {
-                if !shifted {
-                    // Shift cursor
-                    cursors[i] += 1;
-                    shifted = true;
-                }
-
-                if i == 0 && cursors[0] > candidates[0].len() - 1 {
-                    // Candidates are exhausted
-                    break 'LOOP;
-                } else if i > 0 && cursors[i] > candidates[i].len() - 1 {
-                    // Shift previous children
-                    cursors[i] = 0;
-                    cursors[i - 1] += 1;
-                    continue;
-                } else {
-                    break;
-                }
-            }
-        }
-
-        Ok(results)
-    }
-
-    /// Expand a `Pattern` node to an arbitrary `SExpr` with `m_expr` as the root.
-    /// Since we don't care about the actual content of the `Pattern` node, we will
-    /// choose the first `MExpr` in each group to construct the `SExpr`.
-    fn expand_pattern(memo: &Memo, m_expr: &MExpr) -> Result<SExpr> {
-        let mut children = Vec::with_capacity(m_expr.arity());
-        for child in m_expr.children.iter() {
-            let child_group = memo.group(*child)?;
-            let child_m_expr = child_group
-                .m_exprs
-                .first()
-                .ok_or_else(|| ErrorCode::Internal(format!("No MExpr in group {child}")))?;
-            children.push(Arc::new(Self::expand_pattern(memo, child_m_expr)?));
-        }
-
-        Ok(SExpr::create(
-            m_expr.plan.clone(),
-            children,
-            Some(m_expr.group_index),
-            Some(memo.group(m_expr.group_index)?.relational_prop.clone()),
-            Some(memo.group(m_expr.group_index)?.stat_info.clone()),
-        ))
-    }
-}
diff --git a/src/query/sql/src/planner/optimizer/rule/factory.rs b/src/query/sql/src/planner/optimizer/rule/factory.rs
index e57b176e31c8..55426f0476fc 100644
--- a/src/query/sql/src/planner/optimizer/rule/factory.rs
+++ b/src/query/sql/src/planner/optimizer/rule/factory.rs
@@ -17,15 +17,13 @@ use databend_common_exception::Result;
 use super::rewrite::RuleCommuteJoin;
 use super::rewrite::RuleEliminateEvalScalar;
 use super::rewrite::RuleFoldCountAggregate;
-use super::rewrite::RuleInferFilter;
-use super::rewrite::RuleNormalizeDisjunctiveFilter;
 use super::rewrite::RuleNormalizeScalarFilter;
 use super::rewrite::RulePushDownFilterAggregate;
 use super::rewrite::RulePushDownFilterEvalScalar;
 use super::rewrite::RulePushDownFilterJoin;
 use super::rewrite::RulePushDownFilterWindow;
 use super::rewrite::RulePushDownLimitAggregate;
-use super::rewrite::RulePushDownLimitExpression;
+use super::rewrite::RulePushDownLimitEvalScalar;
 use super::rewrite::RulePushDownPrewhere;
 use super::rewrite::RuleTryApplyAggIndex;
 use crate::optimizer::rule::rewrite::RuleEliminateFilter;
@@ -41,6 +39,7 @@ use crate::optimizer::rule::rewrite::RulePushDownLimitOuterJoin;
 use crate::optimizer::rule::rewrite::RulePushDownLimitScan;
 use crate::optimizer::rule::rewrite::RulePushDownLimitSort;
 use crate::optimizer::rule::rewrite::RulePushDownLimitUnion;
+use crate::optimizer::rule::rewrite::RulePushDownLimitWindow;
 use crate::optimizer::rule::rewrite::RulePushDownSortScan;
 use crate::optimizer::rule::rewrite::RuleSemiToInnerJoin;
 use crate::optimizer::rule::rewrite::RuleSplitAggregate;
@@ -58,10 +57,8 @@ impl RuleFactory {
         match id {
             RuleID::EliminateEvalScalar => Ok(Box::new(RuleEliminateEvalScalar::new())),
             RuleID::PushDownFilterUnion => Ok(Box::new(RulePushDownFilterUnion::new())),
-            RuleID::PushDownFilterEvalScalar => {
-                Ok(Box::new(RulePushDownFilterEvalScalar::new(metadata)))
-            }
-            RuleID::PushDownFilterJoin => Ok(Box::new(RulePushDownFilterJoin::new(metadata))),
+            RuleID::PushDownFilterEvalScalar => Ok(Box::new(RulePushDownFilterEvalScalar::new())),
+            RuleID::PushDownFilterJoin => Ok(Box::new(RulePushDownFilterJoin::new())),
             RuleID::PushDownFilterScan => Ok(Box::new(RulePushDownFilterScan::new(metadata))),
             RuleID::PushDownFilterSort => Ok(Box::new(RulePushDownFilterSort::new())),
             RuleID::PushDownFilterProjectSet => Ok(Box::new(RulePushDownFilterProjectSet::new())),
@@ -69,8 +66,9 @@ impl RuleFactory {
             RuleID::PushDownLimitScan => Ok(Box::new(RulePushDownLimitScan::new())),
             RuleID::PushDownSortScan => Ok(Box::new(RulePushDownSortScan::new())),
             RuleID::PushDownLimitOuterJoin => Ok(Box::new(RulePushDownLimitOuterJoin::new())),
-            RuleID::PushDownLimitExpression => Ok(Box::new(RulePushDownLimitExpression::new())),
+            RuleID::PushDownLimitEvalScalar => Ok(Box::new(RulePushDownLimitEvalScalar::new())),
             RuleID::PushDownLimitSort => Ok(Box::new(RulePushDownLimitSort::new())),
+            RuleID::PushDownLimitWindow => Ok(Box::new(RulePushDownLimitWindow::new())),
             RuleID::PushDownLimitAggregate => Ok(Box::new(RulePushDownLimitAggregate::new())),
             RuleID::PushDownFilterAggregate => Ok(Box::new(RulePushDownFilterAggregate::new())),
             RuleID::PushDownFilterWindow => Ok(Box::new(RulePushDownFilterWindow::new())),
@@ -81,10 +79,6 @@ impl RuleFactory {
             RuleID::NormalizeAggregate => Ok(Box::new(RuleNormalizeAggregate::new())),
             RuleID::SplitAggregate => Ok(Box::new(RuleSplitAggregate::new())),
             RuleID::FoldCountAggregate => Ok(Box::new(RuleFoldCountAggregate::new())),
-            RuleID::NormalizeDisjunctiveFilter => {
-                Ok(Box::new(RuleNormalizeDisjunctiveFilter::new()))
-            }
-            RuleID::InferFilter => Ok(Box::new(RuleInferFilter::new())),
             RuleID::CommuteJoin => Ok(Box::new(RuleCommuteJoin::new())),
             RuleID::CommuteJoinBaseTable => Ok(Box::new(RuleCommuteJoinBaseTable::new())),
             RuleID::LeftExchangeJoin => Ok(Box::new(RuleLeftExchangeJoin::new())),
@@ -92,7 +86,7 @@ impl RuleFactory {
             RuleID::PushDownPrewhere => Ok(Box::new(RulePushDownPrewhere::new(metadata))),
             RuleID::TryApplyAggIndex => Ok(Box::new(RuleTryApplyAggIndex::new(metadata))),
             RuleID::EliminateSort => Ok(Box::new(RuleEliminateSort::new())),
-            RuleID::SemiToInnerJoin => Ok(Box::new(RuleSemiToInnerJoin::new(metadata))),
+            RuleID::SemiToInnerJoin => Ok(Box::new(RuleSemiToInnerJoin::new())),
         }
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/derive_filter.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/derive_filter.rs
deleted file mode 100644
index b2d448218888..000000000000
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/derive_filter.rs
+++ /dev/null
@@ -1,156 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::collections::HashMap;
-use std::sync::Arc;
-
-use databend_common_exception::Result;
-
-use crate::optimizer::SExpr;
-use crate::plans::walk_expr_mut;
-use crate::plans::Filter;
-use crate::plans::Join;
-use crate::plans::JoinType;
-use crate::plans::VisitorMut;
-use crate::ScalarExpr;
-
-/// Derive filter to push down
-pub fn try_derive_predicates(
-    s_expr: &SExpr,
-    join: Join,
-    mut left_push_down: Vec<ScalarExpr>,
-    mut right_push_down: Vec<ScalarExpr>,
-) -> Result<SExpr> {
-    let join_expr = s_expr.child(0)?;
-    let mut left_child = join_expr.child(0)?.clone();
-    let mut right_child = join_expr.child(1)?.clone();
-
-    if join.join_type == JoinType::Inner {
-        let mut new_left_push_down = vec![];
-        let mut new_right_push_down = vec![];
-        for predicate in left_push_down.iter() {
-            let used_columns = predicate.used_columns();
-            let mut equi_conditions_map = HashMap::new();
-            for (idx, left_condition) in join.left_conditions.iter().enumerate() {
-                if left_condition.used_columns().len() > 1
-                    || !left_condition.used_columns().is_subset(&used_columns)
-                {
-                    continue;
-                }
-                equi_conditions_map.insert(left_condition, &join.right_conditions[idx]);
-            }
-            if used_columns.len() == equi_conditions_map.len() {
-                derive_predicate(
-                    &mut equi_conditions_map,
-                    predicate,
-                    &mut new_right_push_down,
-                )?;
-            }
-        }
-        for predicate in right_push_down.iter() {
-            let used_columns = predicate.used_columns();
-            let mut equi_conditions_map = HashMap::new();
-            for (idx, right_condition) in join.right_conditions.iter().enumerate() {
-                if right_condition.used_columns().len() > 1
-                    || !right_condition.used_columns().is_subset(&used_columns)
-                {
-                    continue;
-                }
-                equi_conditions_map.insert(right_condition, &join.left_conditions[idx]);
-            }
-            if used_columns.len() == equi_conditions_map.len() {
-                derive_predicate(&mut equi_conditions_map, predicate, &mut new_left_push_down)?;
-            }
-        }
-        left_push_down.extend(new_left_push_down);
-        right_push_down.extend(new_right_push_down);
-    }
-
-    if !left_push_down.is_empty() {
-        left_child = SExpr::create_unary(
-            Arc::new(
-                Filter {
-                    predicates: left_push_down,
-                }
-                .into(),
-            ),
-            Arc::new(left_child),
-        );
-    }
-
-    if !right_push_down.is_empty() {
-        right_child = SExpr::create_unary(
-            Arc::new(
-                Filter {
-                    predicates: right_push_down,
-                }
-                .into(),
-            ),
-            Arc::new(right_child),
-        );
-    }
-    Ok(SExpr::create_binary(
-        Arc::new(join.into()),
-        Arc::new(left_child),
-        Arc::new(right_child),
-    ))
-}
-
-fn derive_predicate(
-    equi_conditions_map: &mut HashMap<&ScalarExpr, &ScalarExpr>,
-    predicate: &ScalarExpr,
-    new_push_down: &mut Vec<ScalarExpr>,
-) -> Result<()> {
-    let mut replaced_predicate = predicate.clone();
-    replace_column(&mut replaced_predicate, equi_conditions_map);
-    if &replaced_predicate != predicate {
-        new_push_down.push(replaced_predicate);
-    }
-    Ok(())
-}
-
-fn replace_column(
-    scalar: &mut ScalarExpr,
-    equi_conditions_map: &mut HashMap<&ScalarExpr, &ScalarExpr>,
-) {
-    struct ReplaceColumn<'a> {
-        equi_conditions_map: &'a HashMap<&'a ScalarExpr, &'a ScalarExpr>,
-    }
-
-    impl<'a> VisitorMut<'_> for ReplaceColumn<'a> {
-        fn visit(&mut self, expr: &mut ScalarExpr) -> Result<()> {
-            if let Some(e) = self.equi_conditions_map.get(expr) {
-                *expr = (**e).clone();
-                return Ok(());
-            } else if let ScalarExpr::BoundColumnRef(col) = expr {
-                for (key, val) in self.equi_conditions_map.iter() {
-                    if let ScalarExpr::BoundColumnRef(key_col) = key {
-                        if key_col.column.index == col.column.index {
-                            *expr = (**val).clone();
-                            return Ok(());
-                        }
-                    }
-                }
-            }
-
-            walk_expr_mut(self, expr)
-        }
-    }
-
-    let mut replace_column = ReplaceColumn {
-        equi_conditions_map,
-    };
-
-    replace_column.visit(scalar).unwrap();
-}
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/outer_join_to_inner_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/outer_join_to_inner_join.rs
deleted file mode 100644
index 3c02c76cb946..000000000000
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/outer_join_to_inner_join.rs
+++ /dev/null
@@ -1,99 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-#[cfg(feature = "z3-prove")]
-use std::sync::Arc;
-
-use databend_common_exception::Result;
-
-use crate::optimizer::SExpr;
-use crate::plans::Join;
-
-pub fn outer_to_inner(s_expr: &SExpr) -> Result<SExpr> {
-    let join: Join = s_expr.child(0)?.plan().clone().try_into()?;
-    let origin_join_type = join.join_type.clone();
-    if !origin_join_type.is_outer_join() {
-        return Ok(s_expr.clone());
-    }
-
-    #[cfg(feature = "z3-prove")]
-    {
-        use crate::optimizer::RelExpr;
-        use crate::plans::Filter;
-        use crate::plans::JoinType;
-
-        let mut join = join;
-        let filter: Filter = s_expr.plan().clone().try_into()?;
-        let constraint_set = crate::optimizer::ConstraintSet::new(&filter.predicates);
-
-        let join_expr = RelExpr::with_s_expr(s_expr.child(0)?);
-        let left_columns = join_expr
-            .derive_relational_prop_child(0)?
-            .output_columns
-            .clone();
-        let right_columns = join_expr
-            .derive_relational_prop_child(1)?
-            .output_columns
-            .clone();
-
-        let eliminate_left_null = left_columns
-            .iter()
-            .any(|col| constraint_set.is_null_reject(col));
-        let eliminate_right_null = right_columns
-            .iter()
-            .any(|col| constraint_set.is_null_reject(col));
-
-        let new_join_type = match join.join_type {
-            JoinType::Left => {
-                if eliminate_right_null {
-                    JoinType::Inner
-                } else {
-                    JoinType::Left
-                }
-            }
-            JoinType::Right => {
-                if eliminate_left_null {
-                    JoinType::Inner
-                } else {
-                    JoinType::Right
-                }
-            }
-            JoinType::Full => {
-                if eliminate_left_null && eliminate_right_null {
-                    JoinType::Inner
-                } else if eliminate_left_null {
-                    JoinType::Left
-                } else if eliminate_right_null {
-                    JoinType::Right
-                } else {
-                    JoinType::Full
-                }
-            }
-            _ => unreachable!(),
-        };
-
-        join.join_type = new_join_type;
-        Ok(SExpr::create_unary(
-            Arc::new(filter.into()),
-            Arc::new(SExpr::create_binary(
-                Arc::new(join.into()),
-                Arc::new(s_expr.child(0)?.child(0)?.clone()),
-                Arc::new(s_expr.child(0)?.child(1)?.clone()),
-            )),
-        ))
-    }
-
-    #[cfg(not(feature = "z3-prove"))]
-    Ok(s_expr.clone())
-}
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/mod.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/mod.rs
index b2805ee5aa53..1a01c7ac98eb 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/mod.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/mod.rs
@@ -13,17 +13,15 @@
 // limitations under the License.
 
 pub mod agg_index;
-mod filter_join;
+mod push_down_filter_join;
 mod rule_commute_join;
 mod rule_eliminate_eval_scalar;
 mod rule_eliminate_filter;
 mod rule_eliminate_sort;
 mod rule_fold_count_aggregate;
-mod rule_infer_filter;
 mod rule_merge_eval_scalar;
 mod rule_merge_filter;
 mod rule_normalize_aggregate;
-mod rule_normalize_disjunctive_filter;
 mod rule_normalize_scalar;
 mod rule_push_down_filter_aggregate;
 mod rule_push_down_filter_eval_scalar;
@@ -39,6 +37,7 @@ mod rule_push_down_limit_join;
 mod rule_push_down_limit_scan;
 mod rule_push_down_limit_sort;
 mod rule_push_down_limit_union;
+mod rule_push_down_limit_window;
 mod rule_push_down_prewhere;
 mod rule_push_down_sort_scan;
 mod rule_semi_to_inner_join;
@@ -50,11 +49,9 @@ pub use rule_eliminate_eval_scalar::RuleEliminateEvalScalar;
 pub use rule_eliminate_filter::RuleEliminateFilter;
 pub use rule_eliminate_sort::RuleEliminateSort;
 pub use rule_fold_count_aggregate::RuleFoldCountAggregate;
-pub use rule_infer_filter::RuleInferFilter;
 pub use rule_merge_eval_scalar::RuleMergeEvalScalar;
 pub use rule_merge_filter::RuleMergeFilter;
 pub use rule_normalize_aggregate::RuleNormalizeAggregate;
-pub use rule_normalize_disjunctive_filter::RuleNormalizeDisjunctiveFilter;
 pub use rule_normalize_scalar::RuleNormalizeScalarFilter;
 pub use rule_push_down_filter_aggregate::RulePushDownFilterAggregate;
 pub use rule_push_down_filter_eval_scalar::RulePushDownFilterEvalScalar;
@@ -66,11 +63,12 @@ pub use rule_push_down_filter_sort::RulePushDownFilterSort;
 pub use rule_push_down_filter_union::RulePushDownFilterUnion;
 pub use rule_push_down_filter_window::RulePushDownFilterWindow;
 pub use rule_push_down_limit_aggregate::RulePushDownLimitAggregate;
-pub use rule_push_down_limit_expression::RulePushDownLimitExpression;
+pub use rule_push_down_limit_expression::RulePushDownLimitEvalScalar;
 pub use rule_push_down_limit_join::RulePushDownLimitOuterJoin;
 pub use rule_push_down_limit_scan::RulePushDownLimitScan;
 pub use rule_push_down_limit_sort::RulePushDownLimitSort;
 pub use rule_push_down_limit_union::RulePushDownLimitUnion;
+pub use rule_push_down_limit_window::RulePushDownLimitWindow;
 pub use rule_push_down_prewhere::RulePushDownPrewhere;
 pub use rule_push_down_sort_scan::RulePushDownSortScan;
 pub use rule_semi_to_inner_join::RuleSemiToInnerJoin;
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/extract_or_predicates.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/extract_or_predicates.rs
similarity index 100%
rename from src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/extract_or_predicates.rs
rename to src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/extract_or_predicates.rs
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/mark_join_to_semi_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/mark_join_to_semi_join.rs
similarity index 86%
rename from src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/mark_join_to_semi_join.rs
rename to src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/mark_join_to_semi_join.rs
index 028e4455666b..bf8a758de8ac 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/mark_join_to_semi_join.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/mark_join_to_semi_join.rs
@@ -22,14 +22,14 @@ use crate::plans::Join;
 use crate::plans::JoinType;
 use crate::ScalarExpr;
 
-pub fn convert_mark_to_semi_join(s_expr: &SExpr) -> Result<SExpr> {
+pub fn convert_mark_to_semi_join(s_expr: &SExpr) -> Result<(SExpr, bool)> {
     let mut filter: Filter = s_expr.plan().clone().try_into()?;
     let mut join: Join = s_expr.child(0)?.plan().clone().try_into()?;
     let has_disjunction = filter.predicates.iter().any(
         |predicate| matches!(predicate, ScalarExpr::FunctionCall(func) if func.func_name == "or"),
     );
     if !join.join_type.is_mark_join() || has_disjunction {
-        return Ok(s_expr.clone());
+        return Ok((s_expr.clone(), false));
     }
 
     let mark_index = join.marker_index.unwrap();
@@ -47,7 +47,7 @@ pub fn convert_mark_to_semi_join(s_expr: &SExpr) -> Result<SExpr> {
                 // Check if the argument is mark index, if so, we won't convert it to semi join
                 if let ScalarExpr::BoundColumnRef(col) = &func.arguments[0] {
                     if col.column.index == mark_index {
-                        return Ok(s_expr.clone());
+                        return Ok((s_expr.clone(), false));
                     }
                 }
             }
@@ -57,7 +57,7 @@ pub fn convert_mark_to_semi_join(s_expr: &SExpr) -> Result<SExpr> {
 
     if !find_mark_index {
         // To be conservative, we do not convert
-        return Ok(s_expr.clone());
+        return Ok((s_expr.clone(), false));
     }
 
     join.join_type = match join.join_type {
@@ -73,6 +73,8 @@ pub fn convert_mark_to_semi_join(s_expr: &SExpr) -> Result<SExpr> {
         Arc::new(s_join_expr.child(1)?.clone()),
     );
 
-    result = SExpr::create_unary(Arc::new(filter.into()), Arc::new(result));
-    Ok(result)
+    if !filter.predicates.is_empty() {
+        result = SExpr::create_unary(Arc::new(filter.into()), Arc::new(result));
+    }
+    Ok((result, true))
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/mod.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/mod.rs
similarity index 87%
rename from src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/mod.rs
rename to src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/mod.rs
index f35ee3d9e5ef..1f1070b49249 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/filter_join/mod.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/mod.rs
@@ -12,12 +12,11 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-mod derive_filter;
 mod extract_or_predicates;
 mod mark_join_to_semi_join;
 mod outer_join_to_inner_join;
 
-pub use derive_filter::try_derive_predicates;
 pub use extract_or_predicates::rewrite_predicates;
 pub use mark_join_to_semi_join::convert_mark_to_semi_join;
-pub use outer_join_to_inner_join::outer_to_inner;
+pub use outer_join_to_inner_join::can_filter_null;
+pub use outer_join_to_inner_join::outer_join_to_inner_join;
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/outer_join_to_inner_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/outer_join_to_inner_join.rs
new file mode 100644
index 000000000000..f9193afb2857
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/push_down_filter_join/outer_join_to_inner_join.rs
@@ -0,0 +1,274 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::sync::Arc;
+
+use databend_common_exception::Result;
+use databend_common_expression::DataBlock;
+use databend_common_expression::DataSchema;
+use databend_common_expression::Evaluator;
+use databend_common_expression::Expr;
+use databend_common_expression::FunctionContext;
+use databend_common_expression::Scalar;
+use databend_common_expression::Value;
+use databend_common_functions::BUILTIN_FUNCTIONS;
+
+use crate::binder::JoinPredicate;
+use crate::executor::cast_expr_to_non_null_boolean;
+use crate::optimizer::RelExpr;
+use crate::optimizer::SExpr;
+use crate::plans::ConstantExpr;
+use crate::plans::Filter;
+use crate::plans::Join;
+use crate::plans::JoinType;
+use crate::ColumnSet;
+use crate::ScalarExpr;
+use crate::TypeCheck;
+
+pub fn outer_join_to_inner_join(s_expr: &SExpr) -> Result<(SExpr, bool)> {
+    let mut join: Join = s_expr.child(0)?.plan().clone().try_into()?;
+    if !join.join_type.is_outer_join() {
+        return Ok((s_expr.clone(), false));
+    }
+
+    let filter: Filter = s_expr.plan().clone().try_into()?;
+    let join_s_expr = s_expr.child(0)?;
+    let join_rel_expr = RelExpr::with_s_expr(join_s_expr);
+
+    let mut can_filter_left_null = false;
+    let mut can_filter_right_null = false;
+    let left_prop = join_rel_expr.derive_relational_prop_child(0)?;
+    let right_prop = join_rel_expr.derive_relational_prop_child(1)?;
+    for predicate in &filter.predicates {
+        let pred = JoinPredicate::new(predicate, &left_prop, &right_prop);
+        match pred {
+            JoinPredicate::Left(_)
+                if can_filter_null(
+                    predicate,
+                    &left_prop.output_columns,
+                    &right_prop.output_columns,
+                )? =>
+            {
+                can_filter_left_null = true;
+            }
+            JoinPredicate::Right(_)
+                if can_filter_null(
+                    predicate,
+                    &left_prop.output_columns,
+                    &right_prop.output_columns,
+                )? =>
+            {
+                can_filter_right_null = true;
+            }
+            JoinPredicate::Both { .. }
+                if can_filter_null(
+                    predicate,
+                    &left_prop.output_columns,
+                    &right_prop.output_columns,
+                )? =>
+            {
+                can_filter_left_null = true;
+                can_filter_right_null = true;
+            }
+            _ => (),
+        }
+    }
+
+    #[cfg(feature = "z3-prove")]
+    {
+        let constraint_set = crate::optimizer::ConstraintSet::new(&filter.predicates);
+        let left_columns = join_rel_expr
+            .derive_relational_prop_child(0)?
+            .output_columns
+            .clone();
+        let right_columns = join_rel_expr
+            .derive_relational_prop_child(1)?
+            .output_columns
+            .clone();
+        can_filter_left_null |= left_columns
+            .iter()
+            .any(|col| constraint_set.is_null_reject(col));
+        can_filter_right_null |= right_columns
+            .iter()
+            .any(|col| constraint_set.is_null_reject(col));
+    }
+
+    let original_join_type = join.join_type.clone();
+    join.join_type =
+        eliminate_outer_join_type(join.join_type, can_filter_left_null, can_filter_right_null);
+    if join.join_type == original_join_type {
+        return Ok((s_expr.clone(), false));
+    }
+
+    if matches!(
+        original_join_type,
+        JoinType::LeftSingle | JoinType::RightSingle
+    ) {
+        join.join_type = original_join_type.clone();
+        join.single_to_inner = Some(original_join_type);
+    }
+
+    let result = SExpr::create_unary(
+        Arc::new(filter.into()),
+        Arc::new(SExpr::create_binary(
+            Arc::new(join.into()),
+            Arc::new(join_s_expr.child(0)?.clone()),
+            Arc::new(join_s_expr.child(1)?.clone()),
+        )),
+    );
+
+    Ok((result, true))
+}
+
+fn eliminate_outer_join_type(
+    join_type: JoinType,
+    can_filter_left_null: bool,
+    can_filter_right_null: bool,
+) -> JoinType {
+    match join_type {
+        JoinType::Left | JoinType::LeftSingle if can_filter_right_null => JoinType::Inner,
+        JoinType::Right | JoinType::RightSingle if can_filter_left_null => JoinType::Inner,
+        JoinType::Full => {
+            if can_filter_left_null && can_filter_right_null {
+                JoinType::Inner
+            } else if can_filter_left_null {
+                JoinType::Left
+            } else if can_filter_right_null {
+                JoinType::Right
+            } else {
+                join_type
+            }
+        }
+        _ => join_type,
+    }
+}
+
+pub fn can_filter_null(
+    predicate: &ScalarExpr,
+    left_output_columns: &ColumnSet,
+    right_output_columns: &ColumnSet,
+) -> Result<bool> {
+    struct ReplaceColumnBindingsNull<'a> {
+        can_replace: bool,
+        left_output_columns: &'a ColumnSet,
+        right_output_columns: &'a ColumnSet,
+    }
+
+    impl<'a> ReplaceColumnBindingsNull<'a> {
+        fn replace(
+            &mut self,
+            expr: &mut ScalarExpr,
+            column_set: &mut Option<ColumnSet>,
+        ) -> Result<()> {
+            if !self.can_replace {
+                return Ok(());
+            }
+            match expr {
+                ScalarExpr::BoundColumnRef(column_ref) => {
+                    if let Some(column_set) = column_set {
+                        column_set.insert(column_ref.column.index);
+                    }
+                    *expr = ScalarExpr::ConstantExpr(ConstantExpr {
+                        span: None,
+                        value: Scalar::Null,
+                    });
+                    Ok(())
+                }
+                ScalarExpr::FunctionCall(func) => {
+                    // If the function is `assume_not_null` or `remove_nullable`, we cannot replace
+                    // the column bindings with `Scalar::Null`.
+                    if matches!(
+                        func.func_name.as_str(),
+                        "assume_not_null" | "remove_nullable"
+                    ) {
+                        self.can_replace = false;
+                        return Ok(());
+                    }
+
+                    if func.func_name != "or" {
+                        for expr in &mut func.arguments {
+                            self.replace(expr, column_set)?;
+                        }
+                        return Ok(());
+                    }
+
+                    let mut children_columns_set = Some(ColumnSet::new());
+                    for expr in &mut func.arguments {
+                        self.replace(expr, &mut children_columns_set)?;
+                    }
+
+                    let mut has_left = false;
+                    let mut has_right = false;
+                    let children_columns_set = children_columns_set.unwrap();
+                    for column in children_columns_set.iter() {
+                        if self.left_output_columns.contains(column) {
+                            has_left = true;
+                        } else if self.right_output_columns.contains(column) {
+                            has_right = true;
+                        }
+                    }
+                    if has_left && has_right {
+                        self.can_replace = false;
+                        return Ok(());
+                    }
+
+                    if let Some(column_set) = column_set {
+                        *column_set = column_set.union(&children_columns_set).cloned().collect();
+                    }
+
+                    Ok(())
+                }
+                ScalarExpr::CastExpr(cast) => self.replace(&mut cast.argument, column_set),
+                ScalarExpr::ConstantExpr(_) => Ok(()),
+                _ => {
+                    self.can_replace = false;
+                    Ok(())
+                }
+            }
+        }
+    }
+
+    // Replace the column bindings of predicate with `Scalar::Null` and evaluate the result.
+    let mut replace = ReplaceColumnBindingsNull {
+        can_replace: true,
+        left_output_columns,
+        right_output_columns,
+    };
+    let mut null_scalar_expr = predicate.clone();
+    replace.replace(&mut null_scalar_expr, &mut None).unwrap();
+    if replace.can_replace {
+        let expr = convert_scalar_expr_to_expr(null_scalar_expr)?;
+        let func_ctx = &FunctionContext::default();
+        let data_block = DataBlock::empty();
+        let evaluator = Evaluator::new(&data_block, func_ctx, &BUILTIN_FUNCTIONS);
+        if let Value::Scalar(scalar) = evaluator.run(&expr)? {
+            // if null column can be filtered, return true.
+            if matches!(scalar, Scalar::Boolean(false) | Scalar::Null) {
+                return Ok(true);
+            }
+        }
+    }
+    Ok(false)
+}
+
+// Convert `ScalarExpr` to `Expr`.
+fn convert_scalar_expr_to_expr(scalar_expr: ScalarExpr) -> Result<Expr> {
+    let schema = Arc::new(DataSchema::new(vec![]));
+    let remote_expr = scalar_expr
+        .type_check(schema.as_ref())?
+        .project_column_ref(|index| schema.index_of(&index.to_string()).unwrap())
+        .as_remote_expr();
+    let expr = remote_expr.as_expr(&BUILTIN_FUNCTIONS);
+    cast_expr_to_non_null_boolean(expr)
+}
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_commute_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_commute_join.rs
index aec744bd1ce6..5b18bd3e37d0 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_commute_join.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_commute_join.rs
@@ -16,6 +16,7 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RelExpr;
@@ -23,7 +24,6 @@ use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Join;
 use crate::plans::JoinType;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 /// Rule to apply commutativity of join operator.
@@ -31,7 +31,7 @@ use crate::plans::RelOp;
 /// rule will help us measure which child is the better one.
 pub struct RuleCommuteJoin {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleCommuteJoin {
@@ -42,16 +42,10 @@ impl RuleCommuteJoin {
             // LogicalJoin
             // | \
             // *  *
-            patterns: vec![SExpr::create_binary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Join,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_pattern_leaf()),
-                Arc::new(SExpr::create_pattern_leaf()),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Join,
+                children: vec![Matcher::Leaf, Matcher::Leaf],
+            }],
         }
     }
 }
@@ -84,6 +78,7 @@ impl Rule for RuleCommuteJoin {
                     | JoinType::LeftAnti
                     | JoinType::RightAnti
                     | JoinType::LeftMark
+                    | JoinType::RightMark
             )
         } else if left_card == right_card {
             matches!(
@@ -109,7 +104,7 @@ impl Rule for RuleCommuteJoin {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_eval_scalar.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_eval_scalar.rs
index 95c89f07871a..6f0a3295c83d 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_eval_scalar.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_eval_scalar.rs
@@ -12,21 +12,19 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::Arc;
-
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::SExpr;
 use crate::plans::EvalScalar;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 pub struct RuleEliminateEvalScalar {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleEliminateEvalScalar {
@@ -36,20 +34,10 @@ impl RuleEliminateEvalScalar {
             // EvalScalar
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -70,7 +58,7 @@ impl Rule for RuleEliminateEvalScalar {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_filter.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_filter.rs
index f6c562c7fbd5..560d58203234 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_filter.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_filter.rs
@@ -17,18 +17,18 @@ use std::sync::Arc;
 use databend_common_exception::Result;
 use itertools::Itertools;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::SExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 
 pub struct RuleEliminateFilter {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleEliminateFilter {
@@ -38,20 +38,10 @@ impl RuleEliminateFilter {
             // Filter
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -103,7 +93,7 @@ impl Rule for RuleEliminateFilter {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_sort.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_sort.rs
index 61b4b1bbaed3..ecf580fa5631 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_sort.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_eliminate_sort.rs
@@ -12,22 +12,20 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::Arc;
-
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RelExpr;
 use crate::optimizer::SExpr;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::Sort;
 
 pub struct RuleEliminateSort {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleEliminateSort {
@@ -37,20 +35,10 @@ impl RuleEliminateSort {
             // Sort
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Sort,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Sort,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -75,7 +63,7 @@ impl Rule for RuleEliminateSort {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_fold_count_aggregate.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_fold_count_aggregate.rs
index 8874c0f6a478..c73f353bba36 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_fold_count_aggregate.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_fold_count_aggregate.rs
@@ -18,6 +18,7 @@ use databend_common_exception::Result;
 use databend_common_expression::types::NumberScalar;
 use databend_common_expression::Scalar;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
@@ -28,14 +29,13 @@ use crate::plans::AggregateMode;
 use crate::plans::ConstantExpr;
 use crate::plans::DummyTableScan;
 use crate::plans::EvalScalar;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 
 /// Fold simple `COUNT(*)` aggregate with statistics information.
 pub struct RuleFoldCountAggregate {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleFoldCountAggregate {
@@ -45,20 +45,10 @@ impl RuleFoldCountAggregate {
             //  Aggregate
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Aggregate,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Aggregate,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -127,7 +117,7 @@ impl Rule for RuleFoldCountAggregate {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_eval_scalar.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_eval_scalar.rs
index b624b737d4fa..aabcb55c8e09 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_eval_scalar.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_eval_scalar.rs
@@ -16,6 +16,7 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
@@ -23,13 +24,12 @@ use crate::optimizer::ColumnSet;
 use crate::optimizer::RelExpr;
 use crate::optimizer::SExpr;
 use crate::plans::EvalScalar;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 // Merge two adjacent `EvalScalar`s into one
 pub struct RuleMergeEvalScalar {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleMergeEvalScalar {
@@ -41,28 +41,13 @@ impl RuleMergeEvalScalar {
             //  EvalScalar
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -107,7 +92,7 @@ impl Rule for RuleMergeEvalScalar {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_filter.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_filter.rs
index 57c816ea863f..bb9ac746f260 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_filter.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_merge_filter.rs
@@ -16,18 +16,18 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::SExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 // Merge two adjacent `Filter`s into one
 pub struct RuleMergeFilter {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleMergeFilter {
@@ -39,28 +39,13 @@ impl RuleMergeFilter {
             //  Filter
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Filter,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Filter,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -89,7 +74,7 @@ impl Rule for RuleMergeFilter {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_aggregate.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_aggregate.rs
index 1e7eaf08c4a8..bebbd7089415 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_aggregate.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_aggregate.rs
@@ -16,6 +16,7 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
@@ -23,7 +24,6 @@ use crate::optimizer::SExpr;
 use crate::plans::Aggregate;
 use crate::plans::BoundColumnRef;
 use crate::plans::EvalScalar;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 use crate::plans::ScalarItem;
@@ -32,7 +32,7 @@ use crate::Visibility;
 
 pub struct RuleNormalizeAggregate {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleNormalizeAggregate {
@@ -42,20 +42,10 @@ impl RuleNormalizeAggregate {
             // Aggregate
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Aggregate,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Aggregate,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -195,7 +185,7 @@ impl Rule for RuleNormalizeAggregate {
         }
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_scalar.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_scalar.rs
index ec29f7d55715..981a6638ef55 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_scalar.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_normalize_scalar.rs
@@ -17,6 +17,7 @@ use std::sync::Arc;
 use databend_common_exception::Result;
 use databend_common_expression::Scalar;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::constant::is_falsy;
 use crate::optimizer::rule::constant::is_true;
 use crate::optimizer::rule::Rule;
@@ -24,7 +25,6 @@ use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::ConstantExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 
@@ -58,7 +58,7 @@ fn normalize_falsy_predicate(predicates: Vec<ScalarExpr>) -> Vec<ScalarExpr> {
 /// whole filter with FALSE
 pub struct RuleNormalizeScalarFilter {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleNormalizeScalarFilter {
@@ -68,20 +68,10 @@ impl RuleNormalizeScalarFilter {
             // Filter
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -114,7 +104,7 @@ impl Rule for RuleNormalizeScalarFilter {
         }
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_aggregate.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_aggregate.rs
index 258805a70598..599122e2a97d 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_aggregate.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_aggregate.rs
@@ -14,6 +14,7 @@
 
 use std::sync::Arc;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RelExpr;
@@ -21,9 +22,7 @@ use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Aggregate;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Pattern;
 
 /// Input:   Filter
 ///           \
@@ -48,32 +47,21 @@ use crate::plans::RelOp::Pattern;
 ///                *
 pub struct RulePushDownFilterAggregate {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterAggregate {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownFilterAggregate,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Aggregate,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Aggregate,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -140,7 +128,7 @@ impl Rule for RulePushDownFilterAggregate {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_eval_scalar.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_eval_scalar.rs
index c293d334d95a..431803d0a88d 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_eval_scalar.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_eval_scalar.rs
@@ -16,6 +16,7 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
@@ -29,7 +30,6 @@ use crate::plans::FunctionCall;
 use crate::plans::LagLeadFunction;
 use crate::plans::LambdaFunc;
 use crate::plans::NthValueFunction;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 use crate::plans::ScalarItem;
@@ -37,16 +37,14 @@ use crate::plans::UDFServerCall;
 use crate::plans::WindowFunc;
 use crate::plans::WindowFuncType;
 use crate::plans::WindowOrderBy;
-use crate::MetadataRef;
 
 pub struct RulePushDownFilterEvalScalar {
     id: RuleID,
-    patterns: Vec<SExpr>,
-    _metadata: MetadataRef,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterEvalScalar {
-    pub fn new(metadata: MetadataRef) -> Self {
+    pub fn new() -> Self {
         Self {
             id: RuleID::PushDownFilterEvalScalar,
             // Filter
@@ -54,29 +52,13 @@ impl RulePushDownFilterEvalScalar {
             //   EvalScalar
             //    \
             //     *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
-            _metadata: metadata,
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 
@@ -293,7 +275,7 @@ impl Rule for RulePushDownFilterEvalScalar {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_join.rs
index 32783bb75dc2..3bddc292c3a4 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_join.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_join.rs
@@ -17,12 +17,15 @@ use std::sync::Arc;
 use databend_common_exception::Result;
 
 use crate::binder::JoinPredicate;
+use crate::optimizer::extract::Matcher;
+use crate::optimizer::filter::InferFilterOptimizer;
+use crate::optimizer::filter::JoinProperty;
 use crate::optimizer::rule::constant::false_constant;
 use crate::optimizer::rule::constant::is_falsy;
-use crate::optimizer::rule::rewrite::filter_join::convert_mark_to_semi_join;
-use crate::optimizer::rule::rewrite::filter_join::outer_to_inner;
-use crate::optimizer::rule::rewrite::filter_join::rewrite_predicates;
-use crate::optimizer::rule::rewrite::filter_join::try_derive_predicates;
+use crate::optimizer::rule::rewrite::push_down_filter_join::can_filter_null;
+use crate::optimizer::rule::rewrite::push_down_filter_join::convert_mark_to_semi_join;
+use crate::optimizer::rule::rewrite::push_down_filter_join::outer_join_to_inner_join;
+use crate::optimizer::rule::rewrite::push_down_filter_join::rewrite_predicates;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RelExpr;
@@ -30,58 +33,35 @@ use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::ComparisonOp;
 use crate::plans::Filter;
+use crate::plans::FunctionCall;
 use crate::plans::Join;
 use crate::plans::JoinType;
-use crate::plans::PatternPlan;
+use crate::plans::Operator;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
-use crate::MetadataRef;
 
 pub struct RulePushDownFilterJoin {
     id: RuleID,
-    patterns: Vec<SExpr>,
-    _metadata: MetadataRef,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterJoin {
-    pub fn new(metadata: MetadataRef) -> Self {
+    pub fn new() -> Self {
         Self {
             id: RuleID::PushDownFilterJoin,
             // Filter
             //  \
-            //   InnerJoin
+            //   Join
             //   | \
             //   |  *
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_binary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Join,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
-            _metadata: metadata,
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Join,
+                    children: vec![Matcher::Leaf, Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -93,38 +73,44 @@ impl Rule for RulePushDownFilterJoin {
 
     fn apply(&self, s_expr: &SExpr, state: &mut TransformResult) -> Result<()> {
         // First, try to convert outer join to inner join
-        let mut s_expr = outer_to_inner(s_expr)?;
+        let (s_expr, outer_to_inner) = outer_join_to_inner_join(s_expr)?;
+
         // Second, check if can convert mark join to semi join
-        s_expr = convert_mark_to_semi_join(&s_expr)?;
+        let (s_expr, mark_to_semi) = convert_mark_to_semi_join(&s_expr)?;
+        if s_expr.plan().rel_op() != RelOp::Filter {
+            state.add_result(s_expr);
+            return Ok(());
+        }
         let filter: Filter = s_expr.plan().clone().try_into()?;
         if filter.predicates.is_empty() {
             state.add_result(s_expr);
             return Ok(());
         }
-        // Finally, extract or predicates from Filter to push down them to join.
-        // For example: `select * from t1, t2 where (t1.a=1 and t2.b=2) or (t1.a=2 and t2.b=1)`
-        // The predicate will be rewritten to `((t1.a=1 and t2.b=2) or (t1.a=2 and t2.b=1)) and (t1.a=1 or t1.a=2) and (t2.b=2 or t2.b=1)`
-        // So `(t1.a=1 or t1.a=1), (t2.b=2 or t2.b=1)` may be pushed down join and reduce rows between join
-        let predicates = rewrite_predicates(&s_expr)?;
-        let (need_push, mut result) = try_push_down_filter_join(&s_expr, predicates)?;
-        if !need_push {
+
+        // Finally, push down filter to join.
+        let (need_push, mut result) = try_push_down_filter_join(&s_expr)?;
+        if !need_push && !outer_to_inner && !mark_to_semi {
             return Ok(());
         }
+
         result.set_applied_rule(&self.id);
         state.add_result(result);
 
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
 
-pub fn try_push_down_filter_join(
-    s_expr: &SExpr,
-    predicates: Vec<ScalarExpr>,
-) -> Result<(bool, SExpr)> {
+pub fn try_push_down_filter_join(s_expr: &SExpr) -> Result<(bool, SExpr)> {
+    // Extract or predicates from Filter to push down them to join.
+    // For example: `select * from t1, t2 where (t1.a=1 and t2.b=2) or (t1.a=2 and t2.b=1)`
+    // The predicate will be rewritten to `((t1.a=1 and t2.b=2) or (t1.a=2 and t2.b=1)) and (t1.a=1 or t1.a=2) and (t2.b=2 or t2.b=1)`
+    // So `(t1.a=1 or t1.a=1), (t2.b=2 or t2.b=1)` may be pushed down join and reduce rows between join
+    let predicates = rewrite_predicates(s_expr)?;
+
     let join_expr = s_expr.child(0)?;
     let mut join: Join = join_expr.plan().clone().try_into()?;
 
@@ -132,62 +118,67 @@ pub fn try_push_down_filter_join(
     let left_prop = rel_expr.derive_relational_prop_child(0)?;
     let right_prop = rel_expr.derive_relational_prop_child(1)?;
 
+    let original_predicates_count = predicates.len();
+    let mut original_predicates = vec![];
     let mut left_push_down = vec![];
     let mut right_push_down = vec![];
-    let mut original_predicates = vec![];
-
-    let mut need_push = false;
-
+    let mut push_down_predicates = vec![];
+    let mut non_equi_predicates = vec![];
     for predicate in predicates.into_iter() {
         if is_falsy(&predicate) {
-            left_push_down = vec![false_constant()];
-            right_push_down = vec![false_constant()];
-            need_push = true;
+            push_down_predicates = vec![false_constant()];
             break;
         }
         let pred = JoinPredicate::new(&predicate, &left_prop, &right_prop);
         match pred {
             JoinPredicate::ALL(_) => {
-                need_push = true;
-                left_push_down.push(predicate.clone());
-                right_push_down.push(predicate.clone());
+                push_down_predicates.push(predicate);
             }
             JoinPredicate::Left(_) => {
                 if matches!(
                     join.join_type,
                     JoinType::Right | JoinType::RightSingle | JoinType::Full
                 ) {
-                    original_predicates.push(predicate);
-                    continue;
+                    if can_filter_null(
+                        &predicate,
+                        &left_prop.output_columns,
+                        &right_prop.output_columns,
+                    )? {
+                        left_push_down.push(predicate);
+                    } else {
+                        original_predicates.push(predicate);
+                    }
+                } else {
+                    left_push_down.push(predicate);
                 }
-                need_push = true;
-                left_push_down.push(predicate);
             }
             JoinPredicate::Right(_) => {
                 if matches!(
                     join.join_type,
                     JoinType::Left | JoinType::LeftSingle | JoinType::Full
                 ) {
-                    original_predicates.push(predicate);
-                    continue;
+                    if can_filter_null(
+                        &predicate,
+                        &left_prop.output_columns,
+                        &right_prop.output_columns,
+                    )? {
+                        right_push_down.push(predicate);
+                    } else {
+                        original_predicates.push(predicate);
+                    }
+                } else {
+                    right_push_down.push(predicate);
                 }
-                need_push = true;
-                right_push_down.push(predicate);
             }
             JoinPredicate::Other(_) => original_predicates.push(predicate),
-
-            JoinPredicate::Both { left, right, op } => {
-                if op == ComparisonOp::Equal {
-                    if matches!(join.join_type, JoinType::Inner | JoinType::Cross) {
-                        join.join_type = JoinType::Inner;
-                        join.left_conditions.push(left.clone());
-                        join.right_conditions.push(right.clone());
-                        need_push = true;
+            JoinPredicate::Both { is_equal_op, .. } => {
+                if matches!(join.join_type, JoinType::Inner | JoinType::Cross) {
+                    if is_equal_op {
+                        push_down_predicates.push(predicate);
+                    } else {
+                        non_equi_predicates.push(predicate);
                     }
-                } else if matches!(join.join_type, JoinType::Inner | JoinType::Cross) {
                     join.join_type = JoinType::Inner;
-                    join.non_equi_conditions.push(predicate.clone());
-                    need_push = true;
                 } else {
                     original_predicates.push(predicate);
                 }
@@ -195,12 +186,110 @@ pub fn try_push_down_filter_join(
         }
     }
 
-    if !need_push {
+    if original_predicates.len() == original_predicates_count {
         return Ok((false, s_expr.clone()));
     }
 
-    // try to derive new predicate and push down filter
-    let mut result = try_derive_predicates(s_expr, join, left_push_down, right_push_down)?;
+    if !matches!(join.join_type, JoinType::Full) {
+        // Infer new predicate and push down filter.
+        for (left_condition, right_condition) in join
+            .left_conditions
+            .iter()
+            .zip(join.right_conditions.iter())
+        {
+            push_down_predicates.push(ScalarExpr::FunctionCall(FunctionCall {
+                span: None,
+                func_name: String::from(ComparisonOp::Equal.to_func_name()),
+                params: vec![],
+                arguments: vec![left_condition.clone(), right_condition.clone()],
+            }));
+        }
+        join.left_conditions.clear();
+        join.right_conditions.clear();
+        match join.join_type {
+            JoinType::Left | JoinType::LeftSingle => {
+                push_down_predicates.extend(left_push_down);
+                left_push_down = vec![];
+            }
+            JoinType::Right | JoinType::RightSingle => {
+                push_down_predicates.extend(right_push_down);
+                right_push_down = vec![];
+            }
+            _ => {
+                push_down_predicates.extend(left_push_down);
+                left_push_down = vec![];
+                push_down_predicates.extend(right_push_down);
+                right_push_down = vec![];
+            }
+        }
+        let join_prop = JoinProperty::new(&left_prop.output_columns, &right_prop.output_columns);
+        let infer_filter = InferFilterOptimizer::new(Some(join_prop));
+        push_down_predicates = infer_filter.run(push_down_predicates)?;
+    }
+
+    let mut all_push_down = vec![];
+    for predicate in push_down_predicates.into_iter() {
+        if is_falsy(&predicate) {
+            left_push_down = vec![false_constant()];
+            right_push_down = vec![false_constant()];
+            break;
+        }
+        let pred = JoinPredicate::new(&predicate, &left_prop, &right_prop);
+        match pred {
+            JoinPredicate::ALL(_) => {
+                all_push_down.push(predicate);
+            }
+            JoinPredicate::Left(_) => {
+                left_push_down.push(predicate);
+            }
+            JoinPredicate::Right(_) => {
+                right_push_down.push(predicate);
+            }
+            JoinPredicate::Both { left, right, .. } => {
+                join.left_conditions.push(left.clone());
+                join.right_conditions.push(right.clone());
+            }
+            _ => original_predicates.push(predicate),
+        }
+    }
+    join.non_equi_conditions.extend(non_equi_predicates);
+    if !all_push_down.is_empty() {
+        left_push_down.extend(all_push_down.to_vec());
+        right_push_down.extend(all_push_down);
+    }
+
+    let mut left_child = join_expr.child(0)?.clone();
+    let mut right_child = join_expr.child(1)?.clone();
+
+    if !left_push_down.is_empty() {
+        left_child = SExpr::create_unary(
+            Arc::new(
+                Filter {
+                    predicates: left_push_down,
+                }
+                .into(),
+            ),
+            Arc::new(left_child),
+        );
+    }
+
+    if !right_push_down.is_empty() {
+        right_child = SExpr::create_unary(
+            Arc::new(
+                Filter {
+                    predicates: right_push_down,
+                }
+                .into(),
+            ),
+            Arc::new(right_child),
+        );
+    }
+
+    let mut result = SExpr::create_binary(
+        Arc::new(join.into()),
+        Arc::new(left_child),
+        Arc::new(right_child),
+    );
 
     if !original_predicates.is_empty() {
         result = SExpr::create_unary(
@@ -213,5 +302,6 @@ pub fn try_push_down_filter_join(
             Arc::new(result),
         );
     }
-    Ok((need_push, result))
+
+    Ok((true, result))
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_project_set.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_project_set.rs
index 8c492d73df96..d826d529a250 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_project_set.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_project_set.rs
@@ -16,16 +16,15 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RelExpr;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::ProjectSet;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Pattern;
 
 /// Input:   Filter
 ///           \
@@ -50,32 +49,20 @@ use crate::plans::RelOp::Pattern;
 ///                *
 pub struct RulePushDownFilterProjectSet {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterProjectSet {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownFilterProjectSet,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::ProjectSet,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::ProjectSet,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -134,7 +121,7 @@ impl Rule for RulePushDownFilterProjectSet {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_scan.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_scan.rs
index 88c5e540ede4..9ca0e1d6505d 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_scan.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_scan.rs
@@ -17,6 +17,7 @@ use std::sync::Arc;
 use databend_common_exception::Result;
 
 use crate::binder::ColumnBindingBuilder;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
@@ -29,7 +30,6 @@ use crate::plans::FunctionCall;
 use crate::plans::LagLeadFunction;
 use crate::plans::LambdaFunc;
 use crate::plans::NthValueFunction;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::Scan;
 use crate::plans::UDFServerCall;
@@ -43,7 +43,7 @@ use crate::TableEntry;
 
 pub struct RulePushDownFilterScan {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
     metadata: MetadataRef,
 }
 
@@ -54,20 +54,13 @@ impl RulePushDownFilterScan {
             // Filter
             //  \
             //   LogicalGet
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Scan,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Scan,
+                    children: vec![],
+                }],
+            }],
             metadata,
         }
     }
@@ -409,7 +402,7 @@ impl Rule for RulePushDownFilterScan {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_sort.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_sort.rs
index b1e03e266b38..19bcc2b34fba 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_sort.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_sort.rs
@@ -16,14 +16,13 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Pattern;
 use crate::plans::Sort;
 
 /// Input:  Filter
@@ -39,32 +38,20 @@ use crate::plans::Sort;
 ///              *
 pub struct RulePushDownFilterSort {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterSort {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownFilterSort,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Sort,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Sort,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -91,7 +78,7 @@ impl Rule for RulePushDownFilterSort {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_union.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_union.rs
index afccc3e11ad8..ae9f77c8b0fa 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_union.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_union.rs
@@ -18,13 +18,13 @@ use ahash::HashMap;
 use databend_common_exception::Result;
 
 use crate::binder::ColumnBindingBuilder;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::BoundColumnRef;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
 use crate::plans::UnionAll;
@@ -40,7 +40,7 @@ use crate::Visibility;
 // So it'll be efficient to push down `filter` to `union`, reduce the size of data to pull from table.
 pub struct RulePushDownFilterUnion {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterUnion {
@@ -52,34 +52,13 @@ impl RulePushDownFilterUnion {
             //   UnionAll
             //     /  \
             //   ...   ...
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_binary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::UnionAll,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::UnionAll,
+                    children: vec![Matcher::Leaf, Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -124,8 +103,8 @@ impl Rule for RulePushDownFilterUnion {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
 
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_window.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_window.rs
index 6c75e2134f09..0e8fceb77474 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_window.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_filter_window.rs
@@ -14,14 +14,13 @@
 
 use std::sync::Arc;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Pattern;
 use crate::plans::Window;
 
 /// Input:   Filter
@@ -49,32 +48,20 @@ use crate::plans::Window;
 /// note that only push down filter used in `Window.partition_by` columns
 pub struct RulePushDownFilterWindow {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownFilterWindow {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownFilterWindow,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Window,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Window,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -140,7 +127,7 @@ impl Rule for RulePushDownFilterWindow {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_aggregate.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_aggregate.rs
index 7cb786c6378b..d73fc600f3de 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_aggregate.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_aggregate.rs
@@ -15,16 +15,14 @@
 use std::cmp;
 use std::sync::Arc;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Aggregate;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Aggregate as OpAggregate;
-use crate::plans::RelOp::Pattern;
 use crate::plans::RelOperator;
 
 /// Input:  Limit
@@ -40,32 +38,20 @@ use crate::plans::RelOperator;
 ///               *
 pub struct RulePushDownLimitAggregate {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownLimitAggregate {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownLimitAggregate,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Limit,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: OpAggregate,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Aggregate,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -99,7 +85,7 @@ impl Rule for RulePushDownLimitAggregate {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_expression.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_expression.rs
index 9b53d386bd1a..c86354e5ace0 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_expression.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_expression.rs
@@ -14,15 +14,14 @@
 
 use std::sync::Arc;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::EvalScalar;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Pattern;
 use crate::plans::RelOperator;
 
 /// Input:  Limit
@@ -36,39 +35,27 @@ use crate::plans::RelOperator;
 ///          limit
 ///             \
 ///               *
-pub struct RulePushDownLimitExpression {
+pub struct RulePushDownLimitEvalScalar {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
-impl RulePushDownLimitExpression {
+impl RulePushDownLimitEvalScalar {
     pub fn new() -> Self {
         Self {
-            id: RuleID::PushDownLimitExpression,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Limit,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+            id: RuleID::PushDownLimitEvalScalar,
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
 
-impl Rule for RulePushDownLimitExpression {
+impl Rule for RulePushDownLimitEvalScalar {
     fn id(&self) -> RuleID {
         self.id
     }
@@ -96,7 +83,7 @@ impl Rule for RulePushDownLimitExpression {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_join.rs
index 104ff48c1b1f..b2d126d6a212 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_join.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_join.rs
@@ -14,6 +14,7 @@
 
 use std::sync::Arc;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
@@ -21,7 +22,6 @@ use crate::optimizer::SExpr;
 use crate::plans::Join;
 use crate::plans::JoinType;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 
@@ -41,41 +41,20 @@ use crate::plans::RelOperator;
 ///          *
 pub struct RulePushDownLimitOuterJoin {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownLimitOuterJoin {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownLimitOuterJoin,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Limit,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_binary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Join,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Join,
+                    children: vec![Matcher::Leaf, Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -139,7 +118,7 @@ impl Rule for RulePushDownLimitOuterJoin {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_scan.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_scan.rs
index 5b6705bcdd74..693558918fa4 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_scan.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_scan.rs
@@ -17,12 +17,12 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::plans::Scan;
@@ -35,30 +35,22 @@ use crate::plans::Scan;
 ///         Limit
 ///           \
 ///           LogicalGet(padding limit)
-
 pub struct RulePushDownLimitScan {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownLimitScan {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownLimitScan,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Limit,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Scan,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Scan,
+                    children: vec![],
+                }],
+            }],
         }
     }
 }
@@ -84,7 +76,7 @@ impl Rule for RulePushDownLimitScan {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_sort.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_sort.rs
index 03011eb1c053..7a06c87c7ca7 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_sort.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_sort.rs
@@ -15,17 +15,15 @@
 use std::cmp;
 use std::sync::Arc;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
-use crate::plans::RelOp::Pattern;
-use crate::plans::RelOp::Sort;
 use crate::plans::RelOperator;
-use crate::plans::Sort as logsort;
+use crate::plans::Sort;
 
 /// Input:  Limit
 ///           \
@@ -40,27 +38,20 @@ use crate::plans::Sort as logsort;
 ///               *
 pub struct RulePushDownLimitSort {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownLimitSort {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownLimitSort,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Limit,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(PatternPlan { plan_type: Sort }.into()),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan { plan_type: Pattern }.into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Sort,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -79,7 +70,7 @@ impl Rule for RulePushDownLimitSort {
         if let Some(mut count) = limit.limit {
             count += limit.offset;
             let sort = s_expr.child(0)?;
-            let mut sort_limit: logsort = sort.plan().clone().try_into()?;
+            let mut sort_limit: Sort = sort.plan().clone().try_into()?;
             sort_limit.limit = Some(sort_limit.limit.map_or(count, |c| cmp::max(c, count)));
             let sort = SExpr::create_unary(
                 Arc::new(RelOperator::Sort(sort_limit)),
@@ -93,7 +84,7 @@ impl Rule for RulePushDownLimitSort {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_union.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_union.rs
index d4a161c31bee..5ee5e801843e 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_union.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_union.rs
@@ -17,18 +17,18 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Limit;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::UnionAll;
 
 pub struct RulePushDownLimitUnion {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownLimitUnion {
@@ -40,34 +40,13 @@ impl RulePushDownLimitUnion {
             //   UnionAll
             //     /  \
             //   ...   ...
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Limit,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_binary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::UnionAll,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Pattern,
-                        }
-                        .into(),
-                    ))),
-                )),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::UnionAll,
+                    children: vec![Matcher::Leaf, Matcher::Leaf],
+                }],
+            }],
         }
     }
 }
@@ -121,7 +100,7 @@ impl Rule for RulePushDownLimitUnion {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_window.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_window.rs
new file mode 100644
index 000000000000..3a65e684d0f6
--- /dev/null
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_limit_window.rs
@@ -0,0 +1,125 @@
+// Copyright 2021 Datafuse Labs
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+use std::cmp;
+use std::sync::Arc;
+
+use databend_common_exception::Result;
+
+use crate::optimizer::extract::Matcher;
+use crate::optimizer::rule::Rule;
+use crate::optimizer::rule::TransformResult;
+use crate::optimizer::RuleID;
+use crate::optimizer::SExpr;
+use crate::plans::Limit;
+use crate::plans::RelOp;
+use crate::plans::RelOperator;
+use crate::plans::Window as LogicalWindow;
+use crate::plans::WindowFuncFrame;
+use crate::plans::WindowFuncFrameBound;
+use crate::plans::WindowFuncFrameUnits;
+use crate::plans::WindowFuncType;
+
+/// Input:  Limit
+///           \
+///          Window
+///             \
+///              *
+///
+/// Output: Limit
+///           \
+///          Window(padding limit)
+///             \
+///               *
+pub struct RulePushDownLimitWindow {
+    id: RuleID,
+    matchers: Vec<Matcher>,
+}
+
+impl RulePushDownLimitWindow {
+    pub fn new() -> Self {
+        Self {
+            id: RuleID::PushDownLimitSort,
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Limit,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Window,
+                    children: vec![Matcher::Leaf],
+                }],
+            }],
+        }
+    }
+}
+
+impl Rule for RulePushDownLimitWindow {
+    fn id(&self) -> RuleID {
+        self.id
+    }
+
+    fn apply(&self, s_expr: &SExpr, state: &mut TransformResult) -> Result<()> {
+        let limit: Limit = s_expr.plan().clone().try_into()?;
+        if let Some(mut count) = limit.limit {
+            count += limit.offset;
+            let window = s_expr.child(0)?;
+            let mut window_limit: LogicalWindow = window.plan().clone().try_into()?;
+            if should_apply(window.child(0)?, &window_limit)? {
+                window_limit.limit = Some(window_limit.limit.map_or(count, |c| cmp::max(c, count)));
+                let sort = SExpr::create_unary(
+                    Arc::new(RelOperator::Window(window_limit)),
+                    Arc::new(window.child(0)?.clone()),
+                );
+
+                let mut result = s_expr.replace_children(vec![Arc::new(sort)]);
+                result.set_applied_rule(&self.id);
+                state.add_result(result);
+            }
+        }
+        Ok(())
+    }
+
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
+    }
+}
+
+fn should_apply(child: &SExpr, window: &LogicalWindow) -> Result<bool> {
+    let child_window_exists = child_has_window(child)?;
+    // ranking functions are frame insensitive
+    if is_ranking_function(&window.function) {
+        Ok(!child_window_exists)
+    } else {
+        Ok(is_valid_frame(&window.frame) && !child_window_exists)
+    }
+}
+
+fn is_ranking_function(func: &WindowFuncType) -> bool {
+    matches!(
+        func,
+        WindowFuncType::RowNumber | WindowFuncType::Rank | WindowFuncType::DenseRank
+    )
+}
+
+fn is_valid_frame(frame: &WindowFuncFrame) -> bool {
+    matches!(frame.units, WindowFuncFrameUnits::Rows)
+        && matches!(frame.start_bound, WindowFuncFrameBound::Preceding(_))
+        && matches!(frame.end_bound, WindowFuncFrameBound::CurrentRow)
+}
+
+fn child_has_window(child: &SExpr) -> Result<bool> {
+    match child.plan() {
+        RelOperator::Window(_) => Ok(true),
+        RelOperator::Scan(_) => Ok(false), // finish recursion
+        _ => child_has_window(child.child(0)?),
+    }
+}
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_prewhere.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_prewhere.rs
index a6f80c599e21..00eb07d22445 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_prewhere.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_prewhere.rs
@@ -18,12 +18,12 @@ use databend_common_exception::ErrorCode;
 use databend_common_exception::Result;
 use databend_common_expression::TableSchemaRef;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::ColumnSet;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Filter;
-use crate::plans::PatternPlan;
 use crate::plans::Prewhere;
 use crate::plans::RelOp;
 use crate::plans::ScalarExpr;
@@ -34,7 +34,7 @@ use crate::Visibility;
 
 pub struct RulePushDownPrewhere {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
     metadata: MetadataRef,
 }
 
@@ -42,20 +42,13 @@ impl RulePushDownPrewhere {
     pub fn new(metadata: MetadataRef) -> Self {
         Self {
             id: RuleID::PushDownPrewhere,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Filter,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Scan,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Filter,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Scan,
+                    children: vec![],
+                }],
+            }],
             metadata,
         }
     }
@@ -164,10 +157,6 @@ impl Rule for RulePushDownPrewhere {
         self.id
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
-    }
-
     fn apply(
         &self,
         s_expr: &SExpr,
@@ -178,4 +167,8 @@ impl Rule for RulePushDownPrewhere {
         state.add_result(result);
         Ok(())
     }
+
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
+    }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_sort_scan.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_sort_scan.rs
index 8348ce13979d..ecb340f56421 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_sort_scan.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_push_down_sort_scan.rs
@@ -17,11 +17,11 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::plans::Scan;
@@ -35,30 +35,22 @@ use crate::plans::Sort;
 ///         Sort
 ///           \
 ///           LogicalGet(padding order_by and limit)
-
 pub struct RulePushDownSortScan {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RulePushDownSortScan {
     pub fn new() -> Self {
         Self {
             id: RuleID::PushDownSortScan,
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Sort,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Scan,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Sort,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Scan,
+                    children: vec![],
+                }],
+            }],
         }
     }
 }
@@ -86,7 +78,7 @@ impl Rule for RulePushDownSortScan {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_semi_to_inner_join.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_semi_to_inner_join.rs
index 3f413823e05f..0c00281a1223 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_semi_to_inner_join.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_semi_to_inner_join.rs
@@ -17,50 +17,34 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
 use crate::plans::Join;
 use crate::plans::JoinType;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::IndexType;
-use crate::MetadataRef;
 use crate::ScalarExpr;
 
 pub struct RuleSemiToInnerJoin {
     id: RuleID,
-    patterns: Vec<SExpr>,
-    _metadata: MetadataRef,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleSemiToInnerJoin {
-    pub fn new(_metadata: MetadataRef) -> Self {
+    pub fn new() -> Self {
         Self {
             id: RuleID::SemiToInnerJoin,
-            patterns: vec![SExpr::create_binary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Join,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
-            _metadata,
+            // Join
+            // |  \
+            // *   *
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Join,
+                children: vec![Matcher::Leaf, Matcher::Leaf],
+            }],
         }
     }
 }
@@ -116,8 +100,8 @@ impl Rule for RuleSemiToInnerJoin {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
 
@@ -145,8 +129,7 @@ fn find_group_by_keys(child: &SExpr, group_by_keys: &mut HashSet<IndexType>) ->
         | RelOperator::Udf(_)
         | RelOperator::Scan(_)
         | RelOperator::CteScan(_)
-        | RelOperator::Join(_)
-        | RelOperator::Pattern(_) => {}
+        | RelOperator::Join(_) => {}
     }
     Ok(())
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_split_aggregate.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_split_aggregate.rs
index d29dfb7ad7f3..61b12b66aed9 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_split_aggregate.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_split_aggregate.rs
@@ -16,19 +16,19 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::SExpr;
 use crate::plans::Aggregate;
 use crate::plans::AggregateMode;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 // Split `Aggregate` into `FinalAggregate` and `PartialAggregate`
 pub struct RuleSplitAggregate {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleSplitAggregate {
@@ -38,20 +38,10 @@ impl RuleSplitAggregate {
             //  Aggregate
             //  \
             //   *
-            patterns: vec![SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Aggregate,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Pattern,
-                    }
-                    .into(),
-                ))),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Aggregate,
+                children: vec![Matcher::Leaf],
+            }],
         }
     }
 }
@@ -81,7 +71,7 @@ impl Rule for RuleSplitAggregate {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 }
diff --git a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_try_apply_agg_index.rs b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_try_apply_agg_index.rs
index 0a3040eaccb0..f0cb62841e10 100644
--- a/src/query/sql/src/planner/optimizer/rule/rewrite/rule_try_apply_agg_index.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rewrite/rule_try_apply_agg_index.rs
@@ -12,15 +12,13 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
-use std::sync::Arc;
-
 use databend_common_exception::Result;
 
 use super::agg_index;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::IndexType;
@@ -30,39 +28,27 @@ pub struct RuleTryApplyAggIndex {
     id: RuleID,
     metadata: MetadataRef,
 
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleTryApplyAggIndex {
-    fn sorted_patterns() -> Vec<SExpr> {
+    fn sorted_matchers() -> Vec<Matcher> {
         vec![
             // Expression
             //     |
             //    Sort
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Sort,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Scan,
-                        }
-                        .into(),
-                    ))),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Sort,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Scan,
+                        children: vec![],
+                    }],
+                }],
+            },
             // Expression
             //     |
             //    Sort
@@ -70,36 +56,19 @@ impl RuleTryApplyAggIndex {
             //   Filter
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Sort,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Filter,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_leaf(Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Scan,
-                            }
-                            .into(),
-                        ))),
-                    )),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Sort,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Filter,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Scan,
+                            children: vec![],
+                        }],
+                    }],
+                }],
+            },
             // Expression
             //     |
             //    Sort
@@ -109,52 +78,25 @@ impl RuleTryApplyAggIndex {
             // Expression
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Sort,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Aggregate,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Aggregate,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_unary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::EvalScalar,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_leaf(Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::Scan,
-                                    }
-                                    .into(),
-                                ))),
-                            )),
-                        )),
-                    )),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Sort,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Aggregate,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Aggregate,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::EvalScalar,
+                                children: vec![Matcher::MatchOp {
+                                    op_type: RelOp::Scan,
+                                    children: vec![],
+                                }],
+                            }],
+                        }],
+                    }],
+                }],
+            },
             // Expression
             //     |
             //    Sort
@@ -166,109 +108,58 @@ impl RuleTryApplyAggIndex {
             //   Filter
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Sort,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Aggregate,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Aggregate,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_unary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::EvalScalar,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_unary(
-                                    Arc::new(
-                                        PatternPlan {
-                                            plan_type: RelOp::Filter,
-                                        }
-                                        .into(),
-                                    ),
-                                    Arc::new(SExpr::create_leaf(Arc::new(
-                                        PatternPlan {
-                                            plan_type: RelOp::Scan,
-                                        }
-                                        .into(),
-                                    ))),
-                                )),
-                            )),
-                        )),
-                    )),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Sort,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Aggregate,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Aggregate,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::EvalScalar,
+                                children: vec![Matcher::MatchOp {
+                                    op_type: RelOp::Filter,
+                                    children: vec![Matcher::MatchOp {
+                                        op_type: RelOp::Scan,
+                                        children: vec![],
+                                    }],
+                                }],
+                            }],
+                        }],
+                    }],
+                }],
+            },
         ]
     }
 
-    fn normal_patterns() -> Vec<SExpr> {
+    fn normal_matchers() -> Vec<Matcher> {
         vec![
             // Expression
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_leaf(Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Scan,
-                    }
-                    .into(),
-                ))),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Scan,
+                    children: vec![],
+                }],
+            },
             // Expression
             //     |
             //   Filter
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Filter,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_leaf(Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Scan,
-                        }
-                        .into(),
-                    ))),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Filter,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Scan,
+                        children: vec![],
+                    }],
+                }],
+            },
             // Expression
             //     |
             // Aggregation
@@ -276,44 +167,22 @@ impl RuleTryApplyAggIndex {
             // Expression
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Aggregate,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Aggregate,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::EvalScalar,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_leaf(Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Scan,
-                                }
-                                .into(),
-                            ))),
-                        )),
-                    )),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Aggregate,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Aggregate,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::EvalScalar,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::Scan,
+                                children: vec![],
+                            }],
+                        }],
+                    }],
+                }],
+            },
             // Expression
             //     |
             // Aggregation
@@ -323,65 +192,39 @@ impl RuleTryApplyAggIndex {
             //   Filter
             //     |
             //    Scan
-            SExpr::create_unary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::EvalScalar,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Aggregate,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Aggregate,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::EvalScalar,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_unary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::Filter,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_leaf(Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::Scan,
-                                    }
-                                    .into(),
-                                ))),
-                            )),
-                        )),
-                    )),
-                )),
-            ),
+            Matcher::MatchOp {
+                op_type: RelOp::EvalScalar,
+                children: vec![Matcher::MatchOp {
+                    op_type: RelOp::Aggregate,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Aggregate,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::EvalScalar,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::Filter,
+                                children: vec![Matcher::MatchOp {
+                                    op_type: RelOp::Scan,
+                                    children: vec![],
+                                }],
+                            }],
+                        }],
+                    }],
+                }],
+            },
         ]
     }
 
-    fn patterns() -> Vec<SExpr> {
-        let mut patterns = Self::normal_patterns();
-        patterns.extend(Self::sorted_patterns());
+    fn matchers() -> Vec<Matcher> {
+        let mut patterns = Self::normal_matchers();
+        patterns.extend(Self::sorted_matchers());
         patterns
     }
+
     pub fn new(metadata: MetadataRef) -> Self {
         Self {
             id: RuleID::TryApplyAggIndex,
             metadata,
-            patterns: Self::patterns(),
+            matchers: Self::matchers(),
         }
     }
 }
@@ -391,10 +234,6 @@ impl Rule for RuleTryApplyAggIndex {
         self.id
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
-    }
-
     fn apply(
         &self,
         s_expr: &SExpr,
@@ -424,6 +263,10 @@ impl Rule for RuleTryApplyAggIndex {
 
         Ok(())
     }
+
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
+    }
 }
 
 impl RuleTryApplyAggIndex {
diff --git a/src/query/sql/src/planner/optimizer/rule/rule.rs b/src/query/sql/src/planner/optimizer/rule/rule.rs
index a96a1d212164..30d98f42cd34 100644
--- a/src/query/sql/src/planner/optimizer/rule/rule.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rule.rs
@@ -20,25 +20,25 @@ use databend_common_exception::Result;
 use num_derive::FromPrimitive;
 use num_derive::ToPrimitive;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::SExpr;
 
 pub static DEFAULT_REWRITE_RULES: LazyLock<Vec<RuleID>> = LazyLock::new(|| {
     vec![
-        RuleID::NormalizeDisjunctiveFilter,
         RuleID::NormalizeScalarFilter,
         RuleID::NormalizeAggregate,
         RuleID::EliminateFilter,
         RuleID::EliminateSort,
         RuleID::MergeFilter,
-        RuleID::InferFilter,
         RuleID::MergeEvalScalar,
         RuleID::PushDownFilterUnion,
         RuleID::PushDownFilterAggregate,
         RuleID::PushDownFilterWindow,
         RuleID::PushDownLimitUnion,
-        RuleID::PushDownLimitExpression,
+        RuleID::PushDownLimitEvalScalar,
         RuleID::PushDownLimitSort,
+        RuleID::PushDownLimitWindow,
         RuleID::PushDownLimitAggregate,
         RuleID::PushDownLimitOuterJoin,
         RuleID::PushDownLimitScan,
@@ -63,7 +63,7 @@ pub trait Rule {
 
     fn apply(&self, s_expr: &SExpr, state: &mut TransformResult) -> Result<()>;
 
-    fn patterns(&self) -> &Vec<SExpr>;
+    fn matchers(&self) -> &[Matcher];
 
     fn transformation(&self) -> bool {
         true
@@ -77,8 +77,6 @@ pub enum RuleID {
     // Rewrite rules
     NormalizeAggregate,
     NormalizeScalarFilter,
-    NormalizeDisjunctiveFilter,
-    InferFilter,
     PushDownFilterAggregate,
     PushDownFilterEvalScalar,
     PushDownFilterUnion,
@@ -89,8 +87,9 @@ pub enum RuleID {
     PushDownFilterWindow,
     PushDownLimitUnion,
     PushDownLimitOuterJoin,
-    PushDownLimitExpression,
+    PushDownLimitEvalScalar,
     PushDownLimitSort,
+    PushDownLimitWindow,
     PushDownLimitAggregate,
     PushDownLimitScan,
     PushDownSortScan,
@@ -123,12 +122,13 @@ impl Display for RuleID {
             RuleID::PushDownFilterProjectSet => write!(f, "PushDownFilterProjectSet"),
             RuleID::PushDownLimitUnion => write!(f, "PushDownLimitUnion"),
             RuleID::PushDownLimitOuterJoin => write!(f, "PushDownLimitOuterJoin"),
-            RuleID::PushDownLimitExpression => write!(f, "PushDownLimitExpression"),
+            RuleID::PushDownLimitEvalScalar => write!(f, "PushDownLimitEvalScalar"),
             RuleID::PushDownLimitSort => write!(f, "PushDownLimitSort"),
             RuleID::PushDownLimitAggregate => write!(f, "PushDownLimitAggregate"),
             RuleID::PushDownFilterAggregate => write!(f, "PushDownFilterAggregate"),
             RuleID::PushDownLimitScan => write!(f, "PushDownLimitScan"),
             RuleID::PushDownSortScan => write!(f, "PushDownSortScan"),
+            RuleID::PushDownLimitWindow => write!(f, "PushDownLimitWindow"),
             RuleID::PushDownFilterWindow => write!(f, "PushDownFilterWindow"),
             RuleID::EliminateEvalScalar => write!(f, "EliminateEvalScalar"),
             RuleID::EliminateFilter => write!(f, "EliminateFilter"),
@@ -138,8 +138,6 @@ impl Display for RuleID {
             RuleID::NormalizeScalarFilter => write!(f, "NormalizeScalarFilter"),
             RuleID::NormalizeAggregate => write!(f, "NormalizeAggregate"),
             RuleID::SplitAggregate => write!(f, "SplitAggregate"),
-            RuleID::NormalizeDisjunctiveFilter => write!(f, "NormalizeDisjunctiveFilter"),
-            RuleID::InferFilter => write!(f, "InferFilter"),
             RuleID::FoldCountAggregate => write!(f, "FoldCountAggregate"),
             RuleID::PushDownPrewhere => write!(f, "PushDownPrewhere"),
 
diff --git a/src/query/sql/src/planner/optimizer/rule/rule_set.rs b/src/query/sql/src/planner/optimizer/rule/rule_set.rs
index c19d615dbd27..2b153f2d4d4d 100644
--- a/src/query/sql/src/planner/optimizer/rule/rule_set.rs
+++ b/src/query/sql/src/planner/optimizer/rule/rule_set.rs
@@ -82,6 +82,10 @@ impl AppliedRules {
     pub fn get(&self, id: &RuleID) -> bool {
         self.rules.contains(id)
     }
+
+    pub fn clear(&mut self) {
+        self.rules = RuleSet::create();
+    }
 }
 
 impl Hash for AppliedRules {
diff --git a/src/query/sql/src/planner/optimizer/rule/transform/rule_commute_join_base_table.rs b/src/query/sql/src/planner/optimizer/rule/transform/rule_commute_join_base_table.rs
index 2cc32b9b9e63..3d32db9ab5bd 100644
--- a/src/query/sql/src/planner/optimizer/rule/transform/rule_commute_join_base_table.rs
+++ b/src/query/sql/src/planner/optimizer/rule/transform/rule_commute_join_base_table.rs
@@ -16,6 +16,7 @@ use std::sync::Arc;
 
 use databend_common_exception::Result;
 
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RuleID;
@@ -23,14 +24,13 @@ use crate::optimizer::SExpr;
 use crate::plans::Join;
 use crate::plans::JoinType;
 use crate::plans::Operator;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 /// Rule to apply commutativity of join operator.
 /// In opposite to RuleCommuteJoin, this rule only applies to base tables.
 pub struct RuleCommuteJoinBaseTable {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleCommuteJoinBaseTable {
@@ -41,16 +41,10 @@ impl RuleCommuteJoinBaseTable {
             // LogicalJoin
             // | \
             // *  *
-            patterns: vec![SExpr::create_binary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Join,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_pattern_leaf()),
-                Arc::new(SExpr::create_pattern_leaf()),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Join,
+                children: vec![Matcher::Leaf, Matcher::Leaf],
+            }],
         }
     }
 }
@@ -104,10 +98,9 @@ impl Rule for RuleCommuteJoinBaseTable {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
-
     fn transformation(&self) -> bool {
         false
     }
diff --git a/src/query/sql/src/planner/optimizer/rule/transform/rule_eager_aggregation.rs b/src/query/sql/src/planner/optimizer/rule/transform/rule_eager_aggregation.rs
index 67b123bb20a7..6fa83d22d1fd 100644
--- a/src/query/sql/src/planner/optimizer/rule/transform/rule_eager_aggregation.rs
+++ b/src/query/sql/src/planner/optimizer/rule/transform/rule_eager_aggregation.rs
@@ -22,6 +22,7 @@ use databend_common_functions::aggregates::AggregateFunctionFactory;
 
 use crate::binder::wrap_cast;
 use crate::binder::ColumnBindingBuilder;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::ColumnSet;
@@ -36,7 +37,6 @@ use crate::plans::EvalScalar;
 use crate::plans::FunctionCall;
 use crate::plans::Join;
 use crate::plans::JoinType;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::plans::ScalarItem;
@@ -128,7 +128,7 @@ use crate::Visibility;
 
 pub struct RuleEagerAggregation {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
     metadata: MetadataRef,
 }
 
@@ -142,7 +142,7 @@ impl RuleEagerAggregation {
             // Expression
             //     |
             //     *
-            patterns: vec![
+            matchers: vec![
                 //     Expression
                 //         |
                 //  Aggregate(final)
@@ -152,40 +152,19 @@ impl RuleEagerAggregation {
                 //        Join
                 //       /    \
                 //      *      *
-                SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Aggregate,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Aggregate,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_binary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::Join,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_pattern_leaf()),
-                                Arc::new(SExpr::create_pattern_leaf()),
-                            )),
-                        )),
-                    )),
-                ),
+                Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Aggregate,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Aggregate,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::Join,
+                                children: vec![Matcher::Leaf, Matcher::Leaf],
+                            }],
+                        }],
+                    }],
+                },
                 //     Expression
                 //         |
                 //  Aggregate(final)
@@ -197,48 +176,22 @@ impl RuleEagerAggregation {
                 //        Join
                 //       /    \
                 //      *      *
-                SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Aggregate,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Aggregate,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_unary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::EvalScalar,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_binary(
-                                    Arc::new(
-                                        PatternPlan {
-                                            plan_type: RelOp::Join,
-                                        }
-                                        .into(),
-                                    ),
-                                    Arc::new(SExpr::create_pattern_leaf()),
-                                    Arc::new(SExpr::create_pattern_leaf()),
-                                )),
-                            )),
-                        )),
-                    )),
-                ),
+                Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Aggregate,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Aggregate,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::EvalScalar,
+                                children: vec![Matcher::MatchOp {
+                                    op_type: RelOp::Join,
+                                    children: vec![Matcher::Leaf, Matcher::Leaf],
+                                }],
+                            }],
+                        }],
+                    }],
+                },
                 //     Expression
                 //         |
                 //        Sort
@@ -250,48 +203,22 @@ impl RuleEagerAggregation {
                 //        Join
                 //       /    \
                 //      *      *
-                SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Sort,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Aggregate,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_unary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::Aggregate,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_binary(
-                                    Arc::new(
-                                        PatternPlan {
-                                            plan_type: RelOp::Join,
-                                        }
-                                        .into(),
-                                    ),
-                                    Arc::new(SExpr::create_pattern_leaf()),
-                                    Arc::new(SExpr::create_pattern_leaf()),
-                                )),
-                            )),
-                        )),
-                    )),
-                ),
+                Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Sort,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Aggregate,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::Aggregate,
+                                children: vec![Matcher::MatchOp {
+                                    op_type: RelOp::Join,
+                                    children: vec![Matcher::Leaf, Matcher::Leaf],
+                                }],
+                            }],
+                        }],
+                    }],
+                },
                 //     Expression
                 //         |
                 //        Sort
@@ -305,56 +232,25 @@ impl RuleEagerAggregation {
                 //        Join
                 //       /    \
                 //      *      *
-                SExpr::create_unary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::EvalScalar,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_unary(
-                        Arc::new(
-                            PatternPlan {
-                                plan_type: RelOp::Sort,
-                            }
-                            .into(),
-                        ),
-                        Arc::new(SExpr::create_unary(
-                            Arc::new(
-                                PatternPlan {
-                                    plan_type: RelOp::Aggregate,
-                                }
-                                .into(),
-                            ),
-                            Arc::new(SExpr::create_unary(
-                                Arc::new(
-                                    PatternPlan {
-                                        plan_type: RelOp::Aggregate,
-                                    }
-                                    .into(),
-                                ),
-                                Arc::new(SExpr::create_unary(
-                                    Arc::new(
-                                        PatternPlan {
-                                            plan_type: RelOp::EvalScalar,
-                                        }
-                                        .into(),
-                                    ),
-                                    Arc::new(SExpr::create_binary(
-                                        Arc::new(
-                                            PatternPlan {
-                                                plan_type: RelOp::Join,
-                                            }
-                                            .into(),
-                                        ),
-                                        Arc::new(SExpr::create_pattern_leaf()),
-                                        Arc::new(SExpr::create_pattern_leaf()),
-                                    )),
-                                )),
-                            )),
-                        )),
-                    )),
-                ),
+                Matcher::MatchOp {
+                    op_type: RelOp::EvalScalar,
+                    children: vec![Matcher::MatchOp {
+                        op_type: RelOp::Sort,
+                        children: vec![Matcher::MatchOp {
+                            op_type: RelOp::Aggregate,
+                            children: vec![Matcher::MatchOp {
+                                op_type: RelOp::Aggregate,
+                                children: vec![Matcher::MatchOp {
+                                    op_type: RelOp::EvalScalar,
+                                    children: vec![Matcher::MatchOp {
+                                        op_type: RelOp::Join,
+                                        children: vec![Matcher::Leaf, Matcher::Leaf],
+                                    }],
+                                }],
+                            }],
+                        }],
+                    }],
+                },
             ],
             metadata,
         }
@@ -372,8 +268,8 @@ impl Rule for RuleEagerAggregation {
         state: &mut TransformResult,
     ) -> databend_common_exception::Result<()> {
         let mut matched_idx = 0;
-        for (idx, pattern) in self.patterns.iter().enumerate() {
-            if s_expr.match_pattern(pattern) {
+        for (idx, matcher) in self.matchers.iter().enumerate() {
+            if matcher.matches(s_expr) {
                 matched_idx = idx + 1;
                 break;
             }
@@ -1254,8 +1150,12 @@ impl Rule for RuleEagerAggregation {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
+    }
+
+    fn transformation(&self) -> bool {
+        false
     }
 }
 
diff --git a/src/query/sql/src/planner/optimizer/rule/transform/rule_left_exchange_join.rs b/src/query/sql/src/planner/optimizer/rule/transform/rule_left_exchange_join.rs
index 78069e63e084..8ae8e949c1b2 100644
--- a/src/query/sql/src/planner/optimizer/rule/transform/rule_left_exchange_join.rs
+++ b/src/query/sql/src/planner/optimizer/rule/transform/rule_left_exchange_join.rs
@@ -20,15 +20,14 @@ use databend_common_exception::Result;
 
 use super::util::get_join_predicates;
 use crate::binder::JoinPredicate;
+use crate::optimizer::extract::Matcher;
 use crate::optimizer::rule::Rule;
 use crate::optimizer::rule::TransformResult;
 use crate::optimizer::RelExpr;
 use crate::optimizer::RuleID;
 use crate::optimizer::SExpr;
-use crate::plans::ComparisonOp;
 use crate::plans::Join;
 use crate::plans::JoinType;
-use crate::plans::PatternPlan;
 use crate::plans::RelOp;
 
 /// Rule to apply swap on a left-deep join.
@@ -48,7 +47,7 @@ use crate::plans::RelOp;
 ///  t1  t3
 pub struct RuleLeftExchangeJoin {
     id: RuleID,
-    patterns: Vec<SExpr>,
+    matchers: Vec<Matcher>,
 }
 
 impl RuleLeftExchangeJoin {
@@ -63,25 +62,16 @@ impl RuleLeftExchangeJoin {
             // | \
             // |  *
             // *
-            patterns: vec![SExpr::create_binary(
-                Arc::new(
-                    PatternPlan {
-                        plan_type: RelOp::Join,
-                    }
-                    .into(),
-                ),
-                Arc::new(SExpr::create_binary(
-                    Arc::new(
-                        PatternPlan {
-                            plan_type: RelOp::Join,
-                        }
-                        .into(),
-                    ),
-                    Arc::new(SExpr::create_pattern_leaf()),
-                    Arc::new(SExpr::create_pattern_leaf()),
-                )),
-                Arc::new(SExpr::create_pattern_leaf()),
-            )],
+            matchers: vec![Matcher::MatchOp {
+                op_type: RelOp::Join,
+                children: vec![
+                    Matcher::MatchOp {
+                        op_type: RelOp::Join,
+                        children: vec![Matcher::Leaf, Matcher::Leaf],
+                    },
+                    Matcher::Leaf,
+                ],
+            }],
         }
     }
 }
@@ -154,8 +144,12 @@ impl Rule for RuleLeftExchangeJoin {
                     // TODO(leiysky): push down the predicate
                     join_3.non_equi_conditions.push(pred.clone());
                 }
-                JoinPredicate::Both { left, right, op } => {
-                    if op == ComparisonOp::Equal {
+                JoinPredicate::Both {
+                    left,
+                    right,
+                    is_equal_op,
+                } => {
+                    if is_equal_op {
                         join_3.left_conditions.push(left.clone());
                         join_3.right_conditions.push(right.clone());
                     } else {
@@ -183,8 +177,12 @@ impl Rule for RuleLeftExchangeJoin {
                     // TODO(leiysky): push down the predicate
                     join_4.non_equi_conditions.push(predicate.clone());
                 }
-                JoinPredicate::Both { left, right, op } => {
-                    if op == ComparisonOp::Equal {
+                JoinPredicate::Both {
+                    left,
+                    right,
+                    is_equal_op,
+                } => {
+                    if is_equal_op {
                         join_4.left_conditions.push(left.clone());
                         join_4.right_conditions.push(right.clone());
                     } else {
@@ -229,8 +227,8 @@ impl Rule for RuleLeftExchangeJoin {
         Ok(())
     }
 
-    fn patterns(&self) -> &Vec<SExpr> {
-        &self.patterns
+    fn matchers(&self) -> &[Matcher] {
+        &self.matchers
     }
 
     fn transformation(&self) -> bool {
diff --git a/src/query/sql/src/planner/optimizer/s_expr.rs b/src/query/sql/src/planner/optimizer/s_expr.rs
index 6239ec3e2036..4a3cc626a24c 100644
--- a/src/query/sql/src/planner/optimizer/s_expr.rs
+++ b/src/query/sql/src/planner/optimizer/s_expr.rs
@@ -25,9 +25,6 @@ use crate::optimizer::rule::AppliedRules;
 use crate::optimizer::rule::RuleID;
 use crate::optimizer::StatInfo;
 use crate::plans::Exchange;
-use crate::plans::Operator;
-use crate::plans::PatternPlan;
-use crate::plans::RelOp;
 use crate::plans::RelOperator;
 use crate::plans::Scan;
 use crate::plans::SubqueryExpr;
@@ -98,21 +95,6 @@ impl SExpr {
         Self::create(plan, vec![], None, None, None)
     }
 
-    pub fn create_pattern_leaf() -> Self {
-        Self::create(
-            Arc::new(
-                PatternPlan {
-                    plan_type: RelOp::Pattern,
-                }
-                .into(),
-            ),
-            vec![],
-            None,
-            None,
-            None,
-        )
-    }
-
     pub fn plan(&self) -> &RelOperator {
         &self.plan
     }
@@ -132,37 +114,10 @@ impl SExpr {
         self.children.len()
     }
 
-    pub fn is_pattern(&self) -> bool {
-        matches!(self.plan.rel_op(), RelOp::Pattern)
-    }
-
     pub fn original_group(&self) -> Option<IndexType> {
         self.original_group
     }
 
-    pub fn match_pattern(&self, pattern: &SExpr) -> bool {
-        if pattern.plan.rel_op() != RelOp::Pattern {
-            // Pattern is plan
-            if self.plan.rel_op() != pattern.plan.rel_op() {
-                return false;
-            }
-
-            if self.arity() != pattern.arity() {
-                // Check if current expression has same arity with current pattern
-                return false;
-            }
-
-            for (e, p) in self.children.iter().zip(pattern.children.iter()) {
-                // Check children
-                if !e.match_pattern(p) {
-                    return false;
-                }
-            }
-        };
-
-        true
-    }
-
     /// Replace children with given new `children`.
     /// Note that this method will keep the `applied_rules` of
     /// current `SExpr` unchanged.
@@ -362,7 +317,6 @@ impl SExpr {
             | RelOperator::DummyTableScan(_)
             | RelOperator::CteScan(_)
             | RelOperator::AddRowNumber(_)
-            | RelOperator::Pattern(_)
             | RelOperator::MaterializedCte(_)
             | RelOperator::ConstantTableScan(_) => {}
         };
@@ -416,6 +370,21 @@ impl SExpr {
 
         add_internal_column_index_into_child(expr, column_index, table_index)
     }
+
+    // The method will clear the applied rules of current SExpr and its children.
+    pub fn clear_applied_rules(&mut self) {
+        self.applied_rules.clear();
+        let children = self
+            .children()
+            .iter()
+            .map(|child| {
+                let mut child = (**child).clone();
+                child.clear_applied_rules();
+                Arc::new(child)
+            })
+            .collect::<Vec<_>>();
+        self.children = children;
+    }
 }
 
 fn find_subquery(rel_op: &RelOperator) -> bool {
@@ -428,7 +397,6 @@ fn find_subquery(rel_op: &RelOperator) -> bool {
         | RelOperator::DummyTableScan(_)
         | RelOperator::CteScan(_)
         | RelOperator::AddRowNumber(_)
-        | RelOperator::Pattern(_)
         | RelOperator::MaterializedCte(_)
         | RelOperator::ConstantTableScan(_) => false,
         RelOperator::Join(op) => {
diff --git a/src/query/sql/src/planner/planner.rs b/src/query/sql/src/planner/planner.rs
index 2878bf93b8ea..0debd05b6491 100644
--- a/src/query/sql/src/planner/planner.rs
+++ b/src/query/sql/src/planner/planner.rs
@@ -93,12 +93,6 @@ impl Planner {
                 // Step 2: Parse the SQL.
                 let (mut stmt, format) = parse_sql(&tokens, sql_dialect)?;
 
-                if matches!(stmt, Statement::CopyIntoLocation(_)) {
-                    // Indicate binder there is no need to collect column statistics for the binding table.
-                    self.ctx
-                        .attach_query_str(QueryKind::CopyIntoTable, String::new());
-                }
-
                 self.replace_stmt(&mut stmt, sql_dialect);
 
                 // Step 3: Bind AST with catalog, and generate a pure logical SExpr
@@ -110,7 +104,14 @@ impl Planner {
                     name_resolution_ctx,
                     metadata.clone(),
                 );
+
+                // Indicate binder there is no need to collect column statistics for the binding table.
+                self.ctx
+                    .attach_query_str(get_query_kind(&stmt), stmt.to_mask_sql());
                 let plan = binder.bind(&stmt).await?;
+                // attach again to avoid the query kind is overwritten by the subquery
+                self.ctx
+                    .attach_query_str(get_query_kind(&stmt), stmt.to_mask_sql());
 
                 // Step 4: Optimize the SExpr with optimizers, and generate optimized physical SExpr
                 let opt_ctx = OptimizerContext::new(self.ctx.clone(), metadata.clone())
@@ -192,3 +193,19 @@ impl Planner {
         self.add_max_rows_limit(stmt);
     }
 }
+
+pub fn get_query_kind(stmt: &Statement) -> QueryKind {
+    match stmt {
+        Statement::Query { .. } => QueryKind::Query,
+        Statement::CopyIntoTable(_) => QueryKind::CopyIntoTable,
+        Statement::CopyIntoLocation(_) => QueryKind::CopyIntoLocation,
+        Statement::Explain { .. } => QueryKind::Explain,
+        Statement::Insert(_) => QueryKind::Insert,
+        Statement::Replace(_)
+        | Statement::Delete(_)
+        | Statement::MergeInto(_)
+        | Statement::OptimizeTable(_)
+        | Statement::Update(_) => QueryKind::Update,
+        _ => QueryKind::Other,
+    }
+}
diff --git a/src/query/sql/src/planner/plans/aggregate.rs b/src/query/sql/src/planner/plans/aggregate.rs
index 9e62b698ddaa..7e1c6e6197dc 100644
--- a/src/query/sql/src/planner/plans/aggregate.rs
+++ b/src/query/sql/src/planner/plans/aggregate.rs
@@ -268,12 +268,13 @@ impl Operator for Aggregate {
             f64::min(res, cardinality)
         };
 
-        let precise_cardinality =
-            if self.group_items.is_empty() && self.mode == AggregateMode::Final {
-                Some(1)
-            } else {
-                None
-            };
+        let precise_cardinality = if self.group_items.is_empty()
+            && matches!(self.mode, AggregateMode::Final | AggregateMode::Initial)
+        {
+            Some(1)
+        } else {
+            None
+        };
         Ok(Arc::new(StatInfo {
             cardinality,
             statistics: Statistics {
diff --git a/src/query/sql/src/planner/plans/ddl/task.rs b/src/query/sql/src/planner/plans/ddl/task.rs
index d3c48b624561..a1fd9c14c556 100644
--- a/src/query/sql/src/planner/plans/ddl/task.rs
+++ b/src/query/sql/src/planner/plans/ddl/task.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::BTreeMap;
 use std::sync::Arc;
 
 use databend_common_ast::ast::AlterTaskOptions;
@@ -47,6 +48,7 @@ pub fn task_schema() -> DataSchemaRef {
         DataField::new("next_schedule_time", DataType::Timestamp.wrap_nullable()),
         DataField::new("last_committed_on", DataType::Timestamp),
         DataField::new("last_suspended_on", DataType::Timestamp.wrap_nullable()),
+        DataField::new("session_parameters", DataType::Variant.wrap_nullable()),
     ]))
 }
 
@@ -69,6 +71,7 @@ pub fn task_run_schema() -> DataSchemaRef {
         DataField::new("completed_time", DataType::Timestamp.wrap_nullable()),
         DataField::new("scheduled_time", DataType::Timestamp),
         DataField::new("root_task_id", DataType::String),
+        DataField::new("session_parameters", DataType::Variant.wrap_nullable()),
     ]))
 }
 
@@ -82,6 +85,7 @@ pub struct CreateTaskPlan {
     pub after: Vec<String>,
     pub when_condition: Option<String>,
     pub suspend_task_after_num_failures: Option<u64>,
+    pub session_parameters: BTreeMap<String, String>,
     pub sql: String,
     pub comment: String,
 }
diff --git a/src/query/sql/src/planner/plans/join.rs b/src/query/sql/src/planner/plans/join.rs
index 8e0cb9213177..f047dc15e964 100644
--- a/src/query/sql/src/planner/plans/join.rs
+++ b/src/query/sql/src/planner/plans/join.rs
@@ -81,7 +81,14 @@ impl JoinType {
     }
 
     pub fn is_outer_join(&self) -> bool {
-        matches!(self, JoinType::Left | JoinType::Right | JoinType::Full)
+        matches!(
+            self,
+            JoinType::Left
+                | JoinType::Right
+                | JoinType::Full
+                | JoinType::LeftSingle
+                | JoinType::RightSingle
+        )
     }
 
     pub fn is_mark_join(&self) -> bool {
@@ -151,8 +158,10 @@ pub struct Join {
     // if we execute distributed merge into, we need to hold the
     // hash table to get not match data from source.
     pub need_hold_hash_table: bool,
-    // Under cluster, mark if the join is broadcast join.
-    pub broadcast: bool,
+    pub is_lateral: bool,
+    // When left/right single join converted to inner join, record the original join type
+    // and do some special processing during runtime.
+    pub single_to_inner: Option<JoinType>,
 }
 
 impl Default for Join {
@@ -165,7 +174,8 @@ impl Default for Join {
             marker_index: Default::default(),
             from_correlated_subquery: Default::default(),
             need_hold_hash_table: false,
-            broadcast: false,
+            is_lateral: false,
+            single_to_inner: None,
         }
     }
 }
@@ -526,7 +536,7 @@ impl Operator for Join {
                 | JoinType::Full
                 | JoinType::RightAnti
                 | JoinType::RightSemi
-                | JoinType::RightMark
+                | JoinType::LeftMark
         ) {
             let left_stat_info = rel_expr.derive_cardinality_child(0)?;
             let right_stat_info = rel_expr.derive_cardinality_child(1)?;
@@ -593,7 +603,7 @@ impl Operator for Join {
                 | JoinType::Full
                 | JoinType::RightAnti
                 | JoinType::RightSemi
-                | JoinType::RightMark
+                | JoinType::LeftMark
                 | JoinType::RightSingle
         ) {
             // (Any, Broadcast)
diff --git a/src/query/sql/src/planner/plans/mod.rs b/src/query/sql/src/planner/plans/mod.rs
index 04e5dff0da5f..2f0127173091 100644
--- a/src/query/sql/src/planner/plans/mod.rs
+++ b/src/query/sql/src/planner/plans/mod.rs
@@ -35,7 +35,6 @@ mod udf;
 
 mod copy_into_location;
 pub mod operator;
-mod pattern;
 mod plan;
 mod presign;
 mod project_set;
@@ -79,7 +78,6 @@ pub use merge_into::DELETE_NAME;
 pub use merge_into::INSERT_NAME;
 pub use merge_into::UPDATE_NAME;
 pub use operator::*;
-pub use pattern::PatternPlan;
 pub use plan::*;
 pub use presign::*;
 pub use project_set::*;
diff --git a/src/query/sql/src/planner/plans/operator.rs b/src/query/sql/src/planner/plans/operator.rs
index 75328d6c594f..9a7cefae3e11 100644
--- a/src/query/sql/src/planner/plans/operator.rs
+++ b/src/query/sql/src/planner/plans/operator.rs
@@ -25,7 +25,6 @@ use super::eval_scalar::EvalScalar;
 use super::filter::Filter;
 use super::join::Join;
 use super::limit::Limit;
-use super::pattern::PatternPlan;
 use super::scan::Scan;
 use super::sort::Sort;
 use super::union_all::UnionAll;
@@ -49,11 +48,6 @@ pub trait Operator {
     /// Get arity of this operator
     fn arity(&self) -> usize;
 
-    /// Is this operator a pattern operator
-    fn is_pattern(&self) -> bool {
-        false
-    }
-
     /// Derive relational property
     fn derive_relational_prop(&self, rel_expr: &RelExpr) -> Result<Arc<RelationalProperty>>;
 
@@ -128,7 +122,6 @@ pub enum RelOperator {
     MaterializedCte(MaterializedCte),
     ConstantTableScan(ConstantTableScan),
     Udf(Udf),
-    Pattern(PatternPlan),
 }
 
 impl Operator for RelOperator {
@@ -141,7 +134,6 @@ impl Operator for RelOperator {
             RelOperator::Aggregate(rel_op) => rel_op.rel_op(),
             RelOperator::Sort(rel_op) => rel_op.rel_op(),
             RelOperator::Limit(rel_op) => rel_op.rel_op(),
-            RelOperator::Pattern(rel_op) => rel_op.rel_op(),
             RelOperator::Exchange(rel_op) => rel_op.rel_op(),
             RelOperator::UnionAll(rel_op) => rel_op.rel_op(),
             RelOperator::DummyTableScan(rel_op) => rel_op.rel_op(),
@@ -174,7 +166,6 @@ impl Operator for RelOperator {
             RelOperator::MaterializedCte(rel_op) => rel_op.arity(),
             RelOperator::ConstantTableScan(rel_op) => rel_op.arity(),
             RelOperator::Udf(rel_op) => rel_op.arity(),
-            RelOperator::Pattern(rel_op) => rel_op.arity(),
         }
     }
 
@@ -187,7 +178,6 @@ impl Operator for RelOperator {
             RelOperator::Aggregate(rel_op) => rel_op.derive_relational_prop(rel_expr),
             RelOperator::Sort(rel_op) => rel_op.derive_relational_prop(rel_expr),
             RelOperator::Limit(rel_op) => rel_op.derive_relational_prop(rel_expr),
-            RelOperator::Pattern(rel_op) => rel_op.derive_relational_prop(rel_expr),
             RelOperator::Exchange(rel_op) => rel_op.derive_relational_prop(rel_expr),
             RelOperator::UnionAll(rel_op) => rel_op.derive_relational_prop(rel_expr),
             RelOperator::DummyTableScan(rel_op) => rel_op.derive_relational_prop(rel_expr),
@@ -210,7 +200,6 @@ impl Operator for RelOperator {
             RelOperator::Aggregate(rel_op) => rel_op.derive_physical_prop(rel_expr),
             RelOperator::Sort(rel_op) => rel_op.derive_physical_prop(rel_expr),
             RelOperator::Limit(rel_op) => rel_op.derive_physical_prop(rel_expr),
-            RelOperator::Pattern(rel_op) => rel_op.derive_physical_prop(rel_expr),
             RelOperator::Exchange(rel_op) => rel_op.derive_physical_prop(rel_expr),
             RelOperator::UnionAll(rel_op) => rel_op.derive_physical_prop(rel_expr),
             RelOperator::DummyTableScan(rel_op) => rel_op.derive_physical_prop(rel_expr),
@@ -233,7 +222,6 @@ impl Operator for RelOperator {
             RelOperator::Aggregate(rel_op) => rel_op.derive_stats(rel_expr),
             RelOperator::Sort(rel_op) => rel_op.derive_stats(rel_expr),
             RelOperator::Limit(rel_op) => rel_op.derive_stats(rel_expr),
-            RelOperator::Pattern(rel_op) => rel_op.derive_stats(rel_expr),
             RelOperator::Exchange(rel_op) => rel_op.derive_stats(rel_expr),
             RelOperator::UnionAll(rel_op) => rel_op.derive_stats(rel_expr),
             RelOperator::DummyTableScan(rel_op) => rel_op.derive_stats(rel_expr),
@@ -276,9 +264,6 @@ impl Operator for RelOperator {
             RelOperator::Limit(rel_op) => {
                 rel_op.compute_required_prop_child(ctx, rel_expr, child_index, required)
             }
-            RelOperator::Pattern(rel_op) => {
-                rel_op.compute_required_prop_child(ctx, rel_expr, child_index, required)
-            }
             RelOperator::Exchange(rel_op) => {
                 rel_op.compute_required_prop_child(ctx, rel_expr, child_index, required)
             }
@@ -370,9 +355,6 @@ impl Operator for RelOperator {
             RelOperator::Udf(rel_op) => {
                 rel_op.compute_required_prop_children(ctx, rel_expr, required)
             }
-            RelOperator::Pattern(rel_op) => {
-                rel_op.compute_required_prop_children(ctx, rel_expr, required)
-            }
         }
     }
 }
@@ -562,25 +544,6 @@ impl TryFrom<RelOperator> for Limit {
     }
 }
 
-impl From<PatternPlan> for RelOperator {
-    fn from(v: PatternPlan) -> Self {
-        Self::Pattern(v)
-    }
-}
-
-impl TryFrom<RelOperator> for PatternPlan {
-    type Error = ErrorCode;
-    fn try_from(value: RelOperator) -> Result<Self> {
-        if let RelOperator::Pattern(value) = value {
-            Ok(value)
-        } else {
-            Err(ErrorCode::Internal(
-                "Cannot downcast RelOperator to Pattern",
-            ))
-        }
-    }
-}
-
 impl From<Exchange> for RelOperator {
     fn from(v: Exchange) -> Self {
         Self::Exchange(v)
diff --git a/src/query/sql/src/planner/plans/pattern.rs b/src/query/sql/src/planner/plans/pattern.rs
deleted file mode 100644
index 2f545d321ed9..000000000000
--- a/src/query/sql/src/planner/plans/pattern.rs
+++ /dev/null
@@ -1,95 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::sync::Arc;
-
-use databend_common_catalog::table_context::TableContext;
-use databend_common_exception::ErrorCode;
-
-use crate::optimizer::PhysicalProperty;
-use crate::optimizer::RelExpr;
-use crate::optimizer::RelationalProperty;
-use crate::optimizer::RequiredProperty;
-use crate::optimizer::StatInfo;
-use crate::plans::Operator;
-use crate::plans::RelOp;
-
-#[derive(Clone, Debug, PartialEq, Eq, Hash)]
-pub struct PatternPlan {
-    pub plan_type: RelOp,
-}
-
-impl Operator for PatternPlan {
-    fn rel_op(&self) -> RelOp {
-        self.plan_type.clone()
-    }
-
-    fn arity(&self) -> usize {
-        0
-    }
-
-    fn is_pattern(&self) -> bool {
-        true
-    }
-
-    fn derive_relational_prop(
-        &self,
-        _rel_expr: &RelExpr,
-    ) -> databend_common_exception::Result<Arc<RelationalProperty>> {
-        Err(ErrorCode::Internal(
-            "Cannot derive relational property for pattern plan",
-        ))
-    }
-
-    fn derive_physical_prop(
-        &self,
-        _rel_expr: &RelExpr,
-    ) -> databend_common_exception::Result<PhysicalProperty> {
-        Err(ErrorCode::Internal(
-            "Cannot derive physical property for pattern plan",
-        ))
-    }
-
-    fn derive_stats(
-        &self,
-        _rel_expr: &RelExpr,
-    ) -> databend_common_exception::Result<Arc<StatInfo>> {
-        Err(ErrorCode::Internal(
-            "Cannot derive cardinality for pattern plan",
-        ))
-    }
-
-    fn compute_required_prop_child(
-        &self,
-        _ctx: Arc<dyn TableContext>,
-        _rel_expr: &RelExpr,
-        _child_index: usize,
-        _required: &RequiredProperty,
-    ) -> databend_common_exception::Result<RequiredProperty> {
-        Err(ErrorCode::Internal(
-            "Cannot compute required property for pattern plan",
-        ))
-    }
-
-    fn compute_required_prop_children(
-        &self,
-        _ctx: Arc<dyn TableContext>,
-        _rel_expr: &RelExpr,
-        _required: &RequiredProperty,
-    ) -> databend_common_exception::Result<Vec<Vec<RequiredProperty>>> {
-        Err(ErrorCode::Internal(
-            "Cannot compute required property for pattern plan",
-        ))
-    }
-}
diff --git a/src/query/sql/src/planner/plans/scalar_expr.rs b/src/query/sql/src/planner/plans/scalar_expr.rs
index 39d3e82fb5ba..1b581c61e684 100644
--- a/src/query/sql/src/planner/plans/scalar_expr.rs
+++ b/src/query/sql/src/planner/plans/scalar_expr.rs
@@ -176,6 +176,28 @@ impl ScalarExpr {
         visitor.visit(self)?;
         Ok(())
     }
+
+    pub fn has_one_column_ref(&self) -> bool {
+        struct BoundColumnRefVisitor {
+            has_column_ref: bool,
+            num_column_ref: usize,
+        }
+
+        impl<'a> Visitor<'a> for BoundColumnRefVisitor {
+            fn visit_bound_column_ref(&mut self, _col: &'a BoundColumnRef) -> Result<()> {
+                self.has_column_ref = true;
+                self.num_column_ref += 1;
+                Ok(())
+            }
+        }
+
+        let mut visitor = BoundColumnRefVisitor {
+            has_column_ref: false,
+            num_column_ref: 0,
+        };
+        visitor.visit(self).unwrap();
+        visitor.has_column_ref && visitor.num_column_ref == 1
+    }
 }
 
 impl From<BoundColumnRef> for ScalarExpr {
@@ -566,6 +588,8 @@ pub struct SubqueryExpr {
     pub(crate) data_type: Box<DataType>,
     #[educe(Hash(method = "hash_column_set"))]
     pub outer_columns: ColumnSet,
+    // If contain aggregation function in scalar subquery output
+    pub contain_agg: Option<bool>,
 }
 
 impl SubqueryExpr {
diff --git a/src/query/sql/src/planner/plans/window.rs b/src/query/sql/src/planner/plans/window.rs
index 57985e74ddb3..fc394de650b8 100644
--- a/src/query/sql/src/planner/plans/window.rs
+++ b/src/query/sql/src/planner/plans/window.rs
@@ -64,6 +64,8 @@ pub struct Window {
     pub order_by: Vec<WindowOrderByInfo>,
     // window frames
     pub frame: WindowFuncFrame,
+    // limit for potentially possible push-down
+    pub limit: Option<usize>,
 }
 
 impl Window {
diff --git a/src/query/sql/src/planner/semantic/type_check.rs b/src/query/sql/src/planner/semantic/type_check.rs
index 2c6558beb74f..72085e023773 100644
--- a/src/query/sql/src/planner/semantic/type_check.rs
+++ b/src/query/sql/src/planner/semantic/type_check.rs
@@ -17,6 +17,7 @@ use std::collections::VecDeque;
 use std::sync::Arc;
 use std::vec;
 
+use databend_common_ast::ast::contain_agg_func;
 use databend_common_ast::ast::BinaryOperator;
 use databend_common_ast::ast::ColumnID;
 use databend_common_ast::ast::Expr;
@@ -26,6 +27,8 @@ use databend_common_ast::ast::Lambda;
 use databend_common_ast::ast::Literal;
 use databend_common_ast::ast::MapAccessor;
 use databend_common_ast::ast::Query;
+use databend_common_ast::ast::SelectTarget;
+use databend_common_ast::ast::SetExpr;
 use databend_common_ast::ast::SubqueryModifier;
 use databend_common_ast::ast::TrimWhere;
 use databend_common_ast::ast::TypeName;
@@ -2349,6 +2352,17 @@ impl<'a> TypeChecker<'a> {
             )));
         }
 
+        let mut contain_agg = None;
+        if let SetExpr::Select(select_stmt) = &subquery.body {
+            if typ == SubqueryType::Scalar {
+                let select = &select_stmt.select_list[0];
+                if let SelectTarget::AliasedExpr { expr, .. } = select {
+                    // Check if contain aggregation function
+                    contain_agg = Some(contain_agg_func(expr));
+                }
+            }
+        }
+
         let mut data_type = output_context.columns[0].data_type.clone();
 
         let rel_expr = RelExpr::with_s_expr(&s_expr);
@@ -2374,6 +2388,7 @@ impl<'a> TypeChecker<'a> {
             data_type: data_type.clone(),
             typ,
             outer_columns: rel_prop.outer_columns.clone(),
+            contain_agg,
         };
 
         let data_type = subquery_expr.data_type();
@@ -2620,8 +2635,7 @@ impl<'a> TypeChecker<'a> {
                 if args.len() >= 2 {
                     let box (arg, _) = self.resolve(args[1]).await.ok()?;
                     if let Ok(arg) = ConstantExpr::try_from(arg) {
-                        if let Scalar::String(val) = arg.value {
-                            let sort_order = unsafe { std::str::from_utf8_unchecked(&val) };
+                        if let Scalar::String(sort_order) = arg.value {
                             if sort_order.eq_ignore_ascii_case("asc") {
                                 asc = true;
                             } else if sort_order.eq_ignore_ascii_case("desc") {
@@ -2645,8 +2659,7 @@ impl<'a> TypeChecker<'a> {
                 if args.len() == 3 {
                     let box (arg, _) = self.resolve(args[2]).await.ok()?;
                     if let Ok(arg) = ConstantExpr::try_from(arg) {
-                        if let Scalar::String(val) = arg.value {
-                            let nulls_order = unsafe { std::str::from_utf8_unchecked(&val) };
+                        if let Scalar::String(nulls_order) = arg.value {
                             if nulls_order.eq_ignore_ascii_case("nulls first") {
                                 nulls_first = true;
                             } else if nulls_order.eq_ignore_ascii_case("nulls last") {
@@ -2685,8 +2698,7 @@ impl<'a> TypeChecker<'a> {
                 }
                 let box (arg, _) = self.resolve(args[1]).await.ok()?;
                 if let Ok(arg) = ConstantExpr::try_from(arg) {
-                    if let Scalar::String(arg) = arg.value {
-                        let aggr_func_name = unsafe { std::str::from_utf8_unchecked(&arg) };
+                    if let Scalar::String(aggr_func_name) = arg.value {
                         let func_name = format!("array_{}", aggr_func_name);
                         let args_ref: Vec<&Expr> = vec![args[0]];
                         return Some(
@@ -2754,7 +2766,7 @@ impl<'a> TypeChecker<'a> {
         } else {
             let trim_scalar = ConstantExpr {
                 span,
-                value: databend_common_expression::Scalar::String(" ".as_bytes().to_vec()),
+                value: databend_common_expression::Scalar::String(" ".to_string()),
             }
             .into();
             ("trim_both", trim_scalar, DataType::String)
@@ -2782,7 +2794,7 @@ impl<'a> TypeChecker<'a> {
                 scale: *scale,
             })),
             Literal::Float64(float) => Scalar::Number(NumberScalar::Float64((*float).into())),
-            Literal::String(string) => Scalar::String(string.as_bytes().to_vec()),
+            Literal::String(string) => Scalar::String(string.clone()),
             Literal::Boolean(boolean) => Scalar::Boolean(*boolean),
             Literal::Null => Scalar::Null,
         };
@@ -3092,7 +3104,7 @@ impl<'a> TypeChecker<'a> {
                 {
                     let key = ConstantExpr {
                         span,
-                        value: Scalar::String(field_name.clone().into_bytes()),
+                        value: Scalar::String(field_name.clone()),
                     }
                     .into();
 
@@ -3411,6 +3423,7 @@ impl<'a> TypeChecker<'a> {
             data_type: data_type.clone(),
             typ: SubqueryType::Any,
             outer_columns: rel_prop.outer_columns.clone(),
+            contain_agg: None,
         };
         let data_type = subquery_expr.data_type();
         Ok(Box::new((subquery_expr.into(), data_type)))
@@ -3448,7 +3461,7 @@ impl<'a> TypeChecker<'a> {
         let keypaths_str = format!("{}", keypaths);
         let path_scalar = ScalarExpr::ConstantExpr(ConstantExpr {
             span: None,
-            value: Scalar::String(keypaths_str.into_bytes()),
+            value: Scalar::String(keypaths_str),
         });
         let args = vec![scalar, path_scalar];
 
diff --git a/src/query/sql/src/planner/semantic/virtual_column_rewriter.rs b/src/query/sql/src/planner/semantic/virtual_column_rewriter.rs
index 41da8545bc0a..b80b61aba648 100644
--- a/src/query/sql/src/planner/semantic/virtual_column_rewriter.rs
+++ b/src/query/sql/src/planner/semantic/virtual_column_rewriter.rs
@@ -203,7 +203,7 @@ impl VirtualColumnRewriter {
                             return Some(());
                         }
                         let name = match constant.value.clone() {
-                            Scalar::String(v) => match parse_key_paths(&v) {
+                            Scalar::String(v) => match parse_key_paths(v.as_bytes()) {
                                 Ok(key_paths) => {
                                     let mut name = String::new();
                                     name.push_str(&base_column.column_name);
diff --git a/src/query/storages/common/cache/src/providers/table_data_cache.rs b/src/query/storages/common/cache/src/providers/table_data_cache.rs
index 403efae7a7c5..8fb86874dc81 100644
--- a/src/query/storages/common/cache/src/providers/table_data_cache.rs
+++ b/src/query/storages/common/cache/src/providers/table_data_cache.rs
@@ -93,6 +93,7 @@ impl CacheAccessor<String, Bytes, DefaultHashBuilder, Count> for TableDataCache
         let k = k.as_ref();
         if let Some(item) = self.external_cache.get(k) {
             metrics_inc_cache_hit_count(1, TABLE_DATA_CACHE_NAME);
+            // Profile::record_usize_profile(ProfileStatisticsName::ScanCacheBytes, process_values.bytes);
             Some(item)
         } else {
             metrics_inc_cache_miss_count(1, TABLE_DATA_CACHE_NAME);
diff --git a/src/query/storages/common/index/tests/it/filters/bloom_filter.rs b/src/query/storages/common/index/tests/it/filters/bloom_filter.rs
index c6f5de7de357..3eed74e3762d 100644
--- a/src/query/storages/common/index/tests/it/filters/bloom_filter.rs
+++ b/src/query/storages/common/index/tests/it/filters/bloom_filter.rs
@@ -82,7 +82,7 @@ fn test_bloom_filter() -> Result<()> {
                 ),
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(b"a".to_vec())),
+                    Value::Scalar(Scalar::String("a".to_string())),
                 ),
                 BlockEntry::new(
                     map_ty.clone(),
@@ -161,7 +161,7 @@ fn test_bloom_filter() -> Result<()> {
             "1",
             bloom_fields.clone(),
             schema.clone(),
-            Scalar::String(b"a".to_vec()),
+            Scalar::String("a".to_string()),
             DataType::String
         )
     );
@@ -172,7 +172,7 @@ fn test_bloom_filter() -> Result<()> {
             "1",
             bloom_fields.clone(),
             schema.clone(),
-            Scalar::String(b"b".to_vec()),
+            Scalar::String("b".to_string()),
             DataType::String
         )
     );
@@ -183,7 +183,7 @@ fn test_bloom_filter() -> Result<()> {
             "1",
             bloom_fields,
             schema.clone(),
-            Scalar::String(b"d".to_vec()),
+            Scalar::String("d".to_string()),
             DataType::String
         )
     );
@@ -197,7 +197,7 @@ fn test_bloom_filter() -> Result<()> {
             map_ty.clone(),
             Scalar::Number(NumberScalar::UInt8(1)),
             DataType::Number(NumberDataType::UInt8),
-            Scalar::String(b"a".to_vec()),
+            Scalar::String("a".to_string()),
             DataType::String
         )
     );
@@ -210,7 +210,7 @@ fn test_bloom_filter() -> Result<()> {
             map_ty.clone(),
             Scalar::Number(NumberScalar::UInt8(2)),
             DataType::Number(NumberDataType::UInt8),
-            Scalar::String(b"b".to_vec()),
+            Scalar::String("b".to_string()),
             DataType::String
         )
     );
@@ -223,7 +223,7 @@ fn test_bloom_filter() -> Result<()> {
             map_ty,
             Scalar::Number(NumberScalar::UInt8(3)),
             DataType::Number(NumberDataType::UInt8),
-            Scalar::String(b"x".to_vec()),
+            Scalar::String("x".to_string()),
             DataType::String
         )
     );
@@ -261,7 +261,7 @@ fn test_specify_bloom_filter() -> Result<()> {
             "1",
             fields,
             schema,
-            Scalar::String(b"d".to_vec()),
+            Scalar::String("d".to_string()),
             DataType::String
         )
     );
@@ -301,7 +301,7 @@ fn test_string_bloom_filter() -> Result<()> {
             "1",
             fields,
             schema,
-            Scalar::String(b"d".to_vec()),
+            Scalar::String("d".to_string()),
             DataType::String
         )
     );
diff --git a/src/query/storages/common/pruner/src/internal_column_pruner.rs b/src/query/storages/common/pruner/src/internal_column_pruner.rs
index 21d283a498e5..3641431215d2 100644
--- a/src/query/storages/common/pruner/src/internal_column_pruner.rs
+++ b/src/query/storages/common/pruner/src/internal_column_pruner.rs
@@ -59,10 +59,9 @@ impl InternalColumnPruner {
     pub fn should_keep(&self, col_name: &str, value: &str) -> bool {
         if self.input_domains.contains_key(col_name) {
             let mut input_domains = self.input_domains.clone();
-            let bytes = value.as_bytes().to_vec();
             let domain = Domain::String(StringDomain {
-                min: bytes.clone(),
-                max: Some(bytes),
+                min: value.to_string(),
+                max: Some(value.to_string()),
             });
             input_domains.insert(col_name.to_string(), domain);
 
diff --git a/src/query/storages/common/table_meta/benches/bench.rs b/src/query/storages/common/table_meta/benches/bench.rs
index 697b88d1c741..6d95b5f6b07d 100644
--- a/src/query/storages/common/table_meta/benches/bench.rs
+++ b/src/query/storages/common/table_meta/benches/bench.rs
@@ -157,8 +157,8 @@ fn build_test_segment_info(
     });
 
     let col_stat = ColumnStatistics::new(
-        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?.into_bytes()),
-        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?.into_bytes()),
+        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?),
+        Scalar::String(String::from_utf8(vec![b'a'; STATS_STRING_PREFIX_LEN])?),
         0,
         0,
         None,
diff --git a/src/query/storages/delta/src/partition_columns/values_serde.rs b/src/query/storages/delta/src/partition_columns/values_serde.rs
index 66f1c1a13d3f..f35b3c0f3beb 100644
--- a/src/query/storages/delta/src/partition_columns/values_serde.rs
+++ b/src/query/storages/delta/src/partition_columns/values_serde.rs
@@ -38,7 +38,7 @@ fn field_to_value(value: &ParquetFieldValue) -> Result<Scalar> {
         ParquetFieldValue::ULong(v) => Ok(Scalar::Number(NumberScalar::UInt64(*v))),
         ParquetFieldValue::Float(v) => Ok(Scalar::Number(NumberScalar::Float32(OrderedFloat(*v)))),
         ParquetFieldValue::Double(v) => Ok(Scalar::Number(NumberScalar::Float64(OrderedFloat(*v)))),
-        ParquetFieldValue::Str(v) => Ok(Scalar::String(v.as_bytes().to_vec())),
+        ParquetFieldValue::Str(v) => Ok(Scalar::String(v.clone())),
         _ => Err(ErrorCode::IllegalDataType(format!(
             "Unsupported parquet type {:?}",
             value
@@ -54,7 +54,7 @@ pub fn str_to_scalar(value: &str, data_type: &DataType) -> Result<Scalar> {
     }
     match data_type {
         DataType::Nullable(t) => str_to_scalar(value, t),
-        DataType::String => Ok(Scalar::String(value.as_bytes().to_vec())),
+        DataType::String => Ok(Scalar::String(value.to_string())),
         DataType::Number(num_ty) => match num_ty {
             NumberDataType::UInt8 => {
                 let num = value.parse::<u8>().unwrap();
diff --git a/src/query/storages/delta/src/table_source.rs b/src/query/storages/delta/src/table_source.rs
index 362da585817a..887a8f3ea625 100644
--- a/src/query/storages/delta/src/table_source.rs
+++ b/src/query/storages/delta/src/table_source.rs
@@ -31,6 +31,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_storages_parquet::ParquetPart;
 use databend_common_storages_parquet::ParquetRSFullReader;
 use opendal::Reader;
@@ -134,6 +136,10 @@ impl Processor for DeltaTableSource {
                     bytes: data_block.memory_size(),
                 };
                 self.scan_progress.incr(&progress_values);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::ScanBytes,
+                    data_block.memory_size(),
+                );
                 self.output.push_data(Ok(data_block));
                 Ok(Event::NeedConsume)
             }
diff --git a/src/query/storages/fuse/src/operations/agg_index_sink.rs b/src/query/storages/fuse/src/operations/agg_index_sink.rs
index 2c902e81e763..45a3a4b679aa 100644
--- a/src/query/storages/fuse/src/operations/agg_index_sink.rs
+++ b/src/query/storages/fuse/src/operations/agg_index_sink.rs
@@ -21,7 +21,6 @@ use async_trait::unboxed_simple;
 use databend_common_catalog::table_context::TableContext;
 use databend_common_exception::Result;
 use databend_common_expression::types::StringType;
-use databend_common_expression::types::ValueType;
 use databend_common_expression::BlockRowIndex;
 use databend_common_expression::DataBlock;
 use databend_common_expression::TableSchemaRef;
@@ -78,11 +77,7 @@ impl AggIndexSink {
         let block_name_col = col.value.try_downcast::<StringType>().unwrap();
         let block_id = self.blocks.len();
         for i in 0..block.num_rows() {
-            let location = unsafe {
-                String::from_utf8_unchecked(StringType::to_owned_scalar(
-                    block_name_col.index(i).unwrap(),
-                ))
-            };
+            let location = block_name_col.index(i).unwrap().to_string();
 
             self.location_data
                 .entry(location)
diff --git a/src/query/storages/fuse/src/operations/read/native_data_source_deserializer.rs b/src/query/storages/fuse/src/operations/read/native_data_source_deserializer.rs
index 54a4ceda2472..80101a78bcc4 100644
--- a/src/query/storages/fuse/src/operations/read/native_data_source_deserializer.rs
+++ b/src/query/storages/fuse/src/operations/read/native_data_source_deserializer.rs
@@ -60,6 +60,8 @@ use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_sql::IndexType;
 use xorf::BinaryFuse16;
 
@@ -414,6 +416,7 @@ impl NativeDeserializeDataTransform {
             bytes: data_block.memory_size(),
         };
         self.scan_progress.incr(&progress_values);
+        Profile::record_usize_profile(ProfileStatisticsName::ScanBytes, data_block.memory_size());
         self.output_data = Some(data_block);
     }
 
diff --git a/src/query/storages/fuse/src/operations/read/parquet_data_source_deserializer.rs b/src/query/storages/fuse/src/operations/read/parquet_data_source_deserializer.rs
index db6438159bc3..dd340ef5dc77 100644
--- a/src/query/storages/fuse/src/operations/read/parquet_data_source_deserializer.rs
+++ b/src/query/storages/fuse/src/operations/read/parquet_data_source_deserializer.rs
@@ -39,6 +39,8 @@ use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_sql::IndexType;
 use xorf::BinaryFuse16;
 
@@ -251,6 +253,10 @@ impl Processor for DeserializeDataTransform {
                         bytes: block.memory_size(),
                     };
                     self.scan_progress.incr(&progress_values);
+                    Profile::record_usize_profile(
+                        ProfileStatisticsName::ScanBytes,
+                        block.memory_size(),
+                    );
 
                     self.output_data = Some(block);
                 }
@@ -299,6 +305,10 @@ impl Processor for DeserializeDataTransform {
                         bytes: data_block.memory_size(),
                     };
                     self.scan_progress.incr(&progress_values);
+                    Profile::record_usize_profile(
+                        ProfileStatisticsName::ScanBytes,
+                        data_block.memory_size(),
+                    );
 
                     let mut data_block =
                         data_block.resort(&self.src_schema, &self.output_schema)?;
diff --git a/src/query/storages/fuse/src/operations/read/runtime_filter_prunner.rs b/src/query/storages/fuse/src/operations/read/runtime_filter_prunner.rs
index 9090ef2da725..15d5526fbad5 100644
--- a/src/query/storages/fuse/src/operations/read/runtime_filter_prunner.rs
+++ b/src/query/storages/fuse/src/operations/read/runtime_filter_prunner.rs
@@ -49,7 +49,6 @@ pub fn runtime_filter_pruner(
     if filters.is_empty() {
         return Ok(false);
     }
-
     let part = FusePartInfo::from_part(part)?;
     let pruned = filters.iter().any(|filter| {
         let column_refs = filter.column_refs();
diff --git a/src/query/storages/fuse/src/operations/replace_into/mutator/column_hash.rs b/src/query/storages/fuse/src/operations/replace_into/mutator/column_hash.rs
index 7518ac607a34..8cbaec3d8c9a 100644
--- a/src/query/storages/fuse/src/operations/replace_into/mutator/column_hash.rs
+++ b/src/query/storages/fuse/src/operations/replace_into/mutator/column_hash.rs
@@ -70,7 +70,7 @@ pub fn row_hash_of_columns(
                 NumberScalar::Float64(v) => sip.write_u64(v.to_bits()),
             },
             ScalarRef::Timestamp(v) => sip.write_i64(v),
-            ScalarRef::String(v) => sip.write(v),
+            ScalarRef::String(v) => sip.write(v.as_bytes()),
             ScalarRef::Bitmap(v) => sip.write(v),
             ScalarRef::Decimal(v) => match v {
                 DecimalScalar::Decimal128(i, DecimalSize { precision, scale }) => {
diff --git a/src/query/storages/fuse/src/operations/replace_into/mutator/merge_into_mutator.rs b/src/query/storages/fuse/src/operations/replace_into/mutator/merge_into_mutator.rs
index f24d4c643b5a..42f8f003f824 100644
--- a/src/query/storages/fuse/src/operations/replace_into/mutator/merge_into_mutator.rs
+++ b/src/query/storages/fuse/src/operations/replace_into/mutator/merge_into_mutator.rs
@@ -788,8 +788,8 @@ mod tests {
             (
                 1,
                 range(
-                    Scalar::String("a".to_string().into_bytes()),
-                    Scalar::String("z".to_string().into_bytes()),
+                    Scalar::String("a".to_string()),
+                    Scalar::String("z".to_string()),
                 ),
             ),
             // range of xx_time [100, 200]
@@ -827,8 +827,8 @@ mod tests {
             ),
             // for xx_type column, overlaps
             (
-                Scalar::String("b".to_string().into_bytes()),
-                Scalar::String("y".to_string().into_bytes()),
+                Scalar::String("b".to_string()),
+                Scalar::String("y".to_string()),
             ),
             // for xx_time column, overlaps
             (
@@ -872,8 +872,8 @@ mod tests {
             ),
             // for xx_type column, overlaps
             (
-                Scalar::String("b".to_string().into_bytes()),
-                Scalar::String("b".to_string().into_bytes()),
+                Scalar::String("b".to_string()),
+                Scalar::String("b".to_string()),
             ),
             // for xx_time column, overlaps
             (
diff --git a/src/query/storages/fuse/src/operations/replace_into/mutator/mutator_replace_into.rs b/src/query/storages/fuse/src/operations/replace_into/mutator/mutator_replace_into.rs
index 6c6ab26e69a7..b8fd2f718fa9 100644
--- a/src/query/storages/fuse/src/operations/replace_into/mutator/mutator_replace_into.rs
+++ b/src/query/storages/fuse/src/operations/replace_into/mutator/mutator_replace_into.rs
@@ -270,11 +270,10 @@ impl ReplaceIntoMutator {
             ScalarRef::Tuple(_) => "[TUPLE]".to_owned(),
             ScalarRef::Variant(_) => "[VARIANT]".to_owned(),
             // for string, return the first 5 chars
-            ScalarRef::String(s) => {
-                let val = String::from_utf8_lossy(s).to_string();
+            ScalarRef::String(val) => {
                 // take the first 5 chars
-                match val.as_str().char_indices().nth(5) {
-                    None => val,
+                match val.char_indices().nth(5) {
+                    None => val.to_string(),
                     Some((idx, _)) => format!("{}...", &val[..idx]),
                 }
             }
diff --git a/src/query/storages/fuse/src/statistics/column_statistic.rs b/src/query/storages/fuse/src/statistics/column_statistic.rs
index ebc00004470c..2ca1ea61998b 100644
--- a/src/query/storages/fuse/src/statistics/column_statistic.rs
+++ b/src/query/storages/fuse/src/statistics/column_statistic.rs
@@ -222,86 +222,74 @@ pub const STATS_STRING_PREFIX_LEN: usize = 16;
 impl Trim for Scalar {
     fn trim_min(self, trim_len: usize) -> Option<Self> {
         match self {
-            Scalar::String(bytes) => match String::from_utf8(bytes) {
-                Ok(mut v) => {
-                    if v.len() <= trim_len {
-                        Some(Scalar::String(v.into_bytes()))
-                    } else {
-                        // find the character boundary to prevent String::truncate from panic
-                        let vs = v.as_str();
-                        let slice = match vs.char_indices().nth(trim_len) {
-                            None => vs,
-                            Some((idx, _)) => &vs[..idx],
-                        };
+            Scalar::String(mut s) => {
+                if s.len() <= trim_len {
+                    Some(Scalar::String(s))
+                } else {
+                    // find the character boundary to prevent String::truncate from panic
+                    let vs = s.as_str();
+                    let slice = match vs.char_indices().nth(trim_len) {
+                        None => vs,
+                        Some((idx, _)) => &vs[..idx],
+                    };
 
-                        // do truncate
-                        Some(Scalar::String({
-                            v.truncate(slice.len());
-                            v.into_bytes()
-                        }))
-                    }
-                }
-                Err(_) => {
-                    // if failed to convert the bytes into (utf-8)string, just ignore it.
-                    None
+                    // do truncate
+                    Some(Scalar::String({
+                        s.truncate(slice.len());
+                        s
+                    }))
                 }
-            },
+            }
             v => Some(v),
         }
     }
 
     fn trim_max(self, trim_len: usize) -> Option<Self> {
         match self {
-            Scalar::String(bytes) => match String::from_utf8(bytes) {
-                Ok(v) => {
-                    if v.len() <= trim_len {
-                        // if number of bytes is lesser, just return
-                        Some(Scalar::String(v.into_bytes()))
-                    } else {
-                        // no need to trim, less than STRING_PREFIX_LEN chars
-                        let number_of_chars = v.as_str().chars().count();
-                        if number_of_chars <= trim_len {
-                            return Some(Scalar::String(v.into_bytes()));
-                        }
+            Scalar::String(v) => {
+                if v.len() <= trim_len {
+                    // if number of bytes is lesser, just return
+                    Some(Scalar::String(v))
+                } else {
+                    // no need to trim, less than STRING_PREFIX_LEN chars
+                    let number_of_chars = v.as_str().chars().count();
+                    if number_of_chars <= trim_len {
+                        return Some(Scalar::String(v));
+                    }
 
-                        // slice the input (at the boundary of chars), takes at most STRING_PREFIX_LEN chars
-                        let vs = v.as_str();
-                        let sliced = match vs.char_indices().nth(trim_len) {
-                            None => vs,
-                            Some((idx, _)) => &vs[..idx],
-                        };
+                    // slice the input (at the boundary of chars), takes at most STRING_PREFIX_LEN chars
+                    let vs = v.as_str();
+                    let sliced = match vs.char_indices().nth(trim_len) {
+                        None => vs,
+                        Some((idx, _)) => &vs[..idx],
+                    };
 
-                        // find the position to replace the char with REPLACEMENT_CHAR
-                        // in reversed order, break at the first one we met
-                        let mut idx = None;
-                        for (i, c) in sliced.char_indices().rev() {
-                            if c < STATS_REPLACEMENT_CHAR {
-                                idx = Some(i);
-                                break;
-                            }
+                    // find the position to replace the char with REPLACEMENT_CHAR
+                    // in reversed order, break at the first one we met
+                    let mut idx = None;
+                    for (i, c) in sliced.char_indices().rev() {
+                        if c < STATS_REPLACEMENT_CHAR {
+                            idx = Some(i);
+                            break;
                         }
+                    }
 
-                        // grab the replacement_point
-                        let replacement_point = idx?;
+                    // grab the replacement_point
+                    let replacement_point = idx?;
 
-                        // rebuild the string (since the len of result string is rather small)
-                        let mut r = String::with_capacity(trim_len);
-                        for (i, c) in sliced.char_indices() {
-                            if i < replacement_point {
-                                r.push(c)
-                            } else {
-                                r.push(STATS_REPLACEMENT_CHAR);
-                            }
+                    // rebuild the string (since the len of result string is rather small)
+                    let mut r = String::with_capacity(trim_len);
+                    for (i, c) in sliced.char_indices() {
+                        if i < replacement_point {
+                            r.push(c)
+                        } else {
+                            r.push(STATS_REPLACEMENT_CHAR);
                         }
-
-                        Some(Scalar::String(r.into_bytes()))
                     }
+
+                    Some(Scalar::String(r))
                 }
-                Err(_) => {
-                    // if failed to convert the bytes into (utf-8)string, just ignore it.
-                    None
-                }
-            },
+            }
             v => Some(v),
         }
     }
diff --git a/src/query/storages/fuse/src/table_functions/clustering_information/clustering_information.rs b/src/query/storages/fuse/src/table_functions/clustering_information/clustering_information.rs
index 9fe14417f126..c88da1bd7802 100644
--- a/src/query/storages/fuse/src/table_functions/clustering_information/clustering_information.rs
+++ b/src/query/storages/fuse/src/table_functions/clustering_information/clustering_information.rs
@@ -218,7 +218,7 @@ impl<'a> ClusteringInformation<'a> {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String(cluster_key.as_bytes().to_vec())),
+                    Value::Scalar(Scalar::String(cluster_key.clone())),
                 ),
                 BlockEntry::new(
                     DataType::Number(NumberDataType::UInt64),
diff --git a/src/query/storages/fuse/src/table_functions/fuse_blocks/fuse_block.rs b/src/query/storages/fuse/src/table_functions/fuse_blocks/fuse_block.rs
index 7eac11930553..65653201fa7c 100644
--- a/src/query/storages/fuse/src/table_functions/fuse_blocks/fuse_block.rs
+++ b/src/query/storages/fuse/src/table_functions/fuse_blocks/fuse_block.rs
@@ -103,7 +103,7 @@ impl<'a> FuseBlock<'a> {
         let limit = self.limit.unwrap_or(usize::MAX);
         let len = std::cmp::min(snapshot.summary.block_count as usize, limit);
 
-        let snapshot_id = snapshot.snapshot_id.simple().to_string().into_bytes();
+        let snapshot_id = snapshot.snapshot_id.simple().to_string();
         let timestamp = snapshot.timestamp.unwrap_or_default().timestamp_micros();
         let mut block_location = StringColumnBuilder::with_capacity(len, len);
         let mut block_size = Vec::with_capacity(len);
@@ -131,7 +131,7 @@ impl<'a> FuseBlock<'a> {
 
                 for block in segment.blocks.iter() {
                     let block = block.as_ref();
-                    block_location.put_slice(block.location.0.as_bytes());
+                    block_location.put_str(&block.location.0);
                     block_location.commit_row();
                     block_size.push(block.block_size);
                     file_size.push(block.file_size);
@@ -140,7 +140,7 @@ impl<'a> FuseBlock<'a> {
                         block
                             .bloom_filter_index_location
                             .as_ref()
-                            .map(|s| s.0.as_bytes().to_vec()),
+                            .map(|s| s.0.clone()),
                     );
                     bloom_filter_size.push(block.bloom_filter_index_size);
 
@@ -159,10 +159,7 @@ impl<'a> FuseBlock<'a> {
 
         Ok(DataBlock::new(
             vec![
-                BlockEntry::new(
-                    DataType::String,
-                    Value::Scalar(Scalar::String(snapshot_id.to_vec())),
-                ),
+                BlockEntry::new(DataType::String, Value::Scalar(Scalar::String(snapshot_id))),
                 BlockEntry::new(
                     DataType::Timestamp,
                     Value::Scalar(Scalar::Timestamp(timestamp)),
diff --git a/src/query/storages/fuse/src/table_functions/fuse_columns/fuse_column.rs b/src/query/storages/fuse/src/table_functions/fuse_columns/fuse_column.rs
index 6df0a169e8e0..b4082995abbd 100644
--- a/src/query/storages/fuse/src/table_functions/fuse_columns/fuse_column.rs
+++ b/src/query/storages/fuse/src/table_functions/fuse_columns/fuse_column.rs
@@ -103,7 +103,7 @@ impl<'a> FuseColumn<'a> {
         let limit = self.limit.unwrap_or(usize::MAX);
         let len = std::cmp::min(snapshot.summary.block_count as usize, limit);
 
-        let snapshot_id = snapshot.snapshot_id.simple().to_string().into_bytes();
+        let snapshot_id = snapshot.snapshot_id.simple().to_string();
         let timestamp = snapshot.timestamp.unwrap_or_default().timestamp_micros();
         let mut block_location = StringColumnBuilder::with_capacity(len, len);
         let mut block_size = vec![];
@@ -141,16 +141,16 @@ impl<'a> FuseColumn<'a> {
 
                     for (id, column) in block.col_metas.iter() {
                         if let Some(f) = leaf_fields.iter().find(|f| f.column_id == *id) {
-                            block_location.put_slice(block.location.0.as_bytes());
+                            block_location.put_str(&block.location.0);
                             block_location.commit_row();
                             block_size.push(block.block_size);
                             file_size.push(block.file_size);
                             row_count.push(column.total_rows() as u64);
 
-                            column_name.put_slice(f.name.as_bytes());
+                            column_name.put_str(&f.name);
                             column_name.commit_row();
 
-                            column_type.put_slice(f.data_type.to_string().as_bytes());
+                            column_type.put_str(&f.data_type.to_string());
                             column_type.commit_row();
 
                             column_id.push(*id);
@@ -177,10 +177,7 @@ impl<'a> FuseColumn<'a> {
 
         Ok(DataBlock::new(
             vec![
-                BlockEntry::new(
-                    DataType::String,
-                    Value::Scalar(Scalar::String(snapshot_id.to_vec())),
-                ),
+                BlockEntry::new(DataType::String, Value::Scalar(Scalar::String(snapshot_id))),
                 BlockEntry::new(
                     DataType::Timestamp,
                     Value::Scalar(Scalar::Timestamp(timestamp)),
diff --git a/src/query/storages/fuse/src/table_functions/fuse_encodings/fuse_encoding.rs b/src/query/storages/fuse/src/table_functions/fuse_encodings/fuse_encoding.rs
index 93bbfc0aca2e..fd88c487e9bd 100644
--- a/src/query/storages/fuse/src/table_functions/fuse_encodings/fuse_encoding.rs
+++ b/src/query/storages/fuse/src/table_functions/fuse_encodings/fuse_encoding.rs
@@ -177,15 +177,14 @@ impl<'a> FuseEncoding<'a> {
                 validity_size.reserve(num_row);
                 compressed_size.reserve(num_row);
                 uncompressed_size.reserve(num_row);
-                let tmp_table_name = StringColumnBuilder::repeat(table.as_bytes(), num_row);
-                let tmp_column_name =
-                    StringColumnBuilder::repeat(column_info.field.name.as_bytes(), num_row);
-                let tmp_column_type = StringColumnBuilder::repeat(type_str.as_bytes(), num_row);
+                let tmp_table_name = StringColumnBuilder::repeat(table, num_row);
+                let tmp_column_name = StringColumnBuilder::repeat(&column_info.field.name, num_row);
+                let tmp_column_type = StringColumnBuilder::repeat(type_str, num_row);
                 for p in pages_info {
                     validity_size.push(p.validity_size);
                     compressed_size.push(p.compressed_size);
                     uncompressed_size.push(p.uncompressed_size);
-                    l1.put_slice(encoding_to_string(&p.body).as_bytes());
+                    l1.put_str(&encoding_to_string(&p.body));
                     l1.commit_row();
                     let l2_encoding = match &p.body {
                         PageBody::Dict(dict) => Some(encoding_to_string(&dict.indices.body)),
@@ -196,7 +195,7 @@ impl<'a> FuseEncoding<'a> {
                         _ => None,
                     };
                     if let Some(l2_encoding) = l2_encoding {
-                        l2.push(l2_encoding.as_bytes());
+                        l2.push(&l2_encoding);
                     } else {
                         l2.push_null();
                     }
diff --git a/src/query/storages/fuse/src/table_functions/fuse_segments/fuse_segment.rs b/src/query/storages/fuse/src/table_functions/fuse_segments/fuse_segment.rs
index 75d1cff25898..8c109b65251d 100644
--- a/src/query/storages/fuse/src/table_functions/fuse_segments/fuse_segment.rs
+++ b/src/query/storages/fuse/src/table_functions/fuse_segments/fuse_segment.rs
@@ -101,7 +101,7 @@ impl<'a> FuseSegment<'a> {
         let mut row_count: Vec<u64> = Vec::with_capacity(len);
         let mut compressed: Vec<u64> = Vec::with_capacity(len);
         let mut uncompressed: Vec<u64> = Vec::with_capacity(len);
-        let mut file_location: Vec<Vec<u8>> = Vec::with_capacity(len);
+        let mut file_location: Vec<String> = Vec::with_capacity(len);
 
         let segments_io = SegmentsIO::create(
             self.ctx.clone(),
@@ -125,7 +125,7 @@ impl<'a> FuseSegment<'a> {
                 row_count.push(segment.summary.row_count);
                 compressed.push(segment.summary.compressed_byte_size);
                 uncompressed.push(segment.summary.uncompressed_byte_size);
-                file_location.push(segment_locations[idx].0.clone().into_bytes());
+                file_location.push(segment_locations[idx].0.clone());
 
                 row_num += 1;
                 if row_num >= limit {
diff --git a/src/query/storages/fuse/src/table_functions/fuse_snapshots/fuse_snapshot.rs b/src/query/storages/fuse/src/table_functions/fuse_snapshots/fuse_snapshot.rs
index 406386f7245f..9c88c36249f9 100644
--- a/src/query/storages/fuse/src/table_functions/fuse_snapshots/fuse_snapshot.rs
+++ b/src/query/storages/fuse/src/table_functions/fuse_snapshots/fuse_snapshot.rs
@@ -93,9 +93,9 @@ impl<'a> FuseSnapshot<'a> {
         latest_snapshot_version: u64,
     ) -> Result<DataBlock> {
         let len = snapshots.len();
-        let mut snapshot_ids: Vec<Vec<u8>> = Vec::with_capacity(len);
-        let mut snapshot_locations: Vec<Vec<u8>> = Vec::with_capacity(len);
-        let mut prev_snapshot_ids: Vec<Option<Vec<u8>>> = Vec::with_capacity(len);
+        let mut snapshot_ids: Vec<String> = Vec::with_capacity(len);
+        let mut snapshot_locations: Vec<String> = Vec::with_capacity(len);
+        let mut prev_snapshot_ids: Vec<Option<String>> = Vec::with_capacity(len);
         let mut format_versions: Vec<u64> = Vec::with_capacity(len);
         let mut segment_count: Vec<u64> = Vec::with_capacity(len);
         let mut block_count: Vec<u64> = Vec::with_capacity(len);
@@ -106,15 +106,14 @@ impl<'a> FuseSnapshot<'a> {
         let mut timestamps: Vec<Option<i64>> = Vec::with_capacity(len);
         let mut current_snapshot_version = latest_snapshot_version;
         for s in snapshots {
-            snapshot_ids.push(s.snapshot_id.simple().to_string().into_bytes());
+            snapshot_ids.push(s.snapshot_id.simple().to_string());
             snapshot_locations.push(
                 location_generator
-                    .snapshot_location_from_uuid(&s.snapshot_id, current_snapshot_version)?
-                    .into_bytes(),
+                    .snapshot_location_from_uuid(&s.snapshot_id, current_snapshot_version)?,
             );
-            let (id, ver) = s.prev_snapshot_id.map_or((None, 0), |(id, v)| {
-                (Some(id.simple().to_string().into_bytes()), v)
-            });
+            let (id, ver) = s
+                .prev_snapshot_id
+                .map_or((None, 0), |(id, v)| (Some(id.simple().to_string()), v));
             prev_snapshot_ids.push(id);
             format_versions.push(s.format_version);
             segment_count.push(s.segment_count);
diff --git a/src/query/storages/fuse/src/table_functions/fuse_statistics/fuse_statistic.rs b/src/query/storages/fuse/src/table_functions/fuse_statistics/fuse_statistic.rs
index 960e106acd75..bc2d33073824 100644
--- a/src/query/storages/fuse/src/table_functions/fuse_statistics/fuse_statistic.rs
+++ b/src/query/storages/fuse/src/table_functions/fuse_statistics/fuse_statistic.rs
@@ -58,13 +58,13 @@ impl<'a> FuseStatistic<'a> {
         _summy: &Statistics,
         table_statistics: &Option<Arc<TableSnapshotStatistics>>,
     ) -> Result<DataBlock> {
-        let mut col_ndvs: Vec<Vec<u8>> = Vec::with_capacity(1);
+        let mut col_ndvs: Vec<String> = Vec::with_capacity(1);
         if let Some(table_statistics) = table_statistics {
             let mut ndvs: String = "".to_string();
             for (i, n) in table_statistics.column_distinct_values.iter() {
                 ndvs.push_str(&format!("({},{});", *i, *n));
             }
-            col_ndvs.push(ndvs.into_bytes());
+            col_ndvs.push(ndvs);
         };
 
         Ok(DataBlock::new_from_columns(vec![StringType::from_data(
diff --git a/src/query/storages/fuse/src/table_functions/table_args.rs b/src/query/storages/fuse/src/table_functions/table_args.rs
index 47d083636e0b..e15fd0cde2be 100644
--- a/src/query/storages/fuse/src/table_functions/table_args.rs
+++ b/src/query/storages/fuse/src/table_functions/table_args.rs
@@ -23,14 +23,13 @@ use crate::table_functions::TableArgs;
 
 pub fn string_value(value: &Scalar) -> Result<String> {
     match value {
-        Scalar::String(val) => String::from_utf8(val.clone())
-            .map_err(|e| ErrorCode::BadArguments(format!("invalid string. {}", e))),
+        Scalar::String(val) => Ok(val.clone()),
         _ => Err(ErrorCode::BadArguments("invalid string.")),
     }
 }
 
 pub fn string_literal(val: &str) -> Scalar {
-    Scalar::String(val.as_bytes().to_vec())
+    Scalar::String(val.to_string())
 }
 
 pub fn cmp_with_null(v1: &Scalar, v2: &Scalar) -> Ordering {
diff --git a/src/query/storages/hive/hive/src/hive_block_filter.rs b/src/query/storages/hive/hive/src/hive_block_filter.rs
index 8df4764e8dbb..9414bf48b695 100644
--- a/src/query/storages/hive/hive/src/hive_block_filter.rs
+++ b/src/query/storages/hive/hive/src/hive_block_filter.rs
@@ -103,7 +103,7 @@ impl HiveBlockFilter {
                         null_count = row_group.num_rows();
                         Scalar::Null
                     } else {
-                        Scalar::String(p_value.as_bytes().to_vec())
+                        Scalar::String(p_value)
                     };
 
                     let col_stats = ColumnStatistics::new(v.clone(), v, null_count as u64, 0, None);
@@ -282,8 +282,12 @@ impl HiveBlockFilter {
                     None
                 } else {
                     let null_count = s.null_count.unwrap();
-                    let max = StringType::upcast_scalar(s.max_value.clone().unwrap());
-                    let min = StringType::upcast_scalar(s.min_value.clone().unwrap());
+                    let max = StringType::upcast_scalar(
+                        String::from_utf8(s.max_value.clone().unwrap()).ok()?,
+                    );
+                    let min = StringType::upcast_scalar(
+                        String::from_utf8(s.min_value.clone().unwrap()).ok()?,
+                    );
                     Some((max, min, null_count))
                 }
             }
diff --git a/src/query/storages/hive/hive/src/hive_table_source.rs b/src/query/storages/hive/hive/src/hive_table_source.rs
index 51ebb86d388f..e27488d22bb1 100644
--- a/src/query/storages/hive/hive/src/hive_table_source.rs
+++ b/src/query/storages/hive/hive/src/hive_table_source.rs
@@ -37,6 +37,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use log::debug;
 use opendal::Operator;
 
@@ -179,6 +181,7 @@ impl HiveTableSource {
             rows: prewhere_datablocks.iter().map(|x| x.num_rows()).sum(),
             bytes: prewhere_datablocks.iter().map(|x| x.memory_size()).sum(),
         };
+        Profile::record_usize_profile(ProfileStatisticsName::ScanBytes, progress_values.bytes);
         self.scan_progress.incr(&progress_values);
 
         if let Some(filter) = self.prewhere_filter.as_ref() {
diff --git a/src/query/storages/hive/hive/src/utils.rs b/src/query/storages/hive/hive/src/utils.rs
index c2b96cf1146e..6312dd015919 100644
--- a/src/query/storages/hive/hive/src/utils.rs
+++ b/src/query/storages/hive/hive/src/utils.rs
@@ -31,7 +31,7 @@ pub(crate) fn str_field_to_scalar(value: &str, data_type: &DataType) -> Result<S
                 str_field_to_scalar(value, c.as_ref())
             }
         }
-        DataType::String => Ok(Scalar::String(value.as_bytes().to_vec())),
+        DataType::String => Ok(Scalar::String(value.to_string())),
         DataType::Number(num_ty) => match num_ty {
             NumberDataType::UInt8 => {
                 let num = value.parse::<u8>().unwrap();
diff --git a/src/query/storages/iceberg/src/stats.rs b/src/query/storages/iceberg/src/stats.rs
index 4a53e2ba88c7..8c9bb2785d1d 100644
--- a/src/query/storages/iceberg/src/stats.rs
+++ b/src/query/storages/iceberg/src/stats.rs
@@ -107,7 +107,7 @@ fn parse_binary_value(ty: &TableDataType, data: &[u8]) -> Option<Scalar> {
             let v = i64::from_le_bytes(data.try_into().ok()?);
             Some(Scalar::Timestamp(v))
         }
-        TableDataType::String => Some(Scalar::String(data.to_vec())),
+        TableDataType::String => Some(Scalar::String(String::from_utf8(data.to_vec()).ok()?)),
         // TODO: support Decimal.
         _ => None, // Not supported.
     }
diff --git a/src/query/storages/iceberg/src/table_source.rs b/src/query/storages/iceberg/src/table_source.rs
index a980b7368dd5..7377d1cbe9b5 100644
--- a/src/query/storages/iceberg/src/table_source.rs
+++ b/src/query/storages/iceberg/src/table_source.rs
@@ -27,6 +27,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_storages_parquet::ParquetPart;
 use databend_common_storages_parquet::ParquetRSFullReader;
 use opendal::Reader;
@@ -102,6 +104,10 @@ impl Processor for IcebergTableSource {
                     bytes: data_block.memory_size(),
                 };
                 self.scan_progress.incr(&progress_values);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::ScanBytes,
+                    data_block.memory_size(),
+                );
                 self.output.push_data(Ok(data_block));
                 Ok(Event::NeedConsume)
             }
diff --git a/src/query/storages/parquet/src/parquet2/parquet_table/create.rs b/src/query/storages/parquet/src/parquet2/parquet_table/create.rs
index 7302ce402a33..ad2c69a1d211 100644
--- a/src/query/storages/parquet/src/parquet2/parquet_table/create.rs
+++ b/src/query/storages/parquet/src/parquet2/parquet_table/create.rs
@@ -58,7 +58,10 @@ impl Parquet2Table {
 
         // If the query is `COPY`, we don't need to collect column statistics.
         // It's because the only transform could be contained in `COPY` command is projection.
-        let need_stats_provider = !matches!(ctx.get_query_kind(), QueryKind::CopyIntoTable);
+        let need_stats_provider = !matches!(
+            ctx.get_query_kind(),
+            QueryKind::CopyIntoTable | QueryKind::CopyIntoLocation
+        );
         let mut table_info = create_parquet_table_info(arrow_schema.clone(), &stage_info);
         let column_statistics_provider = if need_stats_provider {
             let file_metas = get_parquet2_file_meta(
diff --git a/src/query/storages/parquet/src/parquet2/processors/deserialize_transform.rs b/src/query/storages/parquet/src/parquet2/processors/deserialize_transform.rs
index e80f1d70a243..7ab94a0d1fa7 100644
--- a/src/query/storages/parquet/src/parquet2/processors/deserialize_transform.rs
+++ b/src/query/storages/parquet/src/parquet2/processors/deserialize_transform.rs
@@ -42,6 +42,8 @@ use databend_common_pipeline_core::processors::InputPort;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_storage::CopyStatus;
 use databend_common_storage::FileStatus;
 use opendal::services::Memory;
@@ -153,6 +155,7 @@ impl Parquet2DeserializeTransform {
             bytes: data_block.memory_size(),
         };
         self.scan_progress.incr(&progress_values);
+        Profile::record_usize_profile(ProfileStatisticsName::ScanBytes, data_block.memory_size());
         self.output_data.push(data_block);
         Ok(())
     }
diff --git a/src/query/storages/parquet/src/parquet_rs/copy_into_table/source.rs b/src/query/storages/parquet/src/parquet_rs/copy_into_table/source.rs
index 0ff6b85dbde2..d1c2206acc98 100644
--- a/src/query/storages/parquet/src/parquet_rs/copy_into_table/source.rs
+++ b/src/query/storages/parquet/src/parquet_rs/copy_into_table/source.rs
@@ -30,6 +30,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use opendal::Operator;
 
 use crate::parquet_rs::copy_into_table::reader::RowGroupReaderForCopy;
@@ -128,6 +130,10 @@ impl Processor for ParquetCopySource {
                     bytes: data_block.memory_size(),
                 };
                 self.scan_progress.incr(&progress_values);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::ScanBytes,
+                    data_block.memory_size(),
+                );
                 self.output.push_data(Ok(data_block));
                 Ok(Event::NeedConsume)
             }
diff --git a/src/query/storages/parquet/src/parquet_rs/parquet_table/table.rs b/src/query/storages/parquet/src/parquet_rs/parquet_table/table.rs
index 6f92f91da583..be130f82a16c 100644
--- a/src/query/storages/parquet/src/parquet_rs/parquet_table/table.rs
+++ b/src/query/storages/parquet/src/parquet_rs/parquet_table/table.rs
@@ -135,7 +135,10 @@ impl ParquetRSTable {
 
         // If the query is `COPY`, we don't need to collect column statistics.
         // It's because the only transform could be contained in `COPY` command is projection.
-        let need_stats_provider = !matches!(ctx.get_query_kind(), QueryKind::CopyIntoTable);
+        let need_stats_provider = !matches!(
+            ctx.get_query_kind(),
+            QueryKind::CopyIntoTable | QueryKind::CopyIntoLocation
+        );
         let settings = ctx.get_settings();
         let max_threads = settings.get_max_threads()? as usize;
         let max_memory_usage = settings.get_max_memory_usage()?;
diff --git a/src/query/storages/parquet/src/parquet_rs/source.rs b/src/query/storages/parquet/src/parquet_rs/source.rs
index 2fb717009483..4acacadde27c 100644
--- a/src/query/storages/parquet/src/parquet_rs/source.rs
+++ b/src/query/storages/parquet/src/parquet_rs/source.rs
@@ -28,6 +28,8 @@ use databend_common_pipeline_core::processors::Event;
 use databend_common_pipeline_core::processors::OutputPort;
 use databend_common_pipeline_core::processors::Processor;
 use databend_common_pipeline_core::processors::ProcessorPtr;
+use databend_common_pipeline_core::processors::Profile;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 use databend_common_storage::CopyStatus;
 use databend_common_storage::FileStatus;
 
@@ -136,6 +138,10 @@ impl Processor for ParquetSource {
                     bytes: data_block.memory_size(),
                 };
                 self.scan_progress.incr(&progress_values);
+                Profile::record_usize_profile(
+                    ProfileStatisticsName::ScanBytes,
+                    data_block.memory_size(),
+                );
                 self.output.push_data(Ok(data_block));
                 Ok(Event::NeedConsume)
             }
diff --git a/src/query/storages/parquet/src/parquet_rs/statistics/column.rs b/src/query/storages/parquet/src/parquet_rs/statistics/column.rs
index 416c12022c4e..96bdde5ecfc9 100644
--- a/src/query/storages/parquet/src/parquet_rs/statistics/column.rs
+++ b/src/query/storages/parquet/src/parquet_rs/statistics/column.rs
@@ -27,7 +27,7 @@ use super::utils::decode_decimal128_from_bytes;
 use super::utils::decode_decimal256_from_bytes;
 
 /// according to https://github.com/apache/parquet-format/blob/master/LogicalTypes.md
-pub fn convert_column_statistics(s: &Statistics, typ: &TableDataType) -> ColumnStatistics {
+pub fn convert_column_statistics(s: &Statistics, typ: &TableDataType) -> Option<ColumnStatistics> {
     let (max, min) = if s.has_min_max_set() {
         match s {
             Statistics::Boolean(s) => (Scalar::Boolean(*s.max()), Scalar::Boolean(*s.min())),
@@ -110,8 +110,8 @@ pub fn convert_column_statistics(s: &Statistics, typ: &TableDataType) -> ColumnS
             Statistics::Float(s) => (Scalar::from(*s.max()), Scalar::from(*s.min())),
             Statistics::Double(s) => (Scalar::from(*s.max()), Scalar::from(*s.min())),
             Statistics::ByteArray(s) => (
-                Scalar::String(s.max().as_bytes().to_vec()),
-                Scalar::String(s.min().as_bytes().to_vec()),
+                Scalar::String(String::from_utf8(s.max().as_bytes().to_vec()).ok()?),
+                Scalar::String(String::from_utf8(s.min().as_bytes().to_vec()).ok()?),
             ),
             Statistics::FixedLenByteArray(s) => {
                 let (max, min) = (s.max(), s.min());
@@ -131,11 +131,11 @@ pub fn convert_column_statistics(s: &Statistics, typ: &TableDataType) -> ColumnS
     } else {
         (Scalar::Null, Scalar::Null)
     };
-    ColumnStatistics::new(
+    Some(ColumnStatistics::new(
         min,
         max,
         s.null_count(),
         0, // this field is not used.
         s.distinct_count(),
-    )
+    ))
 }
diff --git a/src/query/storages/parquet/src/parquet_rs/statistics/page.rs b/src/query/storages/parquet/src/parquet_rs/statistics/page.rs
index cb449ad944a4..0ae34f62bab0 100644
--- a/src/query/storages/parquet/src/parquet_rs/statistics/page.rs
+++ b/src/query/storages/parquet/src/parquet_rs/statistics/page.rs
@@ -246,8 +246,8 @@ fn convert_page_index_byte_array(
 ) -> Option<ColumnStatistics> {
     match (&index.min, &index.max, index.null_count) {
         (Some(min), Some(max), Some(null_count)) => Some(ColumnStatistics::new(
-            Scalar::String(min.as_bytes().to_vec()),
-            Scalar::String(max.as_bytes().to_vec()),
+            Scalar::String(String::from_utf8(min.as_bytes().to_vec()).ok()?),
+            Scalar::String(String::from_utf8(max.as_bytes().to_vec()).ok()?),
             null_count as u64,
             0,
             None,
diff --git a/src/query/storages/parquet/src/parquet_rs/statistics/row_group.rs b/src/query/storages/parquet/src/parquet_rs/statistics/row_group.rs
index e9ab006dd48e..f1bac90e2a35 100644
--- a/src/query/storages/parquet/src/parquet_rs/statistics/row_group.rs
+++ b/src/query/storages/parquet/src/parquet_rs/statistics/row_group.rs
@@ -53,7 +53,7 @@ pub fn collect_row_group_stats(
                 let column_stats = column.statistics().unwrap();
                 stats_of_columns.insert(
                     *col_idx as u32,
-                    convert_column_statistics(column_stats, &field.data_type().remove_nullable()),
+                    convert_column_statistics(column_stats, &field.data_type().remove_nullable())?,
                 );
             }
         } else {
@@ -63,7 +63,7 @@ pub fn collect_row_group_stats(
                 let column_stats = column.statistics().unwrap();
                 stats_of_columns.insert(
                     col_idx as u32,
-                    convert_column_statistics(column_stats, &field.data_type().remove_nullable()),
+                    convert_column_statistics(column_stats, &field.data_type().remove_nullable())?,
                 );
             }
         }
diff --git a/src/query/storages/result_cache/src/table_function/table.rs b/src/query/storages/result_cache/src/table_function/table.rs
index 9f3a542d55e0..06ece0903236 100644
--- a/src/query/storages/result_cache/src/table_function/table.rs
+++ b/src/query/storages/result_cache/src/table_function/table.rs
@@ -115,7 +115,7 @@ impl Table for ResultScan {
     }
 
     fn table_args(&self) -> Option<TableArgs> {
-        let args = vec![Scalar::String(self.query_id.as_bytes().to_vec())];
+        let args = vec![Scalar::String(self.query_id.clone())];
 
         Some(TableArgs::new_positioned(args))
     }
diff --git a/src/query/storages/stage/src/append/output.rs b/src/query/storages/stage/src/append/output.rs
index b7ba6ea4775b..c5b41fd3324f 100644
--- a/src/query/storages/stage/src/append/output.rs
+++ b/src/query/storages/stage/src/append/output.rs
@@ -154,7 +154,7 @@ fn file_infos_to_block(files: &[OutputFileInfo]) -> DataBlock {
     let mut rows = Vec::with_capacity(files.len());
     let mut sizes = Vec::with_capacity(files.len());
     for file in files {
-        paths.push(file.file_name.clone().as_bytes().to_vec());
+        paths.push(file.file_name.as_str());
         rows.push(file.summary.row_counts as u64);
         sizes.push(file.summary.output_bytes as u64);
     }
diff --git a/src/query/storages/system/Cargo.toml b/src/query/storages/system/Cargo.toml
index 27afdc5eadd9..9c33b9c157ab 100644
--- a/src/query/storages/system/Cargo.toml
+++ b/src/query/storages/system/Cargo.toml
@@ -26,7 +26,6 @@ databend-common-meta-types = { path = "../../../meta/types" }
 databend-common-metrics = { path = "../../../common/metrics" }
 databend-common-pipeline-core = { path = "../../pipeline/core" }
 databend-common-pipeline-sources = { path = "../../pipeline/sources" }
-databend-common-profile = { path = "../../profile" }
 databend-common-sql = { path = "../../sql" }
 databend-common-storage = { path = "../../../common/storage" }
 databend-common-storages-fuse = { path = "../fuse" }
diff --git a/src/query/storages/system/src/background_jobs_table.rs b/src/query/storages/system/src/background_jobs_table.rs
index 1d708850dc7e..e43f403abedb 100644
--- a/src/query/storages/system/src/background_jobs_table.rs
+++ b/src/query/storages/system/src/background_jobs_table.rs
@@ -76,10 +76,10 @@ impl AsyncSystemTable for BackgroundJobTable {
         let mut creator = Vec::with_capacity(jobs.len());
         let mut create_time = Vec::with_capacity(jobs.len());
         for (_, name, job) in jobs {
-            names.push(name.as_bytes().to_vec());
+            names.push(name);
             let job_type = job.job_params.as_ref().map(|x| x.job_type.clone());
             if let Some(job_type) = job_type {
-                job_types.push(Some(job_type.to_string().as_bytes().to_vec()));
+                job_types.push(Some(job_type.to_string()));
                 match job_type {
                     BackgroundJobType::INTERVAL => {
                         scheduled_job_interval_secs.push(Some(
@@ -95,19 +95,13 @@ impl AsyncSystemTable for BackgroundJobTable {
                     BackgroundJobType::CRON => {
                         scheduled_job_interval_secs.push(None);
                         scheduled_job_cron_expression.push(Some(
-                            job.job_params
-                                .as_ref()
-                                .unwrap()
-                                .scheduled_job_cron
-                                .clone()
-                                .as_bytes()
-                                .to_vec(),
+                            job.job_params.as_ref().unwrap().scheduled_job_cron.clone(),
                         ));
                         scheduled_job_cron_timezone.push(
                             job.job_params
                                 .unwrap()
                                 .scheduled_job_timezone
-                                .map(|tz| tz.to_string().as_bytes().to_vec()),
+                                .map(|tz| tz.to_string()),
                         );
                     }
                     BackgroundJobType::ONESHOT => {
@@ -122,17 +116,9 @@ impl AsyncSystemTable for BackgroundJobTable {
                 scheduled_job_cron_expression.push(None);
                 scheduled_job_cron_timezone.push(None);
             }
-            task_types.push(job.task_type.to_string().as_bytes().to_vec());
-            job_states.push(
-                job.job_status
-                    .as_ref()
-                    .map(|x| x.job_state.to_string().as_bytes().to_vec()),
-            );
-            last_task_ids.push(
-                job.job_status
-                    .as_ref()
-                    .and_then(|x| x.last_task_id.clone().map(|x| x.as_bytes().to_vec())),
-            );
+            task_types.push(job.task_type.to_string());
+            job_states.push(job.job_status.as_ref().map(|x| x.job_state.to_string()));
+            last_task_ids.push(job.job_status.as_ref().and_then(|x| x.last_task_id.clone()));
             last_task_run_at.push(
                 job.job_status
                     .as_ref()
@@ -143,9 +129,9 @@ impl AsyncSystemTable for BackgroundJobTable {
                     .as_ref()
                     .and_then(|x| x.next_task_scheduled_time.map(|x| x.timestamp_micros())),
             );
-            message.push(job.message.as_bytes().to_vec());
+            message.push(job.message);
             last_updated.push(job.last_updated.map(|t| t.timestamp_micros()));
-            creator.push(job.creator.map(|x| x.to_string().as_bytes().to_vec()));
+            creator.push(job.creator.map(|x| x.to_string()));
             create_time.push(job.created_at.timestamp_micros());
         }
 
diff --git a/src/query/storages/system/src/background_tasks_table.rs b/src/query/storages/system/src/background_tasks_table.rs
index 0b006941931a..b0e80364ce3c 100644
--- a/src/query/storages/system/src/background_tasks_table.rs
+++ b/src/query/storages/system/src/background_tasks_table.rs
@@ -75,10 +75,10 @@ impl AsyncSystemTable for BackgroundTaskTable {
         let mut create_timestamps = Vec::with_capacity(tasks.len());
         let mut update_timestamps = Vec::with_capacity(tasks.len());
         for (_, name, task) in tasks {
-            names.push(name.as_bytes().to_vec());
-            types.push(task.task_type.to_string().as_bytes().to_vec());
-            stats.push(task.task_state.to_string().as_bytes().to_vec());
-            messages.push(task.message.as_bytes().to_vec());
+            names.push(name);
+            types.push(task.task_type.to_string());
+            stats.push(task.task_state.to_string());
+            messages.push(task.message);
             compaction_stats.push(
                 task.compaction_task_stats
                     .as_ref()
@@ -98,11 +98,8 @@ impl AsyncSystemTable for BackgroundTaskTable {
                 table_ids.push(0);
                 task_run_secs.push(None);
             }
-            creators.push(task.creator.map(|s| s.to_string().as_bytes().to_vec()));
-            trigger.push(
-                task.manual_trigger
-                    .map(|s| s.trigger.to_string().as_bytes().to_vec()),
-            );
+            creators.push(task.creator.map(|s| s.to_string()));
+            trigger.push(task.manual_trigger.map(|s| s.trigger.to_string()));
             create_timestamps.push(task.created_at.timestamp_micros());
             update_timestamps.push(task.last_updated.unwrap_or_default().timestamp_micros());
         }
diff --git a/src/query/storages/system/src/backtrace_table.rs b/src/query/storages/system/src/backtrace_table.rs
index 6ec64a5b82d9..311d003ff1f1 100644
--- a/src/query/storages/system/src/backtrace_table.rs
+++ b/src/query/storages/system/src/backtrace_table.rs
@@ -51,16 +51,19 @@ impl SyncSystemTable for BacktraceTable {
     }
 
     fn get_full_data(&self, ctx: Arc<dyn TableContext>) -> Result<DataBlock> {
-        let local_node = ctx.get_cluster().local_id.clone().into_bytes();
+        let local_node = ctx.get_cluster().local_id.clone();
         let (tasks, polling_tasks) = get_all_tasks(false);
         let tasks_size = tasks.len() + polling_tasks.len();
 
-        let mut nodes: Vec<Vec<u8>> = Vec::with_capacity(tasks_size);
-        let mut queries_id: Vec<Vec<u8>> = Vec::with_capacity(tasks_size);
-        let mut queries_status: Vec<Vec<u8>> = Vec::with_capacity(tasks_size);
-        let mut stacks: Vec<Vec<u8>> = Vec::with_capacity(tasks_size);
+        let mut nodes: Vec<String> = Vec::with_capacity(tasks_size);
+        let mut queries_id: Vec<String> = Vec::with_capacity(tasks_size);
+        let mut queries_status: Vec<String> = Vec::with_capacity(tasks_size);
+        let mut stacks: Vec<String> = Vec::with_capacity(tasks_size);
 
-        for (status, mut tasks) in [("PENDING", tasks), ("RUNNING", polling_tasks)] {
+        for (status, mut tasks) in [
+            ("PENDING".to_string(), tasks),
+            ("RUNNING".to_string(), polling_tasks),
+        ] {
             tasks.sort_by(|l, r| Ord::cmp(&l.stack_frames.len(), &r.stack_frames.len()));
 
             for item in tasks.into_iter().rev() {
@@ -102,9 +105,9 @@ impl SyncSystemTable for BacktraceTable {
                 }
 
                 nodes.push(local_node.clone());
-                stacks.push(stack_frames.into_bytes());
-                queries_id.push(query_id.into_bytes());
-                queries_status.push(status.as_bytes().to_vec());
+                stacks.push(stack_frames);
+                queries_id.push(query_id);
+                queries_status.push(status.clone());
             }
         }
 
diff --git a/src/query/storages/system/src/build_options_table.rs b/src/query/storages/system/src/build_options_table.rs
index 0de2cb47fda3..6d1953a3245f 100644
--- a/src/query/storages/system/src/build_options_table.rs
+++ b/src/query/storages/system/src/build_options_table.rs
@@ -43,26 +43,26 @@ impl SyncSystemTable for BuildOptionsTable {
     }
 
     fn get_full_data(&self, _: Arc<dyn TableContext>) -> Result<DataBlock> {
-        let mut cargo_features: Vec<Vec<u8>>;
+        let mut cargo_features: Vec<String>;
 
         if let Some(features) = option_env!("VERGEN_CARGO_FEATURES") {
             cargo_features = features
                 .split_terminator(',')
-                .map(|x| x.trim().as_bytes().to_vec())
+                .map(|x| x.trim().to_string())
                 .collect();
         } else {
-            cargo_features = vec!["not available".as_bytes().to_vec()];
+            cargo_features = vec!["not available".to_string()];
         }
 
-        let mut target_features: Vec<Vec<u8>> = env!("DATABEND_CARGO_CFG_TARGET_FEATURE")
+        let mut target_features: Vec<String> = env!("DATABEND_CARGO_CFG_TARGET_FEATURE")
             .split_terminator(',')
-            .map(|x| x.trim().as_bytes().to_vec())
+            .map(|x| x.trim().to_string())
             .collect();
 
         let length = max(cargo_features.len(), target_features.len());
 
-        cargo_features.resize(length, "".as_bytes().to_vec());
-        target_features.resize(length, "".as_bytes().to_vec());
+        cargo_features.resize(length, "".to_string());
+        target_features.resize(length, "".to_string());
 
         Ok(DataBlock::new_from_columns(vec![
             StringType::from_data(cargo_features),
diff --git a/src/query/storages/system/src/caches_table.rs b/src/query/storages/system/src/caches_table.rs
index aebb0ea006cb..848da0dfe0bb 100644
--- a/src/query/storages/system/src/caches_table.rs
+++ b/src/query/storages/system/src/caches_table.rs
@@ -68,68 +68,67 @@ impl SyncSystemTable for CachesTable {
         let table_column_array_cache = cache_manager.get_table_data_array_cache();
 
         if let Some(table_snapshot_cache) = table_snapshot_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("table_snapshot_cache");
+            nodes.push(local_node.clone());
+            names.push("table_snapshot_cache".to_string());
             num_items.push(table_snapshot_cache.len() as u64);
             size.push(table_snapshot_cache.size());
         }
         if let Some(table_snapshot_statistic_cache) = table_snapshot_statistic_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("table_snapshot_statistic_cache");
+            nodes.push(local_node.clone());
+            names.push("table_snapshot_statistic_cache".to_string());
             num_items.push(table_snapshot_statistic_cache.len() as u64);
             size.push(table_snapshot_statistic_cache.size());
         }
 
         if let Some(segment_info_cache) = segment_info_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("segment_info_cache");
+            nodes.push(local_node.clone());
+            names.push("segment_info_cache".to_string());
             num_items.push(segment_info_cache.len() as u64);
             size.push(segment_info_cache.size());
         }
 
         if let Some(bloom_index_filter_cache) = bloom_index_filter_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("bloom_index_filter_cache");
+            nodes.push(local_node.clone());
+            names.push("bloom_index_filter_cache".to_string());
             num_items.push(bloom_index_filter_cache.len() as u64);
             size.push(bloom_index_filter_cache.size());
         }
 
         if let Some(bloom_index_meta_cache) = bloom_index_meta_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("bloom_index_meta_cache");
+            nodes.push(local_node.clone());
+            names.push("bloom_index_meta_cache".to_string());
             num_items.push(bloom_index_meta_cache.len() as u64);
             size.push(bloom_index_meta_cache.size());
         }
 
         if let Some(prune_partitions_cache) = prune_partitions_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("prune_partitions_cache");
+            nodes.push(local_node.clone());
+            names.push("prune_partitions_cache".to_string());
             num_items.push(prune_partitions_cache.len() as u64);
             size.push(prune_partitions_cache.size());
         }
 
         if let Some(file_meta_data_cache) = file_meta_data_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("file_meta_data_cache");
+            nodes.push(local_node.clone());
+            names.push("file_meta_data_cache".to_string());
             num_items.push(file_meta_data_cache.len() as u64);
             size.push(file_meta_data_cache.size());
         }
 
         if let Some(table_data_cache) = table_data_cache {
-            nodes.push(local_node.clone().into_bytes());
-            names.push("table_data_cache");
+            nodes.push(local_node.clone());
+            names.push("table_data_cache".to_string());
             num_items.push(table_data_cache.len() as u64);
             size.push(table_data_cache.size());
         }
 
         if let Some(table_column_array_cache) = table_column_array_cache {
-            nodes.push(local_node.into_bytes());
-            names.push("table_column_array_cache");
+            nodes.push(local_node.clone());
+            names.push("table_column_array_cache".to_string());
             num_items.push(table_column_array_cache.len() as u64);
             size.push(table_column_array_cache.size());
         }
 
-        let names: Vec<_> = names.iter().map(|x| x.as_bytes().to_vec()).collect();
         Ok(DataBlock::new_from_columns(vec![
             StringType::from_data(nodes),
             StringType::from_data(names),
diff --git a/src/query/storages/system/src/catalogs_table.rs b/src/query/storages/system/src/catalogs_table.rs
index dbd94dfe4e58..f279ca115b00 100644
--- a/src/query/storages/system/src/catalogs_table.rs
+++ b/src/query/storages/system/src/catalogs_table.rs
@@ -56,7 +56,7 @@ impl AsyncSystemTable for CatalogsTable {
             .list_catalogs(&ctx.get_tenant())
             .await?
             .into_iter()
-            .map(|v| v.name().into_bytes())
+            .map(|v| v.name())
             .collect::<Vec<_>>();
 
         Ok(DataBlock::new_from_columns(vec![StringType::from_data(
diff --git a/src/query/storages/system/src/clustering_history_table.rs b/src/query/storages/system/src/clustering_history_table.rs
index 9831126eb8fd..d3223a334de1 100644
--- a/src/query/storages/system/src/clustering_history_table.rs
+++ b/src/query/storages/system/src/clustering_history_table.rs
@@ -65,11 +65,11 @@ impl SystemLogElement for ClusteringHistoryLogElement {
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.database.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.database.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.table.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.table.clone()).as_ref());
         columns
             .next()
             .unwrap()
diff --git a/src/query/storages/system/src/clusters_table.rs b/src/query/storages/system/src/clusters_table.rs
index f050035131c2..11573f347ec3 100644
--- a/src/query/storages/system/src/clusters_table.rs
+++ b/src/query/storages/system/src/clusters_table.rs
@@ -58,10 +58,10 @@ impl SyncSystemTable for ClustersTable {
         for cluster_node in &cluster_nodes {
             let (ip, port) = cluster_node.ip_port()?;
 
-            names.push(Scalar::String(cluster_node.id.as_bytes().to_vec()).as_ref());
-            addresses.push(Scalar::String(ip.as_bytes().to_vec()).as_ref());
+            names.push(Scalar::String(cluster_node.id.clone()).as_ref());
+            addresses.push(Scalar::String(ip).as_ref());
             addresses_port.push(Scalar::Number(NumberScalar::UInt16(port)).as_ref());
-            versions.push(Scalar::String(cluster_node.binary_version.as_bytes().to_vec()).as_ref());
+            versions.push(Scalar::String(cluster_node.binary_version.clone()).as_ref());
         }
 
         Ok(DataBlock::new_from_columns(vec![
diff --git a/src/query/storages/system/src/columns_table.rs b/src/query/storages/system/src/columns_table.rs
index b3e64031e955..bfb2ca0454ea 100644
--- a/src/query/storages/system/src/columns_table.rs
+++ b/src/query/storages/system/src/columns_table.rs
@@ -58,22 +58,22 @@ impl AsyncSystemTable for ColumnsTable {
         push_downs: Option<PushDownInfo>,
     ) -> Result<DataBlock> {
         let rows = self.dump_table_columns(ctx, push_downs).await?;
-        let mut names: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut tables: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut databases: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut types: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut data_types: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut default_kinds: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut default_exprs: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut is_nullables: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
-        let mut comments: Vec<Vec<u8>> = Vec::with_capacity(rows.len());
+        let mut names: Vec<String> = Vec::with_capacity(rows.len());
+        let mut tables: Vec<String> = Vec::with_capacity(rows.len());
+        let mut databases: Vec<String> = Vec::with_capacity(rows.len());
+        let mut types: Vec<String> = Vec::with_capacity(rows.len());
+        let mut data_types: Vec<String> = Vec::with_capacity(rows.len());
+        let mut default_kinds: Vec<String> = Vec::with_capacity(rows.len());
+        let mut default_exprs: Vec<String> = Vec::with_capacity(rows.len());
+        let mut is_nullables: Vec<String> = Vec::with_capacity(rows.len());
+        let mut comments: Vec<String> = Vec::with_capacity(rows.len());
         for (database_name, table_name, field) in rows.into_iter() {
-            names.push(field.name().clone().into_bytes());
-            tables.push(table_name.into_bytes());
-            databases.push(database_name.into_bytes());
-            types.push(field.data_type().wrapped_display().into_bytes());
+            names.push(field.name().clone());
+            tables.push(table_name);
+            databases.push(database_name);
+            types.push(field.data_type().wrapped_display());
             let data_type = field.data_type().remove_recursive_nullable().sql_name();
-            data_types.push(data_type.into_bytes());
+            data_types.push(data_type);
 
             let mut default_kind = "".to_string();
             let mut default_expr = "".to_string();
@@ -81,15 +81,15 @@ impl AsyncSystemTable for ColumnsTable {
                 default_kind = "DEFAULT".to_string();
                 default_expr = expr.to_string();
             }
-            default_kinds.push(default_kind.into_bytes());
-            default_exprs.push(default_expr.into_bytes());
+            default_kinds.push(default_kind);
+            default_exprs.push(default_expr);
             if field.is_nullable() {
-                is_nullables.push("YES".to_string().into_bytes());
+                is_nullables.push("YES".to_string());
             } else {
-                is_nullables.push("NO".to_string().into_bytes());
+                is_nullables.push("NO".to_string());
             }
 
-            comments.push("".to_string().into_bytes());
+            comments.push("".to_string());
         }
 
         Ok(DataBlock::new_from_columns(vec![
@@ -166,7 +166,7 @@ pub(crate) async fn dump_tables(
     let tenant = ctx.get_tenant();
     let catalog = ctx.get_catalog(CATALOG_DEFAULT).await?;
 
-    let mut tables = Vec::new();
+    let mut tables: Vec<String> = Vec::new();
     let mut databases: Vec<String> = Vec::new();
 
     if let Some(push_downs) = push_downs {
@@ -174,19 +174,15 @@ pub(crate) async fn dump_tables(
             let expr = filter.as_expr(&BUILTIN_FUNCTIONS);
             find_eq_filter(&expr, &mut |col_name, scalar| {
                 if col_name == "database" {
-                    if let Scalar::String(s) = scalar {
-                        if let Ok(database) = String::from_utf8(s.clone()) {
-                            if !databases.contains(&database) {
-                                databases.push(database);
-                            }
+                    if let Scalar::String(database) = scalar {
+                        if !databases.contains(database) {
+                            databases.push(database.clone());
                         }
                     }
                 } else if col_name == "table" {
-                    if let Scalar::String(s) = scalar {
-                        if let Ok(table) = String::from_utf8(s.clone()) {
-                            if !tables.contains(&table) {
-                                tables.push(table);
-                            }
+                    if let Scalar::String(table) = scalar {
+                        if !tables.contains(table) {
+                            tables.push(table.clone());
                         }
                     }
                 }
diff --git a/src/query/storages/system/src/configs_table.rs b/src/query/storages/system/src/configs_table.rs
index ee64b5b7609b..fb2f5b3e5133 100644
--- a/src/query/storages/system/src/configs_table.rs
+++ b/src/query/storages/system/src/configs_table.rs
@@ -134,11 +134,6 @@ impl SyncSystemTable for ConfigsTable {
             storage_config_value,
         );
 
-        let names: Vec<Vec<u8>> = names.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let values: Vec<Vec<u8>> = values.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let groups: Vec<Vec<u8>> = groups.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let descs: Vec<Vec<u8>> = descs.iter().map(|x| x.as_bytes().to_vec()).collect();
-
         Ok(DataBlock::new_from_columns(vec![
             StringType::from_data(groups),
             StringType::from_data(names),
diff --git a/src/query/storages/system/src/contributors_table.rs b/src/query/storages/system/src/contributors_table.rs
index 5d1216e16689..b70b4989ead7 100644
--- a/src/query/storages/system/src/contributors_table.rs
+++ b/src/query/storages/system/src/contributors_table.rs
@@ -42,9 +42,9 @@ impl SyncSystemTable for ContributorsTable {
     }
 
     fn get_full_data(&self, _: Arc<dyn TableContext>) -> Result<DataBlock> {
-        let contributors: Vec<Vec<u8>> = env!("DATABEND_COMMIT_AUTHORS")
+        let contributors: Vec<String> = env!("DATABEND_COMMIT_AUTHORS")
             .split_terminator(',')
-            .map(|x| x.trim().as_bytes().to_vec())
+            .map(|x| x.trim().to_string())
             .collect();
 
         Ok(DataBlock::new_from_columns(vec![StringType::from_data(
diff --git a/src/query/storages/system/src/credits_table.rs b/src/query/storages/system/src/credits_table.rs
index 0b14b38847bb..95daf3c58fc8 100644
--- a/src/query/storages/system/src/credits_table.rs
+++ b/src/query/storages/system/src/credits_table.rs
@@ -42,17 +42,17 @@ impl SyncSystemTable for CreditsTable {
     }
 
     fn get_full_data(&self, _: Arc<dyn TableContext>) -> Result<DataBlock> {
-        let names: Vec<Vec<u8>> = env!("DATABEND_CREDITS_NAMES")
+        let names: Vec<String> = env!("DATABEND_CREDITS_NAMES")
             .split_terminator(',')
-            .map(|x| x.trim().as_bytes().to_vec())
+            .map(|x| x.trim().to_string())
             .collect();
-        let versions: Vec<Vec<u8>> = env!("DATABEND_CREDITS_VERSIONS")
+        let versions: Vec<String> = env!("DATABEND_CREDITS_VERSIONS")
             .split_terminator(',')
-            .map(|x| x.trim().as_bytes().to_vec())
+            .map(|x| x.trim().to_string())
             .collect();
-        let licenses: Vec<Vec<u8>> = env!("DATABEND_CREDITS_LICENSES")
+        let licenses: Vec<String> = env!("DATABEND_CREDITS_LICENSES")
             .split_terminator(',')
-            .map(|x| x.trim().as_bytes().to_vec())
+            .map(|x| x.trim().to_string())
             .collect();
 
         Ok(DataBlock::new_from_columns(vec![
diff --git a/src/query/storages/system/src/databases_table.rs b/src/query/storages/system/src/databases_table.rs
index 2cfcd494a794..ab39d810c7b4 100644
--- a/src/query/storages/system/src/databases_table.rs
+++ b/src/query/storages/system/src/databases_table.rs
@@ -68,7 +68,7 @@ impl AsyncSystemTable for DatabasesTable {
         let mut catalog_names = vec![];
         let mut db_names = vec![];
         let mut db_id = vec![];
-        let mut owners: Vec<Option<Vec<u8>>> = vec![];
+        let mut owners: Vec<Option<String>> = vec![];
 
         let visibility_checker = ctx.get_visibility_checker().await?;
 
@@ -86,8 +86,8 @@ impl AsyncSystemTable for DatabasesTable {
                 .collect::<Vec<_>>();
 
             for db in final_dbs {
-                catalog_names.push(ctl_name.clone().into_bytes());
-                let db_name = db.name().to_string().into_bytes();
+                catalog_names.push(ctl_name.clone());
+                let db_name = db.name().to_string();
                 db_names.push(db_name);
                 let id = db.get_db_info().ident.db_id;
                 db_id.push(id);
@@ -99,7 +99,7 @@ impl AsyncSystemTable for DatabasesTable {
                         })
                         .await
                         .ok()
-                        .and_then(|ownership| ownership.map(|o| o.role.as_bytes().to_vec())),
+                        .and_then(|ownership| ownership.map(|o| o.role.clone())),
                 );
             }
         }
diff --git a/src/query/storages/system/src/engines_table.rs b/src/query/storages/system/src/engines_table.rs
index 91aea753bf8a..520430aeeeb0 100644
--- a/src/query/storages/system/src/engines_table.rs
+++ b/src/query/storages/system/src/engines_table.rs
@@ -55,8 +55,8 @@ impl AsyncSystemTable for EnginesTable {
         let mut engine_name = Vec::with_capacity(table_engine_descriptors.len());
         let mut engine_comment = Vec::with_capacity(table_engine_descriptors.len());
         for descriptor in &table_engine_descriptors {
-            engine_name.push(descriptor.engine_name.as_bytes().to_vec());
-            engine_comment.push(descriptor.comment.as_bytes().to_vec());
+            engine_name.push(descriptor.engine_name.clone());
+            engine_comment.push(descriptor.comment.clone());
         }
 
         Ok(DataBlock::new_from_columns(vec![
diff --git a/src/query/storages/system/src/indexes_table.rs b/src/query/storages/system/src/indexes_table.rs
index e5a3604499c0..75ddd041f867 100644
--- a/src/query/storages/system/src/indexes_table.rs
+++ b/src/query/storages/system/src/indexes_table.rs
@@ -68,10 +68,10 @@ impl AsyncSystemTable for IndexesTable {
         let mut updated_on = Vec::with_capacity(indexes.len());
 
         for (_, name, index) in indexes {
-            names.push(name.as_bytes().to_vec());
-            types.push(index.index_type.to_string().as_bytes().to_vec());
-            originals.push(index.original_query.as_bytes().to_vec());
-            defs.push(index.query.as_bytes().to_vec());
+            names.push(name.clone());
+            types.push(index.index_type.to_string());
+            originals.push(index.original_query.clone());
+            defs.push(index.query.clone());
             created_on.push(index.created_on.timestamp_micros());
             updated_on.push(index.updated_on.map(|u| u.timestamp_micros()));
         }
diff --git a/src/query/storages/system/src/lib.rs b/src/query/storages/system/src/lib.rs
index f7edbdbbdf07..60f242b232a7 100644
--- a/src/query/storages/system/src/lib.rs
+++ b/src/query/storages/system/src/lib.rs
@@ -15,6 +15,7 @@
 #![allow(clippy::uninlined_format_args)]
 #![feature(type_alias_impl_trait)]
 #![feature(impl_trait_in_assoc_type)]
+#![feature(variant_count)]
 
 extern crate core;
 
@@ -45,8 +46,6 @@ mod processes_table;
 mod processor_profile_table;
 mod query_cache_table;
 mod query_log_table;
-mod query_profile_table;
-mod query_summary_table;
 mod roles_table;
 mod settings_table;
 mod stages_table;
@@ -96,8 +95,6 @@ pub use query_log_table::LogType;
 pub use query_log_table::QueryLogElement;
 pub use query_log_table::QueryLogQueue;
 pub use query_log_table::QueryLogTable;
-pub use query_profile_table::QueryProfileTable;
-pub use query_summary_table::QuerySummaryTable;
 pub use roles_table::RolesTable;
 pub use settings_table::SettingsTable;
 pub use stages_table::StagesTable;
diff --git a/src/query/storages/system/src/locks_table.rs b/src/query/storages/system/src/locks_table.rs
index cd9f14e30b2c..577a95d028b1 100644
--- a/src/query/storages/system/src/locks_table.rs
+++ b/src/query/storages/system/src/locks_table.rs
@@ -101,21 +101,21 @@ impl AsyncSystemTable for LocksTable {
             for info in lock_infos {
                 lock_table_id.push(info.key.get_table_id());
                 lock_revision.push(info.revision);
-                lock_type.push(info.meta.lock_type.to_string().as_bytes().to_vec());
+                lock_type.push(info.meta.lock_type.to_string().clone());
                 if info.meta.acquired_on.is_some() {
-                    lock_status.push("HOLDING".as_bytes().to_vec());
+                    lock_status.push("HOLDING");
                 } else {
-                    lock_status.push("WAITING".as_bytes().to_vec());
+                    lock_status.push("WAITING");
                 }
-                lock_user.push(info.meta.user.as_bytes().to_vec());
-                lock_node.push(info.meta.node.as_bytes().to_vec());
-                lock_query_id.push(info.meta.query_id.as_bytes().to_vec());
+                lock_user.push(info.meta.user.clone());
+                lock_node.push(info.meta.node.clone());
+                lock_query_id.push(info.meta.query_id.clone());
                 lock_created_on.push(info.meta.created_on.timestamp_micros());
                 lock_acquired_on.push(info.meta.acquired_on.map(|v| v.timestamp_micros()));
                 if info.meta.extra_info.is_empty() {
-                    lock_extra_info.push(vec![]);
+                    lock_extra_info.push("".to_string());
                 } else {
-                    lock_extra_info.push(format!("{:?}", info.meta.extra_info).as_bytes().to_vec());
+                    lock_extra_info.push(format!("{:?}", info.meta.extra_info));
                 }
             }
         }
diff --git a/src/query/storages/system/src/malloc_stats_totals_table.rs b/src/query/storages/system/src/malloc_stats_totals_table.rs
index a0992029e2ab..9fc9b2b8c13f 100644
--- a/src/query/storages/system/src/malloc_stats_totals_table.rs
+++ b/src/query/storages/system/src/malloc_stats_totals_table.rs
@@ -41,7 +41,7 @@ macro_rules! set_value {
     ($stat:ident, $names:expr, $values:expr) => {
         let mib = $stat::mib()?;
         let value = mib.read()?;
-        $names.put_slice($stat::name().as_bytes());
+        $names.put_str(&String::from_utf8_lossy($stat::name().as_bytes()));
         $names.commit_row();
         $values.push(value as u64);
     };
diff --git a/src/query/storages/system/src/metrics_table.rs b/src/query/storages/system/src/metrics_table.rs
index 2ade31eeea47..a9a6901804d5 100644
--- a/src/query/storages/system/src/metrics_table.rs
+++ b/src/query/storages/system/src/metrics_table.rs
@@ -59,17 +59,17 @@ impl SyncSystemTable for MetricsTable {
         };
         samples.extend(self.custom_metric_samples()?);
 
-        let mut nodes: Vec<Vec<u8>> = Vec::with_capacity(samples.len());
-        let mut metrics: Vec<Vec<u8>> = Vec::with_capacity(samples.len());
-        let mut labels: Vec<Vec<u8>> = Vec::with_capacity(samples.len());
-        let mut kinds: Vec<Vec<u8>> = Vec::with_capacity(samples.len());
-        let mut values: Vec<Vec<u8>> = Vec::with_capacity(samples.len());
+        let mut nodes: Vec<String> = Vec::with_capacity(samples.len());
+        let mut metrics: Vec<String> = Vec::with_capacity(samples.len());
+        let mut labels: Vec<String> = Vec::with_capacity(samples.len());
+        let mut kinds: Vec<String> = Vec::with_capacity(samples.len());
+        let mut values: Vec<String> = Vec::with_capacity(samples.len());
         for sample in samples.into_iter() {
-            nodes.push(local_id.clone().into_bytes());
-            metrics.push(sample.name.clone().into_bytes());
-            kinds.push(sample.value.kind().into_bytes());
-            labels.push(self.display_sample_labels(&sample.labels)?.into_bytes());
-            values.push(self.display_sample_value(&sample.value)?.into_bytes());
+            nodes.push(local_id.clone());
+            metrics.push(sample.name.clone());
+            kinds.push(sample.value.kind());
+            labels.push(self.display_sample_labels(&sample.labels)?);
+            values.push(self.display_sample_value(&sample.value)?);
         }
 
         Ok(DataBlock::new_from_columns(vec![
diff --git a/src/query/storages/system/src/password_policies_table.rs b/src/query/storages/system/src/password_policies_table.rs
index 7b92e2a2fd50..2f2455ad7cc0 100644
--- a/src/query/storages/system/src/password_policies_table.rs
+++ b/src/query/storages/system/src/password_policies_table.rs
@@ -62,8 +62,8 @@ impl AsyncSystemTable for PasswordPoliciesTable {
         let mut created_on_columns = Vec::with_capacity(password_policies.len());
         let mut updated_on_columns = Vec::with_capacity(password_policies.len());
         for password_policy in password_policies {
-            names.push(password_policy.name.as_bytes().to_vec());
-            comments.push(password_policy.comment.as_bytes().to_vec());
+            names.push(password_policy.name.clone());
+            comments.push(password_policy.comment.clone());
 
             let values = vec![
                 format!("MIN_LENGTH={}", password_policy.min_length),
@@ -85,7 +85,7 @@ impl AsyncSystemTable for PasswordPoliciesTable {
                 format!("HISTORY={}", password_policy.history),
             ];
             let option = values.join(", ");
-            options.push(option.as_bytes().to_vec());
+            options.push(option);
 
             created_on_columns.push(password_policy.create_on.timestamp_micros());
             updated_on_columns.push(password_policy.update_on.map(|u| u.timestamp_micros()));
diff --git a/src/query/storages/system/src/processes_table.rs b/src/query/storages/system/src/processes_table.rs
index b608ed91cf59..25e5bcb97bf4 100644
--- a/src/query/storages/system/src/processes_table.rs
+++ b/src/query/storages/system/src/processes_table.rs
@@ -55,7 +55,7 @@ impl SyncSystemTable for ProcessesTable {
 
         let local_node = ctx.get_cluster().local_id.clone();
 
-        let mut nodes: Vec<Vec<u8>> = Vec::with_capacity(processes_info.len());
+        let mut nodes = Vec::with_capacity(processes_info.len());
         let mut processes_id = Vec::with_capacity(processes_info.len());
         let mut processes_type = Vec::with_capacity(processes_info.len());
         let mut processes_host = Vec::with_capacity(processes_info.len());
@@ -81,21 +81,17 @@ impl SyncSystemTable for ProcessesTable {
                 .unwrap_or(Duration::from_secs(0))
                 .as_secs();
 
-            nodes.push(local_node.clone().into_bytes());
-            processes_id.push(process_info.id.clone().into_bytes());
-            processes_type.push(process_info.typ.clone().into_bytes());
-            processes_state.push(process_info.state.to_string().into_bytes());
-            processes_database.push(process_info.database.clone().into_bytes());
+            nodes.push(local_node.clone());
+            processes_id.push(process_info.id.clone());
+            processes_type.push(process_info.typ.clone());
+            processes_state.push(process_info.state.to_string());
+            processes_database.push(process_info.database.clone());
             processes_host.push(ProcessesTable::process_host(&process_info.client_address));
-            processes_user.push(
-                ProcessesTable::process_option_value(process_info.user.clone())
-                    .name
-                    .into_bytes(),
-            );
-            processes_extra_info.push(
-                ProcessesTable::process_option_value(process_info.session_extra_info.clone())
-                    .into_bytes(),
-            );
+            processes_user
+                .push(ProcessesTable::process_option_value(process_info.user.clone()).name);
+            processes_extra_info.push(ProcessesTable::process_option_value(
+                process_info.session_extra_info.clone(),
+            ));
             processes_memory_usage.push(process_info.memory_usage);
             processes_scan_progress_read_rows.push(scan_progress.rows as u64);
             processes_scan_progress_read_bytes.push(scan_progress.bytes as u64);
@@ -111,13 +107,7 @@ impl SyncSystemTable for ProcessesTable {
             }
 
             // Status info.
-            processes_status.push(
-                process_info
-                    .status_info
-                    .clone()
-                    .unwrap_or("".to_owned())
-                    .into_bytes(),
-            );
+            processes_status.push(process_info.status_info.clone().unwrap_or("".to_owned()));
         }
 
         Ok(DataBlock::new_from_columns(vec![
@@ -196,8 +186,8 @@ impl ProcessesTable {
         SyncOneBlockSystemTable::create(ProcessesTable { table_info })
     }
 
-    fn process_host(client_address: &Option<SocketAddr>) -> Option<Vec<u8>> {
-        client_address.as_ref().map(|s| s.to_string().into_bytes())
+    fn process_host(client_address: &Option<SocketAddr>) -> Option<String> {
+        client_address.as_ref().map(|s| s.to_string())
     }
 
     fn process_option_value<T>(opt: Option<T>) -> T
diff --git a/src/query/storages/system/src/processor_profile_table.rs b/src/query/storages/system/src/processor_profile_table.rs
index b551bcae7a88..7c2a2d24cfdf 100644
--- a/src/query/storages/system/src/processor_profile_table.rs
+++ b/src/query/storages/system/src/processor_profile_table.rs
@@ -12,6 +12,7 @@
 // See the License for the specific language governing permissions and
 // limitations under the License.
 
+use std::collections::HashMap;
 use std::sync::atomic::Ordering;
 use std::sync::Arc;
 
@@ -22,6 +23,7 @@ use databend_common_expression::types::NumberDataType;
 use databend_common_expression::types::StringType;
 use databend_common_expression::types::UInt32Type;
 use databend_common_expression::types::UInt64Type;
+use databend_common_expression::types::VariantType;
 use databend_common_expression::DataBlock;
 use databend_common_expression::FromData;
 use databend_common_expression::TableDataType;
@@ -30,6 +32,7 @@ use databend_common_expression::TableSchemaRefExt;
 use databend_common_meta_app::schema::TableIdent;
 use databend_common_meta_app::schema::TableInfo;
 use databend_common_meta_app::schema::TableMeta;
+use databend_common_pipeline_core::processors::ProfileStatisticsName;
 
 use crate::SyncOneBlockSystemTable;
 use crate::SyncSystemTable;
@@ -53,32 +56,34 @@ impl SyncSystemTable for ProcessorProfileTable {
         let local_id = ctx.get_cluster().local_id.clone();
         let total_size = queries_profiles.values().map(Vec::len).sum();
 
-        let mut node: Vec<Vec<u8>> = Vec::with_capacity(total_size);
-        let mut queries_id: Vec<Vec<u8>> = Vec::with_capacity(total_size);
+        let mut node: Vec<String> = Vec::with_capacity(total_size);
+        let mut queries_id: Vec<String> = Vec::with_capacity(total_size);
         let mut pid: Vec<u64> = Vec::with_capacity(total_size);
-        let mut p_name: Vec<Vec<u8>> = Vec::with_capacity(total_size);
+        let mut p_name: Vec<String> = Vec::with_capacity(total_size);
         let mut plan_id: Vec<Option<u32>> = Vec::with_capacity(total_size);
         let mut parent_id: Vec<Option<u32>> = Vec::with_capacity(total_size);
-        let mut plan_name: Vec<Option<Vec<u8>>> = Vec::with_capacity(total_size);
-        let mut cpu_time: Vec<u64> = Vec::with_capacity(total_size);
-        let mut wait_time: Vec<u64> = Vec::with_capacity(total_size);
-        let mut exchange_rows: Vec<u64> = Vec::with_capacity(total_size);
-        let mut exchange_bytes: Vec<u64> = Vec::with_capacity(total_size);
+        let mut plan_name: Vec<Option<String>> = Vec::with_capacity(total_size);
+        let mut statistics = Vec::with_capacity(total_size);
 
         for (query_id, query_profiles) in queries_profiles {
             for query_profile in query_profiles {
-                node.push(local_id.clone().into_bytes());
-                queries_id.push(query_id.clone().into_bytes());
+                node.push(local_id.clone());
+                queries_id.push(query_id.clone());
                 pid.push(query_profile.pid as u64);
-                p_name.push(query_profile.p_name.clone().into_bytes());
+                p_name.push(query_profile.p_name.clone());
                 plan_id.push(query_profile.plan_id);
                 parent_id.push(query_profile.plan_parent_id);
-                plan_name.push(query_profile.plan_name.clone().map(String::into_bytes));
+                plan_name.push(query_profile.plan_name.clone());
 
-                cpu_time.push(query_profile.cpu_time.load(Ordering::Relaxed));
-                wait_time.push(query_profile.wait_time.load(Ordering::Relaxed));
-                exchange_rows.push(query_profile.exchange_rows.load(Ordering::Relaxed) as u64);
-                exchange_bytes.push(query_profile.exchange_bytes.load(Ordering::Relaxed) as u64);
+                let mut statistics_map = HashMap::with_capacity(query_profile.statistics.len());
+                for (idx, item_value) in query_profile.statistics.iter().enumerate() {
+                    statistics_map.insert(
+                        ProfileStatisticsName::from(idx).to_string(),
+                        item_value.load(Ordering::SeqCst),
+                    );
+                }
+
+                statistics.push(serde_json::to_vec(&statistics_map).unwrap());
             }
         }
 
@@ -90,10 +95,7 @@ impl SyncSystemTable for ProcessorProfileTable {
             UInt32Type::from_opt_data(plan_id),
             UInt32Type::from_opt_data(parent_id),
             StringType::from_opt_data(plan_name),
-            UInt64Type::from_data(cpu_time),
-            UInt64Type::from_data(wait_time),
-            UInt64Type::from_data(exchange_rows),
-            UInt64Type::from_data(exchange_bytes),
+            VariantType::from_data(statistics),
         ]))
     }
 }
@@ -117,16 +119,7 @@ impl ProcessorProfileTable {
                 "plan_name",
                 TableDataType::Nullable(Box::new(TableDataType::String)),
             ),
-            TableField::new("cpu_time", TableDataType::Number(NumberDataType::UInt64)),
-            TableField::new("wait_time", TableDataType::Number(NumberDataType::UInt64)),
-            TableField::new(
-                "exchange_rows",
-                TableDataType::Number(NumberDataType::UInt64),
-            ),
-            TableField::new(
-                "exchange_bytes",
-                TableDataType::Number(NumberDataType::UInt64),
-            ),
+            TableField::new("statistics", TableDataType::Variant),
         ]);
 
         let table_info = TableInfo {
diff --git a/src/query/storages/system/src/query_log_table.rs b/src/query/storages/system/src/query_log_table.rs
index f4d18a50ee6d..b2ca6604b136 100644
--- a/src/query/storages/system/src/query_log_table.rs
+++ b/src/query/storages/system/src/query_log_table.rs
@@ -295,45 +295,45 @@ impl SystemLogElement for QueryLogElement {
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.handler_type.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.handler_type.clone()).as_ref());
         // User.
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.tenant_id.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.tenant_id.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.cluster_id.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.cluster_id.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.node_id.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.node_id.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.sql_user.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.sql_user.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.sql_user_quota.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.sql_user_quota.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.sql_user_privileges.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.sql_user_privileges.clone()).as_ref());
         // Query.
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.query_id.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.query_id.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.query_kind.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.query_kind.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.query_text.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.query_text.clone()).as_ref());
         columns
             .next()
             .unwrap()
@@ -354,23 +354,23 @@ impl SystemLogElement for QueryLogElement {
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.current_database.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.current_database.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.databases.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.databases.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.tables.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.tables.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.columns.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.columns.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.projections.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.projections.clone()).as_ref());
         // Stats.
         columns
             .next()
@@ -472,15 +472,15 @@ impl SystemLogElement for QueryLogElement {
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.client_info.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.client_info.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.client_address.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.client_address.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.user_agent.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.user_agent.clone()).as_ref());
         // Exception.
         columns
             .next()
@@ -489,26 +489,26 @@ impl SystemLogElement for QueryLogElement {
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.exception_text.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.exception_text.clone()).as_ref());
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.stack_trace.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.stack_trace.clone()).as_ref());
         // Server.
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.server_version.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.server_version.clone()).as_ref());
         // Session settings
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.session_settings.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.session_settings.clone()).as_ref());
         // Extra.
         columns
             .next()
             .unwrap()
-            .push(Scalar::String(self.extra.as_bytes().to_vec()).as_ref());
+            .push(Scalar::String(self.extra.clone()).as_ref());
         columns
             .next()
             .unwrap()
diff --git a/src/query/storages/system/src/query_profile_table.rs b/src/query/storages/system/src/query_profile_table.rs
deleted file mode 100644
index 59904014c339..000000000000
--- a/src/query/storages/system/src/query_profile_table.rs
+++ /dev/null
@@ -1,117 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::sync::Arc;
-
-use databend_common_catalog::table::Table;
-use databend_common_catalog::table_context::TableContext;
-use databend_common_expression::types::NumberDataType;
-use databend_common_expression::types::StringType;
-use databend_common_expression::types::UInt32Type;
-use databend_common_expression::types::VariantType;
-use databend_common_expression::DataBlock;
-use databend_common_expression::FromData;
-use databend_common_expression::TableDataType;
-use databend_common_expression::TableField;
-use databend_common_expression::TableSchemaRefExt;
-use databend_common_meta_app::schema::TableIdent;
-use databend_common_meta_app::schema::TableInfo;
-use databend_common_meta_app::schema::TableMeta;
-use databend_common_profile::OperatorExecutionInfo;
-use databend_common_profile::QueryProfileManager;
-
-use crate::SyncOneBlockSystemTable;
-use crate::SyncSystemTable;
-
-fn encode_operator_execution_info(info: &OperatorExecutionInfo) -> jsonb::Value {
-    // Process time represent with number of milliseconds.
-    let process_time = info.process_time.as_nanos() as f64 / 1e6;
-    (&serde_json::json!({
-        "process_time": process_time,
-        "input_rows": info.input_rows,
-        "input_bytes": info.input_bytes,
-        "output_rows": info.output_rows,
-        "output_bytes": info.output_bytes,
-    }))
-        .into()
-}
-
-pub struct QueryProfileTable {
-    table_info: TableInfo,
-}
-
-impl QueryProfileTable {
-    pub fn create(table_id: u64) -> Arc<dyn Table> {
-        let schema = TableSchemaRefExt::create(vec![
-            TableField::new("query_id", TableDataType::String),
-            TableField::new("operator_id", TableDataType::Number(NumberDataType::UInt32)),
-            TableField::new("execution_info", TableDataType::Variant),
-        ]);
-
-        let table_info = TableInfo {
-            desc: "'system'.'query_profile'".to_string(),
-            ident: TableIdent::new(table_id, 0),
-            name: "query_profile".to_string(),
-            meta: TableMeta {
-                schema,
-                engine: "QueryProfileTable".to_string(),
-                ..Default::default()
-            },
-            ..Default::default()
-        };
-
-        SyncOneBlockSystemTable::create(Self { table_info })
-    }
-}
-
-impl SyncSystemTable for QueryProfileTable {
-    const NAME: &'static str = "system.query_profile";
-
-    fn get_table_info(&self) -> &TableInfo {
-        &self.table_info
-    }
-
-    fn get_full_data(
-        &self,
-        _ctx: Arc<dyn TableContext>,
-    ) -> databend_common_exception::Result<DataBlock> {
-        let profile_mgr = QueryProfileManager::instance();
-        let query_profs = profile_mgr.list_all();
-
-        let mut query_ids: Vec<Vec<u8>> = Vec::with_capacity(query_profs.len());
-        let mut operator_ids: Vec<u32> = Vec::with_capacity(query_profs.len());
-        let mut execution_infos: Vec<Vec<u8>> = Vec::with_capacity(query_profs.len());
-
-        for prof in query_profs.iter() {
-            for plan_prof in prof.operator_profiles.iter() {
-                query_ids.push(prof.query_id.clone().into_bytes());
-                operator_ids.push(plan_prof.id);
-
-                let execution_info = encode_operator_execution_info(&plan_prof.execution_info);
-                execution_infos.push(execution_info.to_vec());
-            }
-        }
-
-        let block = DataBlock::new_from_columns(vec![
-            // query_id
-            StringType::from_data(query_ids),
-            // operator_id
-            UInt32Type::from_data(operator_ids),
-            // execution_info
-            VariantType::from_data(execution_infos),
-        ]);
-
-        Ok(block)
-    }
-}
diff --git a/src/query/storages/system/src/query_summary_table.rs b/src/query/storages/system/src/query_summary_table.rs
deleted file mode 100644
index 734b44df8ec0..000000000000
--- a/src/query/storages/system/src/query_summary_table.rs
+++ /dev/null
@@ -1,180 +0,0 @@
-// Copyright 2021 Datafuse Labs
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//     http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-use std::sync::Arc;
-
-use databend_common_catalog::table::Table;
-use databend_common_catalog::table_context::TableContext;
-use databend_common_expression::types::ArgType;
-use databend_common_expression::types::ArrayType;
-use databend_common_expression::types::NumberDataType;
-use databend_common_expression::types::StringType;
-use databend_common_expression::types::UInt32Type;
-use databend_common_expression::types::ValueType;
-use databend_common_expression::types::VariantType;
-use databend_common_expression::DataBlock;
-use databend_common_expression::FromData;
-use databend_common_expression::TableDataType;
-use databend_common_expression::TableField;
-use databend_common_expression::TableSchemaRefExt;
-use databend_common_meta_app::schema::TableIdent;
-use databend_common_meta_app::schema::TableInfo;
-use databend_common_meta_app::schema::TableMeta;
-use databend_common_profile::OperatorAttribute;
-use databend_common_profile::QueryProfileManager;
-
-use crate::SyncOneBlockSystemTable;
-use crate::SyncSystemTable;
-
-// Encode an `OperatorAttribute` into jsonb::Value.
-fn encode_operator_attribute(attr: &OperatorAttribute) -> jsonb::Value {
-    match attr {
-        OperatorAttribute::Join(join_attr) => (&serde_json::json! ({
-            "join_type": join_attr.join_type,
-            "equi_conditions": join_attr.equi_conditions,
-            "non_equi_conditions": join_attr.non_equi_conditions,
-        }))
-            .into(),
-        OperatorAttribute::Aggregate(agg_attr) => (&serde_json::json!({
-            "group_keys": agg_attr.group_keys,
-            "functions": agg_attr.functions,
-        }))
-            .into(),
-        OperatorAttribute::AggregateExpand(expand_attr) => (&serde_json::json!({
-            "group_keys": expand_attr.group_keys,
-            "aggr_exprs": expand_attr.aggr_exprs,
-        }))
-            .into(),
-        OperatorAttribute::Filter(filter_attr) => {
-            (&serde_json::json!({ "predicate": filter_attr.predicate })).into()
-        }
-        OperatorAttribute::EvalScalar(scalar_attr) => {
-            (&serde_json::json!({ "scalars": scalar_attr.scalars })).into()
-        }
-        OperatorAttribute::ProjectSet(project_attr) => {
-            (&serde_json::json!({ "functions": project_attr.functions })).into()
-        }
-        OperatorAttribute::Limit(limit_attr) => (&serde_json::json!({
-            "limit": limit_attr.limit,
-            "offset": limit_attr.offset,
-        }))
-            .into(),
-        OperatorAttribute::TableScan(scan_attr) => {
-            (&serde_json::json!({ "qualified_name": scan_attr.qualified_name })).into()
-        }
-        OperatorAttribute::CteScan(cte_scan_attr) => {
-            (&serde_json::json!({ "cte_idx": cte_scan_attr.cte_idx })).into()
-        }
-        OperatorAttribute::Sort(sort_attr) => {
-            (&serde_json::json!({ "sort_keys": sort_attr.sort_keys })).into()
-        }
-        OperatorAttribute::Window(window_attr) => {
-            (&serde_json::json!({ "functions": window_attr.functions })).into()
-        }
-        OperatorAttribute::Exchange(exchange_attr) => {
-            (&serde_json::json!({ "exchange_mode": exchange_attr.exchange_mode })).into()
-        }
-        OperatorAttribute::Udf(udf_attr) => {
-            (&serde_json::json!({ "scalars": udf_attr.scalars })).into()
-        }
-        OperatorAttribute::Empty => jsonb::Value::Null,
-    }
-}
-
-pub struct QuerySummaryTable {
-    table_info: TableInfo,
-}
-
-impl QuerySummaryTable {
-    pub fn create(table_id: u64) -> Arc<dyn Table> {
-        let schema = TableSchemaRefExt::create(vec![
-            TableField::new("query_id", TableDataType::String),
-            TableField::new("operator_id", TableDataType::Number(NumberDataType::UInt32)),
-            TableField::new("operator_type", TableDataType::String),
-            TableField::new(
-                "operator_children",
-                TableDataType::Array(Box::new(TableDataType::Number(NumberDataType::UInt32))),
-            ),
-            TableField::new("operator_attribute", TableDataType::Variant),
-        ]);
-
-        let table_info = TableInfo {
-            desc: "'system'.'query_summary'".to_string(),
-            ident: TableIdent::new(table_id, 0),
-            name: "query_summary".to_string(),
-            meta: TableMeta {
-                schema,
-                engine: "QuerySummaryTable".to_string(),
-                ..Default::default()
-            },
-            ..Default::default()
-        };
-
-        SyncOneBlockSystemTable::create(Self { table_info })
-    }
-}
-
-impl SyncSystemTable for QuerySummaryTable {
-    const NAME: &'static str = "system.query_summary";
-
-    fn get_table_info(&self) -> &TableInfo {
-        &self.table_info
-    }
-
-    fn get_full_data(
-        &self,
-        _ctx: Arc<dyn TableContext>,
-    ) -> databend_common_exception::Result<DataBlock> {
-        let profile_mgr = QueryProfileManager::instance();
-        let query_profs = profile_mgr.list_all();
-
-        let mut query_ids: Vec<Vec<u8>> = Vec::with_capacity(query_profs.len());
-        let mut operator_ids: Vec<u32> = Vec::with_capacity(query_profs.len());
-        let mut operator_types: Vec<Vec<u8>> = Vec::with_capacity(query_profs.len());
-        let mut operator_childrens: Vec<Vec<u32>> = Vec::with_capacity(query_profs.len());
-        let mut operator_attributes: Vec<Vec<u8>> = Vec::with_capacity(query_profs.len());
-
-        for prof in query_profs.iter() {
-            for plan_prof in prof.operator_profiles.iter() {
-                query_ids.push(prof.query_id.clone().into_bytes());
-                operator_ids.push(plan_prof.id);
-                operator_types.push(plan_prof.operator_type.to_string().into_bytes());
-                operator_childrens.push(plan_prof.children.clone());
-
-                let attribute_value = encode_operator_attribute(&plan_prof.attribute);
-                operator_attributes.push(attribute_value.to_vec());
-            }
-        }
-
-        let block = DataBlock::new_from_columns(vec![
-            // query_id
-            StringType::from_data(query_ids),
-            // operator_id
-            UInt32Type::from_data(operator_ids),
-            // operator_type
-            StringType::from_data(operator_types),
-            // operator_children
-            ArrayType::upcast_column(ArrayType::<UInt32Type>::column_from_iter(
-                operator_childrens
-                    .into_iter()
-                    .map(|children| UInt32Type::column_from_iter(children.into_iter(), &[])),
-                &[],
-            )),
-            // operator_attribute
-            VariantType::from_data(operator_attributes),
-        ]);
-
-        Ok(block)
-    }
-}
diff --git a/src/query/storages/system/src/settings_table.rs b/src/query/storages/system/src/settings_table.rs
index ed5c3d145856..44672258d282 100644
--- a/src/query/storages/system/src/settings_table.rs
+++ b/src/query/storages/system/src/settings_table.rs
@@ -83,14 +83,6 @@ impl SyncSystemTable for SettingsTable {
             types.push(typename.to_string());
         }
 
-        let names: Vec<Vec<u8>> = names.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let values: Vec<Vec<u8>> = values.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let defaults: Vec<Vec<u8>> = defaults.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let ranges: Vec<Vec<u8>> = ranges.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let levels: Vec<Vec<u8>> = levels.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let descs: Vec<Vec<u8>> = descs.iter().map(|x| x.as_bytes().to_vec()).collect();
-        let types: Vec<Vec<u8>> = types.iter().map(|x| x.as_bytes().to_vec()).collect();
-
         Ok(DataBlock::new_from_columns(vec![
             StringType::from_data(names),
             StringType::from_data(values),
diff --git a/src/query/storages/system/src/stages_table.rs b/src/query/storages/system/src/stages_table.rs
index 77d5eb893a53..51264f0b555c 100644
--- a/src/query/storages/system/src/stages_table.rs
+++ b/src/query/storages/system/src/stages_table.rs
@@ -74,30 +74,30 @@ impl AsyncSystemTable for StagesTable {
         };
 
         let user_api = UserApiProvider::instance();
-        let mut owners: Vec<Option<Vec<u8>>> = vec![];
-        let mut name: Vec<Vec<u8>> = Vec::with_capacity(stages.len());
-        let mut stage_type: Vec<Vec<u8>> = Vec::with_capacity(stages.len());
-        let mut stage_params: Vec<Vec<u8>> = Vec::with_capacity(stages.len());
-        let mut copy_options: Vec<Vec<u8>> = Vec::with_capacity(stages.len());
-        let mut file_format_options: Vec<Vec<u8>> = Vec::with_capacity(stages.len());
-        let mut comment: Vec<Vec<u8>> = Vec::with_capacity(stages.len());
+        let mut owners: Vec<Option<String>> = vec![];
+        let mut name: Vec<String> = Vec::with_capacity(stages.len());
+        let mut stage_type: Vec<String> = Vec::with_capacity(stages.len());
+        let mut stage_params: Vec<String> = Vec::with_capacity(stages.len());
+        let mut copy_options: Vec<String> = Vec::with_capacity(stages.len());
+        let mut file_format_options: Vec<String> = Vec::with_capacity(stages.len());
+        let mut comment: Vec<String> = Vec::with_capacity(stages.len());
         let mut number_of_files: Vec<Option<u64>> = Vec::with_capacity(stages.len());
-        let mut creator: Vec<Option<Vec<u8>>> = Vec::with_capacity(stages.len());
+        let mut creator: Vec<Option<String>> = Vec::with_capacity(stages.len());
         let mut created_on = Vec::with_capacity(stages.len());
         for stage in stages.into_iter() {
             let stage_name = stage.stage_name;
-            name.push(stage_name.clone().into_bytes());
+            name.push(stage_name.clone());
             owners.push(
                 user_api
                     .get_ownership(&tenant, &OwnershipObject::Stage { name: stage_name })
                     .await
                     .ok()
-                    .and_then(|ownership| ownership.map(|o| o.role.as_bytes().to_vec())),
+                    .and_then(|ownership| ownership.map(|o| o.role.clone())),
             );
-            stage_type.push(stage.stage_type.clone().to_string().into_bytes());
-            stage_params.push(format!("{:?}", stage.stage_params).into_bytes());
-            copy_options.push(format!("{:?}", stage.copy_options).into_bytes());
-            file_format_options.push(format!("{:?}", stage.file_format_params).into_bytes());
+            stage_type.push(stage.stage_type.clone().to_string());
+            stage_params.push(format!("{:?}", stage.stage_params));
+            copy_options.push(format!("{:?}", stage.copy_options));
+            file_format_options.push(format!("{:?}", stage.file_format_params));
             // TODO(xuanwo): we will remove this line.
             match stage.stage_type {
                 StageType::LegacyInternal | StageType::Internal | StageType::User => {
@@ -107,9 +107,9 @@ impl AsyncSystemTable for StagesTable {
                     number_of_files.push(None);
                 }
             };
-            creator.push(stage.creator.map(|c| c.to_string().into_bytes().to_vec()));
+            creator.push(stage.creator.map(|c| c.to_string()));
             created_on.push(stage.created_on.timestamp_micros());
-            comment.push(stage.comment.clone().into_bytes());
+            comment.push(stage.comment.clone());
         }
 
         Ok(DataBlock::new_from_columns(vec![
diff --git a/src/query/storages/system/src/streams_table.rs b/src/query/storages/system/src/streams_table.rs
index 3cc7312f17d1..753176d58fee 100644
--- a/src/query/storages/system/src/streams_table.rs
+++ b/src/query/storages/system/src/streams_table.rs
@@ -83,7 +83,7 @@ impl AsyncSystemTable for StreamsTable {
         let mut table_name = vec![];
         let mut invalid_reason = vec![];
         let mut mode = vec![];
-        let mut names = vec![];
+        let mut names: Vec<String> = vec![];
         let mut stream_id = vec![];
         let mut created_on = vec![];
         let mut updated_on = vec![];
@@ -95,16 +95,14 @@ impl AsyncSystemTable for StreamsTable {
         for (ctl_name, ctl) in ctls.into_iter() {
             let mut dbs = Vec::new();
             if let Some(push_downs) = &push_downs {
-                let mut db_name = Vec::new();
+                let mut db_name: Vec<String> = Vec::new();
                 if let Some(filter) = push_downs.filters.as_ref().map(|f| &f.filter) {
                     let expr = filter.as_expr(&BUILTIN_FUNCTIONS);
                     find_eq_filter(&expr, &mut |col_name, scalar| {
                         if col_name == "database" {
-                            if let Scalar::String(s) = scalar {
-                                if let Ok(database) = String::from_utf8(s.clone()) {
-                                    if !db_name.contains(&database) {
-                                        db_name.push(database);
-                                    }
+                            if let Scalar::String(database) = scalar {
+                                if !db_name.contains(database) {
+                                    db_name.push(database.clone());
                                 }
                             }
                         }
@@ -163,11 +161,11 @@ impl AsyncSystemTable for StreamsTable {
                         t_id,
                     ) && table.engine() == "STREAM"
                     {
-                        catalogs.push(ctl_name.as_bytes().to_vec());
-                        databases.push(name.as_bytes().to_vec());
+                        catalogs.push(ctl_name);
+                        databases.push(name);
 
                         let stream_info = table.get_table_info();
-                        names.push(table.name().as_bytes().to_vec());
+                        names.push(table.name().to_string());
                         stream_id.push(stream_info.ident.table_id);
                         created_on.push(stream_info.meta.created_on.timestamp_micros());
                         updated_on.push(stream_info.meta.updated_on.timestamp_micros());
@@ -180,27 +178,20 @@ impl AsyncSystemTable for StreamsTable {
                                 })
                                 .await
                                 .ok()
-                                .and_then(|ownership| {
-                                    ownership.map(|o| o.role.as_bytes().to_vec())
-                                }),
+                                .and_then(|ownership| ownership.map(|o| o.role.clone())),
                         );
-                        comment.push(stream_info.meta.comment.as_bytes().to_vec());
+                        comment.push(stream_info.meta.comment.clone());
 
                         let stream_table = StreamTable::try_from_table(table.as_ref())?;
-                        table_name.push(
-                            format!(
-                                "{}.{}",
-                                stream_table.source_table_database(),
-                                stream_table.source_table_name()
-                            )
-                            .as_bytes()
-                            .to_vec(),
-                        );
-                        mode.push(stream_table.mode().to_string().as_bytes().to_vec());
+                        table_name.push(format!(
+                            "{}.{}",
+                            stream_table.source_table_database(),
+                            stream_table.source_table_name()
+                        ));
+                        mode.push(stream_table.mode().to_string());
                         table_version.push(stream_table.offset());
                         table_id.push(stream_table.source_table_id());
-                        snapshot_location
-                            .push(stream_table.snapshot_loc().map(|v| v.as_bytes().to_vec()));
+                        snapshot_location.push(stream_table.snapshot_loc());
 
                         let mut reason = "".to_string();
                         match stream_table.source_table(ctx.clone()).await {
@@ -220,7 +211,7 @@ impl AsyncSystemTable for StreamsTable {
                                 reason = e.display_text();
                             }
                         }
-                        invalid_reason.push(reason.as_bytes().to_vec());
+                        invalid_reason.push(reason);
                     }
                 }
             }
diff --git a/src/query/storages/system/src/tables_table.rs b/src/query/storages/system/src/tables_table.rs
index e2d49572d92f..f83b39c792cc 100644
--- a/src/query/storages/system/src/tables_table.rs
+++ b/src/query/storages/system/src/tables_table.rs
@@ -183,22 +183,20 @@ where TablesTable<T>: HistoryAware
         let mut databases = vec![];
 
         let mut database_tables = vec![];
-        let mut owner: Vec<Option<Vec<u8>>> = Vec::new();
+        let mut owner: Vec<Option<String>> = Vec::new();
         let user_api = UserApiProvider::instance();
 
         for (ctl_name, ctl) in ctls.into_iter() {
             let mut dbs = Vec::new();
             if let Some(push_downs) = &push_downs {
-                let mut db_name = Vec::new();
+                let mut db_name: Vec<String> = Vec::new();
                 if let Some(filter) = push_downs.filters.as_ref().map(|f| &f.filter) {
                     let expr = filter.as_expr(&BUILTIN_FUNCTIONS);
                     find_eq_filter(&expr, &mut |col_name, scalar| {
                         if col_name == "database" {
-                            if let Scalar::String(s) = scalar {
-                                if let Ok(database) = String::from_utf8(s.clone()) {
-                                    if !db_name.contains(&database) {
-                                        db_name.push(database);
-                                    }
+                            if let Scalar::String(database) = scalar {
+                                if !db_name.contains(database) {
+                                    db_name.push(database.clone());
                                 }
                             }
                         }
@@ -276,8 +274,8 @@ where TablesTable<T>: HistoryAware
                         table_id,
                     ) && table.engine() != "STREAM"
                     {
-                        catalogs.push(ctl_name.as_bytes().to_vec());
-                        databases.push(name.as_bytes().to_vec());
+                        catalogs.push(ctl_name);
+                        databases.push(name);
                         database_tables.push(table);
                         owner.push(
                             user_api
@@ -288,9 +286,7 @@ where TablesTable<T>: HistoryAware
                                 })
                                 .await
                                 .ok()
-                                .and_then(|ownership| {
-                                    ownership.map(|o| o.role.as_bytes().to_vec())
-                                }),
+                                .and_then(|ownership| ownership.map(|o| o.role.to_string())),
                         );
                     }
                 }
@@ -298,7 +294,6 @@ where TablesTable<T>: HistoryAware
         }
 
         let mut number_of_blocks: Vec<Option<u64>> = Vec::new();
-
         let mut number_of_segments: Vec<Option<u64>> = Vec::new();
         let mut num_rows: Vec<Option<u64>> = Vec::new();
         let mut data_size: Vec<Option<u64>> = Vec::new();
@@ -328,19 +323,19 @@ where TablesTable<T>: HistoryAware
             index_size.push(stats.as_ref().and_then(|v| v.index_size));
         }
 
-        let names: Vec<Vec<u8>> = database_tables
+        let names: Vec<String> = database_tables
             .iter()
-            .map(|v| v.name().as_bytes().to_vec())
+            .map(|v| v.name().to_string())
             .collect();
         let table_id: Vec<u64> = database_tables
             .iter()
             .map(|v| v.get_table_info().ident.table_id)
             .collect();
-        let engines: Vec<Vec<u8>> = database_tables
+        let engines: Vec<String> = database_tables
             .iter()
-            .map(|v| v.engine().as_bytes().to_vec())
+            .map(|v| v.engine().to_string())
             .collect();
-        let engines_full: Vec<Vec<u8>> = engines.clone();
+        let engines_full: Vec<String> = engines.clone();
         let created_on: Vec<i64> = database_tables
             .iter()
             .map(|v| v.get_table_info().meta.created_on.timestamp_micros())
@@ -369,14 +364,13 @@ where TablesTable<T>: HistoryAware
                     .unwrap_or_else(|| "".to_owned())
             })
             .collect();
-        let cluster_bys: Vec<Vec<u8>> = cluster_bys.iter().map(|s| s.as_bytes().to_vec()).collect();
-        let is_transient: Vec<Vec<u8>> = database_tables
+        let is_transient: Vec<String> = database_tables
             .iter()
             .map(|v| {
                 if v.options().contains_key("TRANSIENT") {
-                    "TRANSIENT".as_bytes().to_vec()
+                    "TRANSIENT".to_string()
                 } else {
-                    vec![]
+                    "".to_string()
                 }
             })
             .collect();
diff --git a/src/query/storages/system/src/task_history_table.rs b/src/query/storages/system/src/task_history_table.rs
index b3dc2df7524b..6757208efba0 100644
--- a/src/query/storages/system/src/task_history_table.rs
+++ b/src/query/storages/system/src/task_history_table.rs
@@ -31,6 +31,7 @@ use databend_common_expression::types::Int64Type;
 use databend_common_expression::types::StringType;
 use databend_common_expression::types::TimestampType;
 use databend_common_expression::types::UInt64Type;
+use databend_common_expression::types::VariantType;
 use databend_common_expression::DataBlock;
 use databend_common_expression::FromData;
 use databend_common_meta_app::schema::TableIdent;
@@ -42,46 +43,46 @@ use crate::table::AsyncOneBlockSystemTable;
 use crate::table::AsyncSystemTable;
 
 pub fn parse_task_runs_to_datablock(task_runs: Vec<TaskRun>) -> Result<DataBlock> {
-    let mut name: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
+    let mut name: Vec<String> = Vec::with_capacity(task_runs.len());
     let mut id: Vec<u64> = Vec::with_capacity(task_runs.len());
-    let mut owner: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
-    let mut definition: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
-    let mut condition_text: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
-    let mut comment: Vec<Option<Vec<u8>>> = Vec::with_capacity(task_runs.len());
-    let mut schedule: Vec<Option<Vec<u8>>> = Vec::with_capacity(task_runs.len());
-    let mut warehouse: Vec<Option<Vec<u8>>> = Vec::with_capacity(task_runs.len());
-    let mut state: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
-    let mut exception_text: Vec<Option<Vec<u8>>> = Vec::with_capacity(task_runs.len());
+    let mut owner: Vec<String> = Vec::with_capacity(task_runs.len());
+    let mut definition: Vec<String> = Vec::with_capacity(task_runs.len());
+    let mut condition_text: Vec<String> = Vec::with_capacity(task_runs.len());
+    let mut comment: Vec<Option<String>> = Vec::with_capacity(task_runs.len());
+    let mut schedule: Vec<Option<String>> = Vec::with_capacity(task_runs.len());
+    let mut warehouse: Vec<Option<String>> = Vec::with_capacity(task_runs.len());
+    let mut state: Vec<String> = Vec::with_capacity(task_runs.len());
+    let mut exception_text: Vec<Option<String>> = Vec::with_capacity(task_runs.len());
     let mut exception_code: Vec<i64> = Vec::with_capacity(task_runs.len());
-    let mut run_id: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
-    let mut query_id: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
+    let mut run_id: Vec<String> = Vec::with_capacity(task_runs.len());
+    let mut query_id: Vec<String> = Vec::with_capacity(task_runs.len());
     let mut attempt_number: Vec<i32> = Vec::with_capacity(task_runs.len());
     let mut scheduled_time: Vec<i64> = Vec::with_capacity(task_runs.len());
     let mut completed_time: Vec<Option<i64>> = Vec::with_capacity(task_runs.len());
-    let mut root_task_id: Vec<Vec<u8>> = Vec::with_capacity(task_runs.len());
+    let mut root_task_id: Vec<String> = Vec::with_capacity(task_runs.len());
+    let mut session_params: Vec<Option<Vec<u8>>> = Vec::with_capacity(task_runs.len());
 
     for task_run in task_runs {
         let tr: databend_common_cloud_control::task_utils::TaskRun = task_run.try_into()?;
-        name.push(tr.task_name.into_bytes());
+        name.push(tr.task_name);
         id.push(tr.task_id);
-        owner.push(tr.owner.into_bytes());
-        comment.push(tr.comment.map(|s| s.into_bytes()));
-        schedule.push(tr.schedule_options.map(|s| s.into_bytes()));
-        warehouse.push(
-            tr.warehouse_options
-                .and_then(|s| s.warehouse.map(|v| v.into_bytes())),
-        );
-        state.push(tr.state.to_string().into_bytes());
+        owner.push(tr.owner);
+        comment.push(tr.comment);
+        schedule.push(tr.schedule_options);
+        warehouse.push(tr.warehouse_options.and_then(|s| s.warehouse));
+        state.push(tr.state.to_string());
         exception_code.push(tr.error_code);
-        exception_text.push(tr.error_message.map(|s| s.into_bytes()));
-        definition.push(tr.query_text.into_bytes());
-        condition_text.push(tr.condition_text.into_bytes());
-        run_id.push(tr.run_id.into_bytes());
-        query_id.push(tr.query_id.into_bytes());
+        exception_text.push(tr.error_message);
+        definition.push(tr.query_text);
+        condition_text.push(tr.condition_text);
+        run_id.push(tr.run_id);
+        query_id.push(tr.query_id);
         attempt_number.push(tr.attempt_number);
         completed_time.push(tr.completed_at.map(|t| t.timestamp_micros()));
         scheduled_time.push(tr.scheduled_at.timestamp_micros());
-        root_task_id.push(tr.root_task_id.into_bytes());
+        root_task_id.push(tr.root_task_id);
+        let serialized_params = serde_json::to_vec(&tr.session_params).unwrap();
+        session_params.push(Some(serialized_params));
     }
     Ok(DataBlock::new_from_columns(vec![
         StringType::from_data(name),
@@ -101,6 +102,7 @@ pub fn parse_task_runs_to_datablock(task_runs: Vec<TaskRun>) -> Result<DataBlock
         TimestampType::from_opt_data(completed_time),
         TimestampType::from_data(scheduled_time),
         StringType::from_data(root_task_id),
+        VariantType::from_opt_data(session_params),
     ]))
 }
 
diff --git a/src/query/storages/system/src/tasks_table.rs b/src/query/storages/system/src/tasks_table.rs
index b577b015e3a4..3fba18c13a4f 100644
--- a/src/query/storages/system/src/tasks_table.rs
+++ b/src/query/storages/system/src/tasks_table.rs
@@ -29,6 +29,7 @@ use databend_common_expression::infer_table_schema;
 use databend_common_expression::types::StringType;
 use databend_common_expression::types::TimestampType;
 use databend_common_expression::types::UInt64Type;
+use databend_common_expression::types::VariantType;
 use databend_common_expression::DataBlock;
 use databend_common_expression::FromData;
 use databend_common_meta_app::schema::TableIdent;
@@ -41,48 +42,41 @@ use crate::table::AsyncSystemTable;
 
 pub fn parse_tasks_to_datablock(tasks: Vec<Task>) -> Result<DataBlock> {
     let mut created_on: Vec<i64> = Vec::with_capacity(tasks.len());
-    let mut name: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
+    let mut name: Vec<String> = Vec::with_capacity(tasks.len());
     let mut id: Vec<u64> = Vec::with_capacity(tasks.len());
-    let mut owner: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-    let mut comment: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
-    let mut warehouse: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
-    let mut schedule: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
-    let mut status: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-    let mut definition: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-    let mut condition_text: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
-    let mut after: Vec<Vec<u8>> = Vec::with_capacity(tasks.len());
+    let mut owner: Vec<String> = Vec::with_capacity(tasks.len());
+    let mut comment: Vec<Option<String>> = Vec::with_capacity(tasks.len());
+    let mut warehouse: Vec<Option<String>> = Vec::with_capacity(tasks.len());
+    let mut schedule: Vec<Option<String>> = Vec::with_capacity(tasks.len());
+    let mut status: Vec<String> = Vec::with_capacity(tasks.len());
+    let mut definition: Vec<String> = Vec::with_capacity(tasks.len());
+    let mut condition_text: Vec<String> = Vec::with_capacity(tasks.len());
+    let mut after: Vec<String> = Vec::with_capacity(tasks.len());
     let mut suspend_after_num_failures: Vec<Option<u64>> = Vec::with_capacity(tasks.len());
     let mut last_committed_on: Vec<i64> = Vec::with_capacity(tasks.len());
     let mut next_schedule_time: Vec<Option<i64>> = Vec::with_capacity(tasks.len());
     let mut last_suspended_on: Vec<Option<i64>> = Vec::with_capacity(tasks.len());
-
+    let mut session_params: Vec<Option<Vec<u8>>> = Vec::with_capacity(tasks.len());
     for task in tasks {
         let tsk: databend_common_cloud_control::task_utils::Task = task.try_into()?;
         created_on.push(tsk.created_at.timestamp_micros());
-        name.push(tsk.task_name.into_bytes());
+        name.push(tsk.task_name);
         id.push(tsk.task_id);
-        owner.push(tsk.owner.into_bytes());
-        comment.push(tsk.comment.map(|s| s.into_bytes()));
-        warehouse.push(
-            tsk.warehouse_options
-                .and_then(|s| s.warehouse.map(|v| v.into_bytes())),
-        );
-        schedule.push(tsk.schedule_options.map(|s| s.into_bytes()));
-        status.push(tsk.status.to_string().into_bytes());
-        definition.push(tsk.query_text.into_bytes());
-        condition_text.push(tsk.condition_text.into_bytes());
+        owner.push(tsk.owner);
+        comment.push(tsk.comment);
+        warehouse.push(tsk.warehouse_options.and_then(|s| s.warehouse));
+        schedule.push(tsk.schedule_options);
+        status.push(tsk.status.to_string());
+        definition.push(tsk.query_text);
+        condition_text.push(tsk.condition_text);
         // join by comma
-        after.push(
-            tsk.after
-                .into_iter()
-                .collect::<Vec<_>>()
-                .join(",")
-                .into_bytes(),
-        );
+        after.push(tsk.after.into_iter().collect::<Vec<_>>().join(","));
         suspend_after_num_failures.push(tsk.suspend_task_after_num_failures.map(|v| v as u64));
         next_schedule_time.push(tsk.next_scheduled_at.map(|t| t.timestamp_micros()));
         last_committed_on.push(tsk.updated_at.timestamp_micros());
         last_suspended_on.push(tsk.last_suspended_at.map(|t| t.timestamp_micros()));
+        let serialized_params = serde_json::to_vec(&tsk.session_params).unwrap();
+        session_params.push(Some(serialized_params));
     }
 
     Ok(DataBlock::new_from_columns(vec![
@@ -101,6 +95,7 @@ pub fn parse_tasks_to_datablock(tasks: Vec<Task>) -> Result<DataBlock> {
         TimestampType::from_opt_data(next_schedule_time),
         TimestampType::from_data(last_committed_on),
         TimestampType::from_opt_data(last_suspended_on),
+        VariantType::from_opt_data(session_params),
     ]))
 }
 
diff --git a/src/query/storages/system/src/temp_files_table.rs b/src/query/storages/system/src/temp_files_table.rs
index acb2afcd833a..b2cdaa144905 100644
--- a/src/query/storages/system/src/temp_files_table.rs
+++ b/src/query/storages/system/src/temp_files_table.rs
@@ -64,7 +64,7 @@ impl AsyncSystemTable for TempFilesTable {
         let tenant = ctx.get_tenant();
         let operator = DataOperator::instance().operator();
 
-        let mut temp_files_name = vec![];
+        let mut temp_files_name: Vec<String> = vec![];
         let mut temp_files_content_length = vec![];
         let mut temp_files_last_modified = vec![];
 
@@ -81,7 +81,7 @@ impl AsyncSystemTable for TempFilesTable {
                 let metadata = entry.metadata();
 
                 if metadata.is_file() {
-                    temp_files_name.push(entry.name().as_bytes().to_vec());
+                    temp_files_name.push(entry.name().to_string());
 
                     temp_files_last_modified
                         .push(metadata.last_modified().map(|x| x.timestamp_micros()));
@@ -95,7 +95,7 @@ impl AsyncSystemTable for TempFilesTable {
             vec![
                 BlockEntry::new(
                     DataType::String,
-                    Value::Scalar(Scalar::String("Spill".as_bytes().to_owned())),
+                    Value::Scalar(Scalar::String("Spill".to_string())),
                 ),
                 BlockEntry::new(
                     DataType::String,
diff --git a/src/query/storages/system/src/tracing_table.rs b/src/query/storages/system/src/tracing_table.rs
index 13d06de9fd7a..c6fc69ce35b5 100644
--- a/src/query/storages/system/src/tracing_table.rs
+++ b/src/query/storages/system/src/tracing_table.rs
@@ -181,7 +181,7 @@ impl SyncSource for TracingSource {
 
                         entry_column = ColumnBuilder::with_capacity(&DataType::String, max_rows);
                     }
-                    entry_column.push(Scalar::String(line.unwrap().as_bytes().to_vec()).as_ref());
+                    entry_column.push(Scalar::String(line.unwrap()).as_ref());
                 }
 
                 if entry_column.len() > 0 {
diff --git a/src/query/storages/system/src/users_table.rs b/src/query/storages/system/src/users_table.rs
index 90578915620f..c43112cf5df6 100644
--- a/src/query/storages/system/src/users_table.rs
+++ b/src/query/storages/system/src/users_table.rs
@@ -54,35 +54,25 @@ impl AsyncSystemTable for UsersTable {
         let tenant = ctx.get_tenant();
         let users = UserApiProvider::instance().get_users(&tenant).await?;
 
-        let mut names: Vec<Vec<u8>> = users.iter().map(|x| x.name.as_bytes().to_vec()).collect();
-        let mut hostnames: Vec<Vec<u8>> = users
+        let mut names: Vec<String> = users.iter().map(|x| x.name.clone()).collect();
+        let mut hostnames: Vec<String> = users.iter().map(|x| x.hostname.clone()).collect();
+        let mut auth_types: Vec<String> = users
             .iter()
-            .map(|x| x.hostname.as_bytes().to_vec())
+            .map(|x| x.auth_info.get_type().to_str().to_string())
             .collect();
-        let mut auth_types: Vec<Vec<u8>> = users
+        let mut default_roles: Vec<String> = users
             .iter()
-            .map(|x| x.auth_info.get_type().to_str().as_bytes().to_vec())
+            .map(|x| x.option.default_role().cloned().unwrap_or_default())
             .collect();
-        let mut default_roles: Vec<Vec<u8>> = users
-            .iter()
-            .map(|x| {
-                x.option
-                    .default_role()
-                    .cloned()
-                    .unwrap_or_default()
-                    .as_bytes()
-                    .to_vec()
-            })
-            .collect();
-        let mut is_configureds: Vec<Vec<u8>> = vec!["NO".as_bytes().to_vec(); users.len()];
+        let mut is_configureds: Vec<String> = vec!["NO".to_string(); users.len()];
 
         let configured_users = UserApiProvider::instance().get_configured_users();
         for (name, auth_info) in configured_users {
-            names.push(name.as_bytes().to_vec());
-            hostnames.push("%".as_bytes().to_vec());
-            auth_types.push(auth_info.get_type().to_str().as_bytes().to_vec());
-            default_roles.push(BUILTIN_ROLE_ACCOUNT_ADMIN.as_bytes().to_vec());
-            is_configureds.push("YES".as_bytes().to_vec());
+            names.push(name.clone());
+            hostnames.push("%".to_string());
+            auth_types.push(auth_info.get_type().to_str().to_string());
+            default_roles.push(BUILTIN_ROLE_ACCOUNT_ADMIN.to_string());
+            is_configureds.push("YES".to_string());
         }
 
         // please note that do NOT display the auth_string field in the result, because there're risks of
diff --git a/src/query/storages/system/src/virtual_columns_table.rs b/src/query/storages/system/src/virtual_columns_table.rs
index 670226f8722a..8e8e061bdd15 100644
--- a/src/query/storages/system/src/virtual_columns_table.rs
+++ b/src/query/storages/system/src/virtual_columns_table.rs
@@ -65,7 +65,7 @@ impl AsyncSystemTable for VirtualColumnsTable {
             .await?;
 
         let mut database_names = Vec::with_capacity(virtual_column_metas.len());
-        let mut table_names = Vec::with_capacity(virtual_column_metas.len());
+        let mut table_names: Vec<String> = Vec::with_capacity(virtual_column_metas.len());
         let mut virtual_columns = Vec::with_capacity(virtual_column_metas.len());
         let mut created_on_columns = Vec::with_capacity(virtual_column_metas.len());
         let mut updated_on_columns = Vec::with_capacity(virtual_column_metas.len());
@@ -81,15 +81,9 @@ impl AsyncSystemTable for VirtualColumnsTable {
                 for table in tables {
                     let table_id = table.get_id();
                     if let Some(virtual_column_meta) = virtual_column_meta_map.remove(&table_id) {
-                        database_names.push(database.as_bytes().to_vec());
-                        table_names.push(table.name().as_bytes().to_vec());
-                        virtual_columns.push(
-                            virtual_column_meta
-                                .virtual_columns
-                                .join(", ")
-                                .as_bytes()
-                                .to_vec(),
-                        );
+                        database_names.push(database.clone());
+                        table_names.push(table.name().to_string());
+                        virtual_columns.push(virtual_column_meta.virtual_columns.join(", "));
                         created_on_columns.push(virtual_column_meta.created_on.timestamp_micros());
                         updated_on_columns
                             .push(virtual_column_meta.updated_on.map(|u| u.timestamp_micros()));
diff --git a/tests/cloud_control_server/simple_server.py b/tests/cloud_control_server/simple_server.py
index 2fd225eff44f..1f38ee11b283 100644
--- a/tests/cloud_control_server/simple_server.py
+++ b/tests/cloud_control_server/simple_server.py
@@ -55,7 +55,8 @@ def create_task_request_to_task(id, create_task_request):
     task.after.extend(create_task_request.after)
     task.created_at = datetime.now(timezone.utc).isoformat()
     task.updated_at = datetime.now(timezone.utc).isoformat()
-
+    # add session parameters
+    task.session_parameters.update(create_task_request.session_parameters)
     return task
 
 
@@ -91,7 +92,7 @@ def create_task_run_from_task(task):
     task_run.query_id = "qwert"
     task_run.scheduled_time = datetime.now(timezone.utc).isoformat()
     task_run.completed_time = datetime.now(timezone.utc).isoformat()
-
+    task_run.session_parameters.update(task.session_parameters)
     return task_run
 
 
@@ -210,6 +211,9 @@ def AlterTask(self, request, context):
                     request.suspend_task_after_num_failures
                 )
                 has_options = True
+            if request.set_session_parameters:
+                task.session_parameters.update(request.session_parameters)
+                has_options = True
             if has_options is False:
                 return task_pb2.AlterTaskResponse(
                     error=task_pb2.TaskError(
@@ -265,14 +269,17 @@ def GetTaskDependents(self, request, context):
             root = TASK_DB[root.after[0]]
             l.insert(0, root)
         return task_pb2.GetTaskDependentsResponse(task=l)
+
     def EnableTaskDependents(self, request, context):
         print("EnableTaskDependents", request)
         task_name = request.task_name
         if task_name not in TASK_DB:
-           return task_pb2.EnableTaskDependentsResponse()
+            return task_pb2.EnableTaskDependentsResponse()
         task = TASK_DB[task_name]
         task.status = task_pb2.Task.Started
         return task_pb2.EnableTaskDependentsResponse()
+
+
 def serve():
     server = grpc.server(futures.ThreadPoolExecutor(max_workers=10))
     task_pb2_grpc.add_TaskServiceServicer_to_server(TaskService(), server)
diff --git a/tests/cloud_control_server/task_pb2.py b/tests/cloud_control_server/task_pb2.py
index 3b3ad6c56c62..ba9d8ac91aed 100644
--- a/tests/cloud_control_server/task_pb2.py
+++ b/tests/cloud_control_server/task_pb2.py
@@ -6,75 +6,92 @@
 from google.protobuf import descriptor_pool as _descriptor_pool
 from google.protobuf import symbol_database as _symbol_database
 from google.protobuf.internal import builder as _builder
+
 # @@protoc_insertion_point(imports)
 
 _sym_db = _symbol_database.Default()
 
 
-
-
-DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\ntask.proto\x12\ttaskproto\"\xe9\x01\n\x0fScheduleOptions\x12\x15\n\x08interval\x18\x01 \x01(\x05H\x00\x88\x01\x01\x12\x11\n\x04\x63ron\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x16\n\ttime_zone\x18\x03 \x01(\tH\x02\x88\x01\x01\x12>\n\rschedule_type\x18\x04 \x01(\x0e\x32\'.taskproto.ScheduleOptions.ScheduleType\"0\n\x0cScheduleType\x12\x11\n\rinterval_type\x10\x00\x12\r\n\tcron_type\x10\x01\x42\x0b\n\t_intervalB\x07\n\x05_cronB\x0c\n\n_time_zone\"t\n\x10WarehouseOptions\x12\x16\n\twarehouse\x18\x01 \x01(\tH\x00\x88\x01\x01\x12!\n\x14using_warehouse_size\x18\x02 \x01(\tH\x01\x88\x01\x01\x42\x0c\n\n_warehouseB\x17\n\x15_using_warehouse_size\"\x93\x03\n\x11\x43reateTaskRequest\x12\x11\n\ttask_name\x18\x01 \x01(\t\x12\x11\n\ttenant_id\x18\x02 \x01(\t\x12\x12\n\nquery_text\x18\x03 \x01(\t\x12\r\n\x05owner\x18\x04 \x01(\t\x12\x14\n\x07\x63omment\x18\x05 \x01(\tH\x00\x88\x01\x01\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x36\n\x11warehouse_options\x18\x08 \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12,\n\x1fsuspend_task_after_num_failures\x18\t \x01(\x05H\x01\x88\x01\x01\x12\x14\n\x0cif_not_exist\x18\n \x01(\x08\x12\r\n\x05\x61\x66ter\x18\x0b \x03(\t\x12\x1b\n\x0ewhen_condition\x18\x0c \x01(\tH\x02\x88\x01\x01\x42\n\n\x08_commentB\"\n _suspend_task_after_num_failuresB\x11\n\x0f_when_condition\"8\n\tTaskError\x12\x0c\n\x04kind\x18\x01 \x01(\t\x12\x0f\n\x07message\x18\x02 \x01(\t\x12\x0c\n\x04\x63ode\x18\x03 \x01(\x05\"Y\n\x12\x43reateTaskResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x12\x0f\n\x07task_id\x18\x02 \x01(\x04\x42\x08\n\x06_error\"I\n\x0f\x44ropTaskRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x10\n\x08if_exist\x18\x03 \x01(\x08\"F\n\x10\x44ropTaskResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error\":\n\x12\x45xecuteTaskRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\"I\n\x13\x45xecuteTaskResponse\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error\"M\n\x13\x44\x65scribeTaskRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x10\n\x08if_exist\x18\x03 \x01(\x08\"\xd0\x04\n\x04Task\x12\x0f\n\x07task_id\x18\x01 \x01(\x04\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x12\n\nquery_text\x18\x04 \x01(\t\x12\x14\n\x07\x63omment\x18\x05 \x01(\tH\x00\x88\x01\x01\x12\r\n\x05owner\x18\x06 \x01(\t\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x36\n\x11warehouse_options\x18\x08 \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12\x1e\n\x11next_scheduled_at\x18\t \x01(\tH\x01\x88\x01\x01\x12,\n\x1fsuspend_task_after_num_failures\x18\n \x01(\x05H\x02\x88\x01\x01\x12&\n\x06status\x18\x0c \x01(\x0e\x32\x16.taskproto.Task.Status\x12\x12\n\ncreated_at\x18\x0e \x01(\t\x12\x12\n\nupdated_at\x18\x0f \x01(\t\x12\x1e\n\x11last_suspended_at\x18\x10 \x01(\tH\x03\x88\x01\x01\x12\r\n\x05\x61\x66ter\x18\x11 \x03(\t\x12\x1b\n\x0ewhen_condition\x18\x12 \x01(\tH\x04\x88\x01\x01\"$\n\x06Status\x12\r\n\tSuspended\x10\x00\x12\x0b\n\x07Started\x10\x01\x42\n\n\x08_commentB\x14\n\x12_next_scheduled_atB\"\n _suspend_task_after_num_failuresB\x14\n\x12_last_suspended_atB\x11\n\x0f_when_condition\"i\n\x14\x44\x65scribeTaskResponse\x12\x1d\n\x04task\x18\x01 \x01(\x0b\x32\x0f.taskproto.Task\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error\"p\n\x10ShowTasksRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\tname_like\x18\x02 \x01(\t\x12\x14\n\x0cresult_limit\x18\x04 \x01(\x05\x12\x0e\n\x06owners\x18\x05 \x03(\t\x12\x10\n\x08task_ids\x18\x06 \x03(\t\"g\n\x11ShowTasksResponse\x12\x1e\n\x05tasks\x18\x01 \x03(\x0b\x32\x0f.taskproto.Task\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error\"\xf0\x04\n\x10\x41lterTaskRequest\x12\x11\n\ttask_name\x18\x01 \x01(\t\x12\x11\n\ttenant_id\x18\x02 \x01(\t\x12\x17\n\nquery_text\x18\x03 \x01(\tH\x00\x88\x01\x01\x12\x14\n\x07\x63omment\x18\x04 \x01(\tH\x01\x88\x01\x01\x12\x42\n\x0f\x61lter_task_type\x18\x05 \x01(\x0e\x32).taskproto.AlterTaskRequest.AlterTaskType\x12\r\n\x05owner\x18\x06 \x01(\t\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x10\n\x08if_exist\x18\x08 \x01(\x08\x12\x36\n\x11warehouse_options\x18\t \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12,\n\x1fsuspend_task_after_num_failures\x18\n \x01(\x05H\x02\x88\x01\x01\x12\x1b\n\x0ewhen_condition\x18\x0b \x01(\tH\x03\x88\x01\x01\x12\x11\n\tadd_after\x18\x0c \x03(\t\x12\x14\n\x0cremove_after\x18\r \x03(\t\"n\n\rAlterTaskType\x12\x0b\n\x07Suspend\x10\x00\x12\n\n\x06Resume\x10\x01\x12\x07\n\x03Set\x10\x02\x12\x0c\n\x08ModifyAs\x10\x03\x12\x0e\n\nModifyWhen\x10\x04\x12\x0c\n\x08\x41\x64\x64\x41\x66ter\x10\x05\x12\x0f\n\x0bRemoveAfter\x10\x06\x42\r\n\x0b_query_textB\n\n\x08_commentB\"\n _suspend_task_after_num_failuresB\x11\n\x0f_when_condition\"f\n\x11\x41lterTaskResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x12\x1d\n\x04task\x18\x02 \x01(\x0b\x32\x0f.taskproto.TaskB\x08\n\x06_error\"\xc1\x01\n\x13ShowTaskRunsRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x1c\n\x14scheduled_time_start\x18\x02 \x01(\t\x12\x1a\n\x12scheduled_time_end\x18\x03 \x01(\t\x12\x14\n\x0cresult_limit\x18\x04 \x01(\x05\x12\x12\n\nerror_only\x18\x05 \x01(\x08\x12\x0e\n\x06owners\x18\x06 \x03(\t\x12\x10\n\x08task_ids\x18\x07 \x03(\t\x12\x11\n\ttask_name\x18\x08 \x01(\t\"\xcc\x04\n\x07TaskRun\x12\x0f\n\x07task_id\x18\x01 \x01(\x04\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x12\n\nquery_text\x18\x04 \x01(\t\x12\x14\n\x07\x63omment\x18\x05 \x01(\tH\x00\x88\x01\x01\x12\r\n\x05owner\x18\x06 \x01(\t\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x0e\n\x06run_id\x18\x08 \x01(\t\x12\x16\n\x0e\x61ttempt_number\x18\t \x01(\x05\x12\x36\n\x11warehouse_options\x18\n \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12\'\n\x05state\x18\x0b \x01(\x0e\x32\x18.taskproto.TaskRun.State\x12\x12\n\nerror_code\x18\x0c \x01(\x03\x12\x1a\n\rerror_message\x18\r \x01(\tH\x01\x88\x01\x01\x12\x16\n\x0escheduled_time\x18\x0e \x01(\t\x12\x1b\n\x0e\x63ompleted_time\x18\x10 \x01(\tH\x02\x88\x01\x01\x12\x10\n\x08query_id\x18\x11 \x01(\t\x12\x16\n\x0e\x63ondition_text\x18\x12 \x01(\t\x12\x14\n\x0croot_task_id\x18\x13 \x01(\t\"O\n\x05State\x12\r\n\tSCHEDULED\x10\x00\x12\r\n\tEXECUTING\x10\x01\x12\r\n\tSUCCEEDED\x10\x02\x12\n\n\x06\x46\x41ILED\x10\x03\x12\r\n\tCANCELLED\x10\x04\x42\n\n\x08_commentB\x10\n\x0e_error_messageB\x11\n\x0f_completed_time\"q\n\x14ShowTaskRunsResponse\x12%\n\ttask_runs\x18\x01 \x03(\x0b\x32\x12.taskproto.TaskRun\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error\"S\n\x18GetTaskDependentsRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x11\n\trecursive\x18\x03 \x01(\x08\"n\n\x19GetTaskDependentsResponse\x12\x1d\n\x04task\x18\x01 \x03(\x0b\x32\x0f.taskproto.Task\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error\"C\n\x1b\x45nableTaskDependentsRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\"R\n\x1c\x45nableTaskDependentsResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error2\xe6\x05\n\x0bTaskService\x12I\n\nCreateTask\x12\x1c.taskproto.CreateTaskRequest\x1a\x1d.taskproto.CreateTaskResponse\x12O\n\x0c\x44\x65scribeTask\x12\x1e.taskproto.DescribeTaskRequest\x1a\x1f.taskproto.DescribeTaskResponse\x12L\n\x0b\x45xecuteTask\x12\x1d.taskproto.ExecuteTaskRequest\x1a\x1e.taskproto.ExecuteTaskResponse\x12\x43\n\x08\x44ropTask\x12\x1a.taskproto.DropTaskRequest\x1a\x1b.taskproto.DropTaskResponse\x12\x46\n\tAlterTask\x12\x1b.taskproto.AlterTaskRequest\x1a\x1c.taskproto.AlterTaskResponse\x12\x46\n\tShowTasks\x12\x1b.taskproto.ShowTasksRequest\x1a\x1c.taskproto.ShowTasksResponse\x12O\n\x0cShowTaskRuns\x12\x1e.taskproto.ShowTaskRunsRequest\x1a\x1f.taskproto.ShowTaskRunsResponse\x12^\n\x11GetTaskDependents\x12#.taskproto.GetTaskDependentsRequest\x1a$.taskproto.GetTaskDependentsResponse\x12g\n\x14\x45nableTaskDependents\x12&.taskproto.EnableTaskDependentsRequest\x1a\'.taskproto.EnableTaskDependentsResponseB!Z\x1f\x64\x61tabend.com/cloudcontrol/protob\x06proto3')
+DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(
+    b'\n\ntask.proto\x12\ttaskproto"\xe9\x01\n\x0fScheduleOptions\x12\x15\n\x08interval\x18\x01 \x01(\x05H\x00\x88\x01\x01\x12\x11\n\x04\x63ron\x18\x02 \x01(\tH\x01\x88\x01\x01\x12\x16\n\ttime_zone\x18\x03 \x01(\tH\x02\x88\x01\x01\x12>\n\rschedule_type\x18\x04 \x01(\x0e\x32\'.taskproto.ScheduleOptions.ScheduleType"0\n\x0cScheduleType\x12\x11\n\rinterval_type\x10\x00\x12\r\n\tcron_type\x10\x01\x42\x0b\n\t_intervalB\x07\n\x05_cronB\x0c\n\n_time_zone"t\n\x10WarehouseOptions\x12\x16\n\twarehouse\x18\x01 \x01(\tH\x00\x88\x01\x01\x12!\n\x14using_warehouse_size\x18\x02 \x01(\tH\x01\x88\x01\x01\x42\x0c\n\n_warehouseB\x17\n\x15_using_warehouse_size"\x9e\x04\n\x11\x43reateTaskRequest\x12\x11\n\ttask_name\x18\x01 \x01(\t\x12\x11\n\ttenant_id\x18\x02 \x01(\t\x12\x12\n\nquery_text\x18\x03 \x01(\t\x12\r\n\x05owner\x18\x04 \x01(\t\x12\x14\n\x07\x63omment\x18\x05 \x01(\tH\x00\x88\x01\x01\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x36\n\x11warehouse_options\x18\x08 \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12,\n\x1fsuspend_task_after_num_failures\x18\t \x01(\x05H\x01\x88\x01\x01\x12\x14\n\x0cif_not_exist\x18\n \x01(\x08\x12\r\n\x05\x61\x66ter\x18\x0b \x03(\t\x12\x1b\n\x0ewhen_condition\x18\x0c \x01(\tH\x02\x88\x01\x01\x12O\n\x12session_parameters\x18\r \x03(\x0b\x32\x33.taskproto.CreateTaskRequest.SessionParametersEntry\x1a\x38\n\x16SessionParametersEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01\x42\n\n\x08_commentB"\n _suspend_task_after_num_failuresB\x11\n\x0f_when_condition"8\n\tTaskError\x12\x0c\n\x04kind\x18\x01 \x01(\t\x12\x0f\n\x07message\x18\x02 \x01(\t\x12\x0c\n\x04\x63ode\x18\x03 \x01(\x05"Y\n\x12\x43reateTaskResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x12\x0f\n\x07task_id\x18\x02 \x01(\x04\x42\x08\n\x06_error"I\n\x0f\x44ropTaskRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x10\n\x08if_exist\x18\x03 \x01(\x08"F\n\x10\x44ropTaskResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error":\n\x12\x45xecuteTaskRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t"I\n\x13\x45xecuteTaskResponse\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error"M\n\x13\x44\x65scribeTaskRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x10\n\x08if_exist\x18\x03 \x01(\x08"\xce\x05\n\x04Task\x12\x0f\n\x07task_id\x18\x01 \x01(\x04\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x12\n\nquery_text\x18\x04 \x01(\t\x12\x14\n\x07\x63omment\x18\x05 \x01(\tH\x00\x88\x01\x01\x12\r\n\x05owner\x18\x06 \x01(\t\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x36\n\x11warehouse_options\x18\x08 \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12\x1e\n\x11next_scheduled_at\x18\t \x01(\tH\x01\x88\x01\x01\x12,\n\x1fsuspend_task_after_num_failures\x18\n \x01(\x05H\x02\x88\x01\x01\x12&\n\x06status\x18\x0c \x01(\x0e\x32\x16.taskproto.Task.Status\x12\x12\n\ncreated_at\x18\x0e \x01(\t\x12\x12\n\nupdated_at\x18\x0f \x01(\t\x12\x1e\n\x11last_suspended_at\x18\x10 \x01(\tH\x03\x88\x01\x01\x12\r\n\x05\x61\x66ter\x18\x11 \x03(\t\x12\x1b\n\x0ewhen_condition\x18\x12 \x01(\tH\x04\x88\x01\x01\x12\x42\n\x12session_parameters\x18\x13 \x03(\x0b\x32&.taskproto.Task.SessionParametersEntry\x1a\x38\n\x16SessionParametersEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01"$\n\x06Status\x12\r\n\tSuspended\x10\x00\x12\x0b\n\x07Started\x10\x01\x42\n\n\x08_commentB\x14\n\x12_next_scheduled_atB"\n _suspend_task_after_num_failuresB\x14\n\x12_last_suspended_atB\x11\n\x0f_when_condition"i\n\x14\x44\x65scribeTaskResponse\x12\x1d\n\x04task\x18\x01 \x01(\x0b\x32\x0f.taskproto.Task\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error"p\n\x10ShowTasksRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\tname_like\x18\x02 \x01(\t\x12\x14\n\x0cresult_limit\x18\x04 \x01(\x05\x12\x0e\n\x06owners\x18\x05 \x03(\t\x12\x10\n\x08task_ids\x18\x06 \x03(\t"g\n\x11ShowTasksResponse\x12\x1e\n\x05tasks\x18\x01 \x03(\x0b\x32\x0f.taskproto.Task\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error"\x9a\x06\n\x10\x41lterTaskRequest\x12\x11\n\ttask_name\x18\x01 \x01(\t\x12\x11\n\ttenant_id\x18\x02 \x01(\t\x12\x17\n\nquery_text\x18\x03 \x01(\tH\x00\x88\x01\x01\x12\x14\n\x07\x63omment\x18\x04 \x01(\tH\x01\x88\x01\x01\x12\x42\n\x0f\x61lter_task_type\x18\x05 \x01(\x0e\x32).taskproto.AlterTaskRequest.AlterTaskType\x12\r\n\x05owner\x18\x06 \x01(\t\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x10\n\x08if_exist\x18\x08 \x01(\x08\x12\x36\n\x11warehouse_options\x18\t \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12,\n\x1fsuspend_task_after_num_failures\x18\n \x01(\x05H\x02\x88\x01\x01\x12\x1b\n\x0ewhen_condition\x18\x0b \x01(\tH\x03\x88\x01\x01\x12\x11\n\tadd_after\x18\x0c \x03(\t\x12\x14\n\x0cremove_after\x18\r \x03(\t\x12\x1e\n\x16set_session_parameters\x18\x0e \x01(\x08\x12N\n\x12session_parameters\x18\x0f \x03(\x0b\x32\x32.taskproto.AlterTaskRequest.SessionParametersEntry\x1a\x38\n\x16SessionParametersEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01"n\n\rAlterTaskType\x12\x0b\n\x07Suspend\x10\x00\x12\n\n\x06Resume\x10\x01\x12\x07\n\x03Set\x10\x02\x12\x0c\n\x08ModifyAs\x10\x03\x12\x0e\n\nModifyWhen\x10\x04\x12\x0c\n\x08\x41\x64\x64\x41\x66ter\x10\x05\x12\x0f\n\x0bRemoveAfter\x10\x06\x42\r\n\x0b_query_textB\n\n\x08_commentB"\n _suspend_task_after_num_failuresB\x11\n\x0f_when_condition"f\n\x11\x41lterTaskResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x12\x1d\n\x04task\x18\x02 \x01(\x0b\x32\x0f.taskproto.TaskB\x08\n\x06_error"\xc1\x01\n\x13ShowTaskRunsRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x1c\n\x14scheduled_time_start\x18\x02 \x01(\t\x12\x1a\n\x12scheduled_time_end\x18\x03 \x01(\t\x12\x14\n\x0cresult_limit\x18\x04 \x01(\x05\x12\x12\n\nerror_only\x18\x05 \x01(\x08\x12\x0e\n\x06owners\x18\x06 \x03(\t\x12\x10\n\x08task_ids\x18\x07 \x03(\t\x12\x11\n\ttask_name\x18\x08 \x01(\t"\xcd\x05\n\x07TaskRun\x12\x0f\n\x07task_id\x18\x01 \x01(\x04\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x12\n\nquery_text\x18\x04 \x01(\t\x12\x14\n\x07\x63omment\x18\x05 \x01(\tH\x00\x88\x01\x01\x12\r\n\x05owner\x18\x06 \x01(\t\x12\x34\n\x10schedule_options\x18\x07 \x01(\x0b\x32\x1a.taskproto.ScheduleOptions\x12\x0e\n\x06run_id\x18\x08 \x01(\t\x12\x16\n\x0e\x61ttempt_number\x18\t \x01(\x05\x12\x36\n\x11warehouse_options\x18\n \x01(\x0b\x32\x1b.taskproto.WarehouseOptions\x12\'\n\x05state\x18\x0b \x01(\x0e\x32\x18.taskproto.TaskRun.State\x12\x12\n\nerror_code\x18\x0c \x01(\x03\x12\x1a\n\rerror_message\x18\r \x01(\tH\x01\x88\x01\x01\x12\x16\n\x0escheduled_time\x18\x0e \x01(\t\x12\x1b\n\x0e\x63ompleted_time\x18\x10 \x01(\tH\x02\x88\x01\x01\x12\x10\n\x08query_id\x18\x11 \x01(\t\x12\x16\n\x0e\x63ondition_text\x18\x12 \x01(\t\x12\x14\n\x0croot_task_id\x18\x13 \x01(\t\x12\x45\n\x12session_parameters\x18\x14 \x03(\x0b\x32).taskproto.TaskRun.SessionParametersEntry\x1a\x38\n\x16SessionParametersEntry\x12\x0b\n\x03key\x18\x01 \x01(\t\x12\r\n\x05value\x18\x02 \x01(\t:\x02\x38\x01"O\n\x05State\x12\r\n\tSCHEDULED\x10\x00\x12\r\n\tEXECUTING\x10\x01\x12\r\n\tSUCCEEDED\x10\x02\x12\n\n\x06\x46\x41ILED\x10\x03\x12\r\n\tCANCELLED\x10\x04\x42\n\n\x08_commentB\x10\n\x0e_error_messageB\x11\n\x0f_completed_time"q\n\x14ShowTaskRunsResponse\x12%\n\ttask_runs\x18\x01 \x03(\x0b\x32\x12.taskproto.TaskRun\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error"S\n\x18GetTaskDependentsRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t\x12\x11\n\trecursive\x18\x03 \x01(\x08"n\n\x19GetTaskDependentsResponse\x12\x1d\n\x04task\x18\x01 \x03(\x0b\x32\x0f.taskproto.Task\x12(\n\x05\x65rror\x18\x02 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error"C\n\x1b\x45nableTaskDependentsRequest\x12\x11\n\ttenant_id\x18\x01 \x01(\t\x12\x11\n\ttask_name\x18\x02 \x01(\t"R\n\x1c\x45nableTaskDependentsResponse\x12(\n\x05\x65rror\x18\x01 \x01(\x0b\x32\x14.taskproto.TaskErrorH\x00\x88\x01\x01\x42\x08\n\x06_error2\xe6\x05\n\x0bTaskService\x12I\n\nCreateTask\x12\x1c.taskproto.CreateTaskRequest\x1a\x1d.taskproto.CreateTaskResponse\x12O\n\x0c\x44\x65scribeTask\x12\x1e.taskproto.DescribeTaskRequest\x1a\x1f.taskproto.DescribeTaskResponse\x12L\n\x0b\x45xecuteTask\x12\x1d.taskproto.ExecuteTaskRequest\x1a\x1e.taskproto.ExecuteTaskResponse\x12\x43\n\x08\x44ropTask\x12\x1a.taskproto.DropTaskRequest\x1a\x1b.taskproto.DropTaskResponse\x12\x46\n\tAlterTask\x12\x1b.taskproto.AlterTaskRequest\x1a\x1c.taskproto.AlterTaskResponse\x12\x46\n\tShowTasks\x12\x1b.taskproto.ShowTasksRequest\x1a\x1c.taskproto.ShowTasksResponse\x12O\n\x0cShowTaskRuns\x12\x1e.taskproto.ShowTaskRunsRequest\x1a\x1f.taskproto.ShowTaskRunsResponse\x12^\n\x11GetTaskDependents\x12#.taskproto.GetTaskDependentsRequest\x1a$.taskproto.GetTaskDependentsResponse\x12g\n\x14\x45nableTaskDependents\x12&.taskproto.EnableTaskDependentsRequest\x1a\'.taskproto.EnableTaskDependentsResponseB!Z\x1f\x64\x61tabend.com/cloudcontrol/protob\x06proto3'
+)
 
 _globals = globals()
 _builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
-_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'task_pb2', _globals)
+_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, "task_pb2", _globals)
 if _descriptor._USE_C_DESCRIPTORS == False:
-  DESCRIPTOR._options = None
-  DESCRIPTOR._serialized_options = b'Z\037databend.com/cloudcontrol/proto'
-  _globals['_SCHEDULEOPTIONS']._serialized_start=26
-  _globals['_SCHEDULEOPTIONS']._serialized_end=259
-  _globals['_SCHEDULEOPTIONS_SCHEDULETYPE']._serialized_start=175
-  _globals['_SCHEDULEOPTIONS_SCHEDULETYPE']._serialized_end=223
-  _globals['_WAREHOUSEOPTIONS']._serialized_start=261
-  _globals['_WAREHOUSEOPTIONS']._serialized_end=377
-  _globals['_CREATETASKREQUEST']._serialized_start=380
-  _globals['_CREATETASKREQUEST']._serialized_end=783
-  _globals['_TASKERROR']._serialized_start=785
-  _globals['_TASKERROR']._serialized_end=841
-  _globals['_CREATETASKRESPONSE']._serialized_start=843
-  _globals['_CREATETASKRESPONSE']._serialized_end=932
-  _globals['_DROPTASKREQUEST']._serialized_start=934
-  _globals['_DROPTASKREQUEST']._serialized_end=1007
-  _globals['_DROPTASKRESPONSE']._serialized_start=1009
-  _globals['_DROPTASKRESPONSE']._serialized_end=1079
-  _globals['_EXECUTETASKREQUEST']._serialized_start=1081
-  _globals['_EXECUTETASKREQUEST']._serialized_end=1139
-  _globals['_EXECUTETASKRESPONSE']._serialized_start=1141
-  _globals['_EXECUTETASKRESPONSE']._serialized_end=1214
-  _globals['_DESCRIBETASKREQUEST']._serialized_start=1216
-  _globals['_DESCRIBETASKREQUEST']._serialized_end=1293
-  _globals['_TASK']._serialized_start=1296
-  _globals['_TASK']._serialized_end=1888
-  _globals['_TASK_STATUS']._serialized_start=1741
-  _globals['_TASK_STATUS']._serialized_end=1777
-  _globals['_DESCRIBETASKRESPONSE']._serialized_start=1890
-  _globals['_DESCRIBETASKRESPONSE']._serialized_end=1995
-  _globals['_SHOWTASKSREQUEST']._serialized_start=1997
-  _globals['_SHOWTASKSREQUEST']._serialized_end=2109
-  _globals['_SHOWTASKSRESPONSE']._serialized_start=2111
-  _globals['_SHOWTASKSRESPONSE']._serialized_end=2214
-  _globals['_ALTERTASKREQUEST']._serialized_start=2217
-  _globals['_ALTERTASKREQUEST']._serialized_end=2841
-  _globals['_ALTERTASKREQUEST_ALTERTASKTYPE']._serialized_start=2649
-  _globals['_ALTERTASKREQUEST_ALTERTASKTYPE']._serialized_end=2759
-  _globals['_ALTERTASKRESPONSE']._serialized_start=2843
-  _globals['_ALTERTASKRESPONSE']._serialized_end=2945
-  _globals['_SHOWTASKRUNSREQUEST']._serialized_start=2948
-  _globals['_SHOWTASKRUNSREQUEST']._serialized_end=3141
-  _globals['_TASKRUN']._serialized_start=3144
-  _globals['_TASKRUN']._serialized_end=3732
-  _globals['_TASKRUN_STATE']._serialized_start=3604
-  _globals['_TASKRUN_STATE']._serialized_end=3683
-  _globals['_SHOWTASKRUNSRESPONSE']._serialized_start=3734
-  _globals['_SHOWTASKRUNSRESPONSE']._serialized_end=3847
-  _globals['_GETTASKDEPENDENTSREQUEST']._serialized_start=3849
-  _globals['_GETTASKDEPENDENTSREQUEST']._serialized_end=3932
-  _globals['_GETTASKDEPENDENTSRESPONSE']._serialized_start=3934
-  _globals['_GETTASKDEPENDENTSRESPONSE']._serialized_end=4044
-  _globals['_ENABLETASKDEPENDENTSREQUEST']._serialized_start=4046
-  _globals['_ENABLETASKDEPENDENTSREQUEST']._serialized_end=4113
-  _globals['_ENABLETASKDEPENDENTSRESPONSE']._serialized_start=4115
-  _globals['_ENABLETASKDEPENDENTSRESPONSE']._serialized_end=4197
-  _globals['_TASKSERVICE']._serialized_start=4200
-  _globals['_TASKSERVICE']._serialized_end=4942
+    DESCRIPTOR._options = None
+    DESCRIPTOR._serialized_options = b"Z\037databend.com/cloudcontrol/proto"
+    _CREATETASKREQUEST_SESSIONPARAMETERSENTRY._options = None
+    _CREATETASKREQUEST_SESSIONPARAMETERSENTRY._serialized_options = b"8\001"
+    _TASK_SESSIONPARAMETERSENTRY._options = None
+    _TASK_SESSIONPARAMETERSENTRY._serialized_options = b"8\001"
+    _ALTERTASKREQUEST_SESSIONPARAMETERSENTRY._options = None
+    _ALTERTASKREQUEST_SESSIONPARAMETERSENTRY._serialized_options = b"8\001"
+    _TASKRUN_SESSIONPARAMETERSENTRY._options = None
+    _TASKRUN_SESSIONPARAMETERSENTRY._serialized_options = b"8\001"
+    _globals["_SCHEDULEOPTIONS"]._serialized_start = 26
+    _globals["_SCHEDULEOPTIONS"]._serialized_end = 259
+    _globals["_SCHEDULEOPTIONS_SCHEDULETYPE"]._serialized_start = 175
+    _globals["_SCHEDULEOPTIONS_SCHEDULETYPE"]._serialized_end = 223
+    _globals["_WAREHOUSEOPTIONS"]._serialized_start = 261
+    _globals["_WAREHOUSEOPTIONS"]._serialized_end = 377
+    _globals["_CREATETASKREQUEST"]._serialized_start = 380
+    _globals["_CREATETASKREQUEST"]._serialized_end = 922
+    _globals["_CREATETASKREQUEST_SESSIONPARAMETERSENTRY"]._serialized_start = 799
+    _globals["_CREATETASKREQUEST_SESSIONPARAMETERSENTRY"]._serialized_end = 855
+    _globals["_TASKERROR"]._serialized_start = 924
+    _globals["_TASKERROR"]._serialized_end = 980
+    _globals["_CREATETASKRESPONSE"]._serialized_start = 982
+    _globals["_CREATETASKRESPONSE"]._serialized_end = 1071
+    _globals["_DROPTASKREQUEST"]._serialized_start = 1073
+    _globals["_DROPTASKREQUEST"]._serialized_end = 1146
+    _globals["_DROPTASKRESPONSE"]._serialized_start = 1148
+    _globals["_DROPTASKRESPONSE"]._serialized_end = 1218
+    _globals["_EXECUTETASKREQUEST"]._serialized_start = 1220
+    _globals["_EXECUTETASKREQUEST"]._serialized_end = 1278
+    _globals["_EXECUTETASKRESPONSE"]._serialized_start = 1280
+    _globals["_EXECUTETASKRESPONSE"]._serialized_end = 1353
+    _globals["_DESCRIBETASKREQUEST"]._serialized_start = 1355
+    _globals["_DESCRIBETASKREQUEST"]._serialized_end = 1432
+    _globals["_TASK"]._serialized_start = 1435
+    _globals["_TASK"]._serialized_end = 2153
+    _globals["_TASK_SESSIONPARAMETERSENTRY"]._serialized_start = 799
+    _globals["_TASK_SESSIONPARAMETERSENTRY"]._serialized_end = 855
+    _globals["_TASK_STATUS"]._serialized_start = 2006
+    _globals["_TASK_STATUS"]._serialized_end = 2042
+    _globals["_DESCRIBETASKRESPONSE"]._serialized_start = 2155
+    _globals["_DESCRIBETASKRESPONSE"]._serialized_end = 2260
+    _globals["_SHOWTASKSREQUEST"]._serialized_start = 2262
+    _globals["_SHOWTASKSREQUEST"]._serialized_end = 2374
+    _globals["_SHOWTASKSRESPONSE"]._serialized_start = 2376
+    _globals["_SHOWTASKSRESPONSE"]._serialized_end = 2479
+    _globals["_ALTERTASKREQUEST"]._serialized_start = 2482
+    _globals["_ALTERTASKREQUEST"]._serialized_end = 3276
+    _globals["_ALTERTASKREQUEST_SESSIONPARAMETERSENTRY"]._serialized_start = 799
+    _globals["_ALTERTASKREQUEST_SESSIONPARAMETERSENTRY"]._serialized_end = 855
+    _globals["_ALTERTASKREQUEST_ALTERTASKTYPE"]._serialized_start = 3084
+    _globals["_ALTERTASKREQUEST_ALTERTASKTYPE"]._serialized_end = 3194
+    _globals["_ALTERTASKRESPONSE"]._serialized_start = 3278
+    _globals["_ALTERTASKRESPONSE"]._serialized_end = 3380
+    _globals["_SHOWTASKRUNSREQUEST"]._serialized_start = 3383
+    _globals["_SHOWTASKRUNSREQUEST"]._serialized_end = 3576
+    _globals["_TASKRUN"]._serialized_start = 3579
+    _globals["_TASKRUN"]._serialized_end = 4296
+    _globals["_TASKRUN_SESSIONPARAMETERSENTRY"]._serialized_start = 799
+    _globals["_TASKRUN_SESSIONPARAMETERSENTRY"]._serialized_end = 855
+    _globals["_TASKRUN_STATE"]._serialized_start = 4168
+    _globals["_TASKRUN_STATE"]._serialized_end = 4247
+    _globals["_SHOWTASKRUNSRESPONSE"]._serialized_start = 4298
+    _globals["_SHOWTASKRUNSRESPONSE"]._serialized_end = 4411
+    _globals["_GETTASKDEPENDENTSREQUEST"]._serialized_start = 4413
+    _globals["_GETTASKDEPENDENTSREQUEST"]._serialized_end = 4496
+    _globals["_GETTASKDEPENDENTSRESPONSE"]._serialized_start = 4498
+    _globals["_GETTASKDEPENDENTSRESPONSE"]._serialized_end = 4608
+    _globals["_ENABLETASKDEPENDENTSREQUEST"]._serialized_start = 4610
+    _globals["_ENABLETASKDEPENDENTSREQUEST"]._serialized_end = 4677
+    _globals["_ENABLETASKDEPENDENTSRESPONSE"]._serialized_start = 4679
+    _globals["_ENABLETASKDEPENDENTSRESPONSE"]._serialized_end = 4761
+    _globals["_TASKSERVICE"]._serialized_start = 4764
+    _globals["_TASKSERVICE"]._serialized_end = 5506
 # @@protoc_insertion_point(module_scope)
diff --git a/tests/cloud_control_server/task_pb2_grpc.py b/tests/cloud_control_server/task_pb2_grpc.py
index 52d0f5c54be2..274c826bb069 100644
--- a/tests/cloud_control_server/task_pb2_grpc.py
+++ b/tests/cloud_control_server/task_pb2_grpc.py
@@ -15,50 +15,50 @@ def __init__(self, channel):
             channel: A grpc.Channel.
         """
         self.CreateTask = channel.unary_unary(
-                '/taskproto.TaskService/CreateTask',
-                request_serializer=task__pb2.CreateTaskRequest.SerializeToString,
-                response_deserializer=task__pb2.CreateTaskResponse.FromString,
-                )
+            "/taskproto.TaskService/CreateTask",
+            request_serializer=task__pb2.CreateTaskRequest.SerializeToString,
+            response_deserializer=task__pb2.CreateTaskResponse.FromString,
+        )
         self.DescribeTask = channel.unary_unary(
-                '/taskproto.TaskService/DescribeTask',
-                request_serializer=task__pb2.DescribeTaskRequest.SerializeToString,
-                response_deserializer=task__pb2.DescribeTaskResponse.FromString,
-                )
+            "/taskproto.TaskService/DescribeTask",
+            request_serializer=task__pb2.DescribeTaskRequest.SerializeToString,
+            response_deserializer=task__pb2.DescribeTaskResponse.FromString,
+        )
         self.ExecuteTask = channel.unary_unary(
-                '/taskproto.TaskService/ExecuteTask',
-                request_serializer=task__pb2.ExecuteTaskRequest.SerializeToString,
-                response_deserializer=task__pb2.ExecuteTaskResponse.FromString,
-                )
+            "/taskproto.TaskService/ExecuteTask",
+            request_serializer=task__pb2.ExecuteTaskRequest.SerializeToString,
+            response_deserializer=task__pb2.ExecuteTaskResponse.FromString,
+        )
         self.DropTask = channel.unary_unary(
-                '/taskproto.TaskService/DropTask',
-                request_serializer=task__pb2.DropTaskRequest.SerializeToString,
-                response_deserializer=task__pb2.DropTaskResponse.FromString,
-                )
+            "/taskproto.TaskService/DropTask",
+            request_serializer=task__pb2.DropTaskRequest.SerializeToString,
+            response_deserializer=task__pb2.DropTaskResponse.FromString,
+        )
         self.AlterTask = channel.unary_unary(
-                '/taskproto.TaskService/AlterTask',
-                request_serializer=task__pb2.AlterTaskRequest.SerializeToString,
-                response_deserializer=task__pb2.AlterTaskResponse.FromString,
-                )
+            "/taskproto.TaskService/AlterTask",
+            request_serializer=task__pb2.AlterTaskRequest.SerializeToString,
+            response_deserializer=task__pb2.AlterTaskResponse.FromString,
+        )
         self.ShowTasks = channel.unary_unary(
-                '/taskproto.TaskService/ShowTasks',
-                request_serializer=task__pb2.ShowTasksRequest.SerializeToString,
-                response_deserializer=task__pb2.ShowTasksResponse.FromString,
-                )
+            "/taskproto.TaskService/ShowTasks",
+            request_serializer=task__pb2.ShowTasksRequest.SerializeToString,
+            response_deserializer=task__pb2.ShowTasksResponse.FromString,
+        )
         self.ShowTaskRuns = channel.unary_unary(
-                '/taskproto.TaskService/ShowTaskRuns',
-                request_serializer=task__pb2.ShowTaskRunsRequest.SerializeToString,
-                response_deserializer=task__pb2.ShowTaskRunsResponse.FromString,
-                )
+            "/taskproto.TaskService/ShowTaskRuns",
+            request_serializer=task__pb2.ShowTaskRunsRequest.SerializeToString,
+            response_deserializer=task__pb2.ShowTaskRunsResponse.FromString,
+        )
         self.GetTaskDependents = channel.unary_unary(
-                '/taskproto.TaskService/GetTaskDependents',
-                request_serializer=task__pb2.GetTaskDependentsRequest.SerializeToString,
-                response_deserializer=task__pb2.GetTaskDependentsResponse.FromString,
-                )
+            "/taskproto.TaskService/GetTaskDependents",
+            request_serializer=task__pb2.GetTaskDependentsRequest.SerializeToString,
+            response_deserializer=task__pb2.GetTaskDependentsResponse.FromString,
+        )
         self.EnableTaskDependents = channel.unary_unary(
-                '/taskproto.TaskService/EnableTaskDependents',
-                request_serializer=task__pb2.EnableTaskDependentsRequest.SerializeToString,
-                response_deserializer=task__pb2.EnableTaskDependentsResponse.FromString,
-                )
+            "/taskproto.TaskService/EnableTaskDependents",
+            request_serializer=task__pb2.EnableTaskDependentsRequest.SerializeToString,
+            response_deserializer=task__pb2.EnableTaskDependentsResponse.FromString,
+        )
 
 
 class TaskServiceServicer(object):
@@ -67,264 +67,373 @@ class TaskServiceServicer(object):
     def CreateTask(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def DescribeTask(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def ExecuteTask(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def DropTask(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def AlterTask(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def ShowTasks(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def ShowTaskRuns(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def GetTaskDependents(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
     def EnableTaskDependents(self, request, context):
         """Missing associated documentation comment in .proto file."""
         context.set_code(grpc.StatusCode.UNIMPLEMENTED)
-        context.set_details('Method not implemented!')
-        raise NotImplementedError('Method not implemented!')
+        context.set_details("Method not implemented!")
+        raise NotImplementedError("Method not implemented!")
 
 
 def add_TaskServiceServicer_to_server(servicer, server):
     rpc_method_handlers = {
-            'CreateTask': grpc.unary_unary_rpc_method_handler(
-                    servicer.CreateTask,
-                    request_deserializer=task__pb2.CreateTaskRequest.FromString,
-                    response_serializer=task__pb2.CreateTaskResponse.SerializeToString,
-            ),
-            'DescribeTask': grpc.unary_unary_rpc_method_handler(
-                    servicer.DescribeTask,
-                    request_deserializer=task__pb2.DescribeTaskRequest.FromString,
-                    response_serializer=task__pb2.DescribeTaskResponse.SerializeToString,
-            ),
-            'ExecuteTask': grpc.unary_unary_rpc_method_handler(
-                    servicer.ExecuteTask,
-                    request_deserializer=task__pb2.ExecuteTaskRequest.FromString,
-                    response_serializer=task__pb2.ExecuteTaskResponse.SerializeToString,
-            ),
-            'DropTask': grpc.unary_unary_rpc_method_handler(
-                    servicer.DropTask,
-                    request_deserializer=task__pb2.DropTaskRequest.FromString,
-                    response_serializer=task__pb2.DropTaskResponse.SerializeToString,
-            ),
-            'AlterTask': grpc.unary_unary_rpc_method_handler(
-                    servicer.AlterTask,
-                    request_deserializer=task__pb2.AlterTaskRequest.FromString,
-                    response_serializer=task__pb2.AlterTaskResponse.SerializeToString,
-            ),
-            'ShowTasks': grpc.unary_unary_rpc_method_handler(
-                    servicer.ShowTasks,
-                    request_deserializer=task__pb2.ShowTasksRequest.FromString,
-                    response_serializer=task__pb2.ShowTasksResponse.SerializeToString,
-            ),
-            'ShowTaskRuns': grpc.unary_unary_rpc_method_handler(
-                    servicer.ShowTaskRuns,
-                    request_deserializer=task__pb2.ShowTaskRunsRequest.FromString,
-                    response_serializer=task__pb2.ShowTaskRunsResponse.SerializeToString,
-            ),
-            'GetTaskDependents': grpc.unary_unary_rpc_method_handler(
-                    servicer.GetTaskDependents,
-                    request_deserializer=task__pb2.GetTaskDependentsRequest.FromString,
-                    response_serializer=task__pb2.GetTaskDependentsResponse.SerializeToString,
-            ),
-            'EnableTaskDependents': grpc.unary_unary_rpc_method_handler(
-                    servicer.EnableTaskDependents,
-                    request_deserializer=task__pb2.EnableTaskDependentsRequest.FromString,
-                    response_serializer=task__pb2.EnableTaskDependentsResponse.SerializeToString,
-            ),
+        "CreateTask": grpc.unary_unary_rpc_method_handler(
+            servicer.CreateTask,
+            request_deserializer=task__pb2.CreateTaskRequest.FromString,
+            response_serializer=task__pb2.CreateTaskResponse.SerializeToString,
+        ),
+        "DescribeTask": grpc.unary_unary_rpc_method_handler(
+            servicer.DescribeTask,
+            request_deserializer=task__pb2.DescribeTaskRequest.FromString,
+            response_serializer=task__pb2.DescribeTaskResponse.SerializeToString,
+        ),
+        "ExecuteTask": grpc.unary_unary_rpc_method_handler(
+            servicer.ExecuteTask,
+            request_deserializer=task__pb2.ExecuteTaskRequest.FromString,
+            response_serializer=task__pb2.ExecuteTaskResponse.SerializeToString,
+        ),
+        "DropTask": grpc.unary_unary_rpc_method_handler(
+            servicer.DropTask,
+            request_deserializer=task__pb2.DropTaskRequest.FromString,
+            response_serializer=task__pb2.DropTaskResponse.SerializeToString,
+        ),
+        "AlterTask": grpc.unary_unary_rpc_method_handler(
+            servicer.AlterTask,
+            request_deserializer=task__pb2.AlterTaskRequest.FromString,
+            response_serializer=task__pb2.AlterTaskResponse.SerializeToString,
+        ),
+        "ShowTasks": grpc.unary_unary_rpc_method_handler(
+            servicer.ShowTasks,
+            request_deserializer=task__pb2.ShowTasksRequest.FromString,
+            response_serializer=task__pb2.ShowTasksResponse.SerializeToString,
+        ),
+        "ShowTaskRuns": grpc.unary_unary_rpc_method_handler(
+            servicer.ShowTaskRuns,
+            request_deserializer=task__pb2.ShowTaskRunsRequest.FromString,
+            response_serializer=task__pb2.ShowTaskRunsResponse.SerializeToString,
+        ),
+        "GetTaskDependents": grpc.unary_unary_rpc_method_handler(
+            servicer.GetTaskDependents,
+            request_deserializer=task__pb2.GetTaskDependentsRequest.FromString,
+            response_serializer=task__pb2.GetTaskDependentsResponse.SerializeToString,
+        ),
+        "EnableTaskDependents": grpc.unary_unary_rpc_method_handler(
+            servicer.EnableTaskDependents,
+            request_deserializer=task__pb2.EnableTaskDependentsRequest.FromString,
+            response_serializer=task__pb2.EnableTaskDependentsResponse.SerializeToString,
+        ),
     }
     generic_handler = grpc.method_handlers_generic_handler(
-            'taskproto.TaskService', rpc_method_handlers)
+        "taskproto.TaskService", rpc_method_handlers
+    )
     server.add_generic_rpc_handlers((generic_handler,))
 
 
- # This class is part of an EXPERIMENTAL API.
+# This class is part of an EXPERIMENTAL API.
 class TaskService(object):
     """Missing associated documentation comment in .proto file."""
 
     @staticmethod
-    def CreateTask(request,
+    def CreateTask(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/CreateTask',
+            "/taskproto.TaskService/CreateTask",
             task__pb2.CreateTaskRequest.SerializeToString,
             task__pb2.CreateTaskResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def DescribeTask(request,
+    def DescribeTask(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/DescribeTask',
+            "/taskproto.TaskService/DescribeTask",
             task__pb2.DescribeTaskRequest.SerializeToString,
             task__pb2.DescribeTaskResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def ExecuteTask(request,
+    def ExecuteTask(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/ExecuteTask',
+            "/taskproto.TaskService/ExecuteTask",
             task__pb2.ExecuteTaskRequest.SerializeToString,
             task__pb2.ExecuteTaskResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def DropTask(request,
+    def DropTask(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/DropTask',
+            "/taskproto.TaskService/DropTask",
             task__pb2.DropTaskRequest.SerializeToString,
             task__pb2.DropTaskResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def AlterTask(request,
+    def AlterTask(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/AlterTask',
+            "/taskproto.TaskService/AlterTask",
             task__pb2.AlterTaskRequest.SerializeToString,
             task__pb2.AlterTaskResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def ShowTasks(request,
+    def ShowTasks(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/ShowTasks',
+            "/taskproto.TaskService/ShowTasks",
             task__pb2.ShowTasksRequest.SerializeToString,
             task__pb2.ShowTasksResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def ShowTaskRuns(request,
+    def ShowTaskRuns(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/ShowTaskRuns',
+            "/taskproto.TaskService/ShowTaskRuns",
             task__pb2.ShowTaskRunsRequest.SerializeToString,
             task__pb2.ShowTaskRunsResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def GetTaskDependents(request,
+    def GetTaskDependents(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/GetTaskDependents',
+            "/taskproto.TaskService/GetTaskDependents",
             task__pb2.GetTaskDependentsRequest.SerializeToString,
             task__pb2.GetTaskDependentsResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
 
     @staticmethod
-    def EnableTaskDependents(request,
+    def EnableTaskDependents(
+        request,
+        target,
+        options=(),
+        channel_credentials=None,
+        call_credentials=None,
+        insecure=False,
+        compression=None,
+        wait_for_ready=None,
+        timeout=None,
+        metadata=None,
+    ):
+        return grpc.experimental.unary_unary(
+            request,
             target,
-            options=(),
-            channel_credentials=None,
-            call_credentials=None,
-            insecure=False,
-            compression=None,
-            wait_for_ready=None,
-            timeout=None,
-            metadata=None):
-        return grpc.experimental.unary_unary(request, target, '/taskproto.TaskService/EnableTaskDependents',
+            "/taskproto.TaskService/EnableTaskDependents",
             task__pb2.EnableTaskDependentsRequest.SerializeToString,
             task__pb2.EnableTaskDependentsResponse.FromString,
-            options, channel_credentials,
-            insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
+            options,
+            channel_credentials,
+            insecure,
+            call_credentials,
+            compression,
+            wait_for_ready,
+            timeout,
+            metadata,
+        )
diff --git a/tests/sqllogictests/suites/base/03_common/03_0034_select_star_columns.test b/tests/sqllogictests/suites/base/03_common/03_0034_select_star_columns.test
index 55146b7c7155..c9880e14f3c4 100644
--- a/tests/sqllogictests/suites/base/03_common/03_0034_select_star_columns.test
+++ b/tests/sqllogictests/suites/base/03_common/03_0034_select_star_columns.test
@@ -257,4 +257,4 @@ drop table default.t
 
 
 statement ok
-drop table default.t1
\ No newline at end of file
+drop table default.t1
diff --git a/tests/sqllogictests/suites/base/09_fuse_engine/09_0026_merge_into.test b/tests/sqllogictests/suites/base/09_fuse_engine/09_0026_merge_into.test
index e2f26d9ebc90..b56efaf7794d 100644
--- a/tests/sqllogictests/suites/base/09_fuse_engine/09_0026_merge_into.test
+++ b/tests/sqllogictests/suites/base/09_fuse_engine/09_0026_merge_into.test
@@ -1,13 +1,7 @@
 statement ok
 set enable_distributed_merge_into = 1;
 
-statement ok
-set enable_runtime_filter = 1;
-
 include ./09_0036_merge_into_without_distributed_enable.test
 
 statement ok
 set enable_distributed_merge_into = 0;
-
-statement ok
-set enable_runtime_filter = 0;
diff --git a/tests/sqllogictests/suites/mode/cluster/memo/aggregate_property.test b/tests/sqllogictests/suites/mode/cluster/memo/aggregate_property.test
index 36ded546c793..ed22625675bc 100644
--- a/tests/sqllogictests/suites/mode/cluster/memo/aggregate_property.test
+++ b/tests/sqllogictests/suites/mode/cluster/memo/aggregate_property.test
@@ -7,6 +7,9 @@ create database aggregate_property
 statement ok
 use aggregate_property
 
+statement ok
+set max_threads = 1
+
 statement ok
 create table t_10(a int) as select * from numbers(10)
 
@@ -23,22 +26,22 @@ where t_10.a = t_1000.a and t_100.a = t_1000.a
 ----
 Memo
 ├── root group: #8
-├── estimated memory: 9912 bytes
+├── estimated memory: 10560 bytes
 ├── Group #0
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
-│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
+│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#0]
+│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#0]
 │   └── #2 Exchange: (Merge) [#0]
 ├── Group #1
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#1]
+│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#1]
 │   └── #2 Exchange: (Merge) [#1]
 ├── Group #2
 │   ├── Best properties
@@ -52,14 +55,16 @@ Memo
 │   └── #3 Exchange: (Merge) [#2]
 ├── Group #3
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 3110.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   │   ├── { dist: Broadcast }: expr: #1, cost: 4110.000, children: [{ dist: Any }]
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #0, cost: 54110.000, children: [{ dist: Hash(t_1000.a (#2)::Int32 NULL) }, { dist: Hash(t_10.a (#0)::Int32 NULL) }]
-│   │   └── { dist: Serial }: expr: #0, cost: 355610.000, children: [{ dist: Serial }, { dist: Serial }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1310.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   │   ├── { dist: Broadcast }: expr: #1, cost: 2310.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #3, cost: 1820.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #0, cost: 6410.000, children: [{ dist: Hash(t_100.a (#1)::Int32 NULL) }, { dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Serial }: expr: #0, cost: 38810.000, children: [{ dist: Serial }, { dist: Serial }]
 │   ├── #0 Join [#1, #2]
 │   ├── #1 Exchange: (Broadcast) [#3]
 │   ├── #2 Exchange: (Merge) [#3]
-│   └── #3 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#3]
+│   ├── #3 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#3]
+│   └── #4 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#3]
 ├── Group #4
 │   ├── Best properties
 │   │   └── { dist: Any }: expr: #0, cost: 4410.000, children: [{ dist: Any }, { dist: Broadcast }]
@@ -92,22 +97,22 @@ group by t_10.a, t_100.a
 ----
 Memo
 ├── root group: #8
-├── estimated memory: 23128 bytes
+├── estimated memory: 24480 bytes
 ├── Group #0
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
-│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
+│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#0]
+│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#0]
 │   └── #2 Exchange: (Merge) [#0]
 ├── Group #1
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#1]
+│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#1]
 │   └── #2 Exchange: (Merge) [#1]
 ├── Group #2
 │   ├── Best properties
@@ -121,14 +126,16 @@ Memo
 │   └── #3 Exchange: (Merge) [#2]
 ├── Group #3
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 3110.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   │   ├── { dist: Broadcast }: expr: #1, cost: 4110.000, children: [{ dist: Any }]
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #0, cost: 53620.000, children: [{ dist: Hash(t_1000.a (#2)::Int32 NULL) }, { dist: Hash(t_10.a (#0)::Int32 NULL) }]
-│   │   └── { dist: Serial }: expr: #0, cost: 355610.000, children: [{ dist: Serial }, { dist: Serial }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1310.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   │   ├── { dist: Broadcast }: expr: #1, cost: 2310.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #3, cost: 1820.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #0, cost: 5920.000, children: [{ dist: Hash(t_100.a (#1)::Int32 NULL) }, { dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Serial }: expr: #0, cost: 38810.000, children: [{ dist: Serial }, { dist: Serial }]
 │   ├── #0 Join [#1, #2]
 │   ├── #1 Exchange: (Broadcast) [#3]
 │   ├── #2 Exchange: (Merge) [#3]
-│   └── #3 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#3]
+│   ├── #3 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#3]
+│   └── #4 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#3]
 ├── Group #4
 │   ├── Best properties
 │   │   └── { dist: Any }: expr: #0, cost: 4410.000, children: [{ dist: Any }, { dist: Broadcast }]
@@ -158,85 +165,87 @@ Memo
 │   └── #4 Exchange: (Merge) [#8]
 ├── Group #9
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 3120.000, children: [{ dist: Any }]
-│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 3630.000, children: [{ dist: Any }]
-│   ├── #0 EvalScalar [#3]
-│   └── #1 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#9]
+│   │   ├── { dist: Any }: expr: #0, cost: 2000.000, children: [{ dist: Any }]
+│   │   └── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 53000.000, children: [{ dist: Any }]
+│   ├── #0 EvalScalar [#0]
+│   └── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#9]
 ├── Group #10
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 3680.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Any }: expr: #0, cost: 58000.000, children: [{ dist: Hash(t_1000.a (#2)::Int32 NULL) }]
 │   └── #0 Aggregate [#9]
 ├── Group #11
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 3730.000, children: [{ dist: Any }]
-│   │   ├── { dist: Broadcast }: expr: #1, cost: 4730.000, children: [{ dist: Any }]
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #3, cost: 4240.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 7230.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 63000.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 114000.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 413000.000, children: [{ dist: Any }]
 │   ├── #0 Aggregate [#10]
-│   ├── #1 Exchange: (Broadcast) [#11]
-│   ├── #2 Exchange: (Merge) [#11]
-│   └── #3 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#11]
+│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#11]
+│   └── #2 Exchange: (Merge) [#11]
 ├── Group #12
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 5030.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 5489.000, children: [{ dist: Any }]
-│   ├── #0 Join [#0, #11]
+│   │   ├── { dist: Any }: expr: #0, cost: 66410.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 66869.000, children: [{ dist: Any }]
+│   ├── #0 Join [#11, #3]
 │   └── #1 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#12]
 ├── Group #13
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 5534.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Any }: expr: #0, cost: 66914.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
 │   └── #0 Aggregate [#12]
 ├── Group #14
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 5579.000, children: [{ dist: Any }]
+│   │   └── { dist: Any }: expr: #0, cost: 66959.000, children: [{ dist: Any }]
 │   └── #0 Aggregate [#13]
 ├── Group #15
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 5700.000, children: [{ dist: Hash(t_100.a (#1)::Int32 NULL) }]
-│   └── #0 Aggregate [#0]
+│   │   └── { dist: Any }: expr: #0, cost: 1870.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   └── #0 Aggregate [#3]
 ├── Group #16
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 6200.000, children: [{ dist: Any }]
-│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 11300.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 41200.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1920.000, children: [{ dist: Any }]
+│   │   ├── { dist: Broadcast }: expr: #3, cost: 2920.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 2430.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #2, cost: 2430.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #4, cost: 5420.000, children: [{ dist: Any }]
 │   ├── #0 Aggregate [#15]
-│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#16]
-│   └── #2 Exchange: (Merge) [#16]
+│   ├── #1 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#16]
+│   ├── #2 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#16]
+│   ├── #3 Exchange: (Broadcast) [#16]
+│   └── #4 Exchange: (Merge) [#16]
 ├── Group #17
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 10510.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   └── #0 Join [#16, #3]
+│   │   └── { dist: Any }: expr: #0, cost: 5020.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   └── #0 Join [#0, #16]
 ├── Group #18
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 10519.000, children: [{ dist: Any }]
-│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 10978.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 5029.000, children: [{ dist: Any }]
+│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 5488.000, children: [{ dist: Any }]
 │   ├── #0 EvalScalar [#17]
 │   └── #1 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#18]
 ├── Group #19
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 11023.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Any }: expr: #0, cost: 5533.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
 │   └── #0 Aggregate [#18]
 ├── Group #20
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 11068.000, children: [{ dist: Any }]
+│   │   └── { dist: Any }: expr: #0, cost: 5578.000, children: [{ dist: Any }]
 │   └── #0 Aggregate [#19]
 ├── Group #21
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 11130.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   └── #0 Join [#16, #11]
+│   │   └── { dist: Any }: expr: #0, cost: 67020.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   └── #0 Join [#11, #16]
 ├── Group #22
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 11139.000, children: [{ dist: Any }]
-│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 11598.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 67029.000, children: [{ dist: Any }]
+│   │   └── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #1, cost: 67488.000, children: [{ dist: Any }]
 │   ├── #0 EvalScalar [#21]
 │   └── #1 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#22]
 ├── Group #23
 │   ├── Best properties
-│   │   └── { dist: Any }: expr: #0, cost: 11643.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Any }: expr: #0, cost: 67533.000, children: [{ dist: Hash(t_10.a (#0)::Int32 NULL) }]
 │   └── #0 Aggregate [#22]
 └── Group #24
     ├── Best properties
-    │   └── { dist: Any }: expr: #0, cost: 11688.000, children: [{ dist: Any }]
+    │   └── { dist: Any }: expr: #0, cost: 67578.000, children: [{ dist: Any }]
     └── #0 Aggregate [#23]
 
 
diff --git a/tests/sqllogictests/suites/mode/cluster/memo/join_property.test b/tests/sqllogictests/suites/mode/cluster/memo/join_property.test
index 8b2d4577d15e..eda22f0a7791 100644
--- a/tests/sqllogictests/suites/mode/cluster/memo/join_property.test
+++ b/tests/sqllogictests/suites/mode/cluster/memo/join_property.test
@@ -7,6 +7,9 @@ create database join_property
 statement ok
 use join_property
 
+statement ok
+set max_threads = 1
+
 statement ok
 create table t_10(a int) as select * from numbers(10)
 
@@ -22,22 +25,22 @@ select * from t_10, t_100, t_1000 where t_10.a = t_1000.a and t_100.a = t_1000.a
 ----
 Memo
 ├── root group: #5
-├── estimated memory: 8024 bytes
+├── estimated memory: 8640 bytes
 ├── Group #0
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
-│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
+│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#0]
+│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#0]
 │   └── #2 Exchange: (Merge) [#0]
 ├── Group #1
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#1]
+│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#1]
 │   └── #2 Exchange: (Merge) [#1]
 ├── Group #2
 │   ├── Best properties
@@ -51,14 +54,16 @@ Memo
 │   └── #3 Exchange: (Merge) [#2]
 ├── Group #3
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 3110.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   │   ├── { dist: Broadcast }: expr: #1, cost: 4110.000, children: [{ dist: Any }]
-│   │   ├── { dist: Hash(t_1000.a (#2)::Int32 NULL) }: expr: #0, cost: 54110.000, children: [{ dist: Hash(t_1000.a (#2)::Int32 NULL) }, { dist: Hash(t_10.a (#0)::Int32 NULL) }]
-│   │   └── { dist: Serial }: expr: #0, cost: 355610.000, children: [{ dist: Serial }, { dist: Serial }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1310.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   │   ├── { dist: Broadcast }: expr: #1, cost: 2310.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_10.a (#0)::Int32 NULL) }: expr: #3, cost: 1820.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #0, cost: 6410.000, children: [{ dist: Hash(t_100.a (#1)::Int32 NULL) }, { dist: Hash(t_10.a (#0)::Int32 NULL) }]
+│   │   └── { dist: Serial }: expr: #0, cost: 38810.000, children: [{ dist: Serial }, { dist: Serial }]
 │   ├── #0 Join [#1, #2]
 │   ├── #1 Exchange: (Broadcast) [#3]
 │   ├── #2 Exchange: (Merge) [#3]
-│   └── #3 Exchange: (Hash(t_1000.a (#2)::Int32 NULL)) [#3]
+│   ├── #3 Exchange: (Hash(t_10.a (#0)::Int32 NULL)) [#3]
+│   └── #4 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#3]
 ├── Group #4
 │   ├── Best properties
 │   │   └── { dist: Any }: expr: #0, cost: 4410.000, children: [{ dist: Any }, { dist: Broadcast }]
@@ -76,7 +81,7 @@ select * from t_1000 left join t_10 on t_1000.a = t_10.a left join t_100 on t_10
 ----
 Memo
 ├── root group: #5
-├── estimated memory: 8024 bytes
+├── estimated memory: 8160 bytes
 ├── Group #0
 │   ├── Best properties
 │   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
@@ -130,7 +135,7 @@ select * from t_1000 right join t_10 on t_1000.a = t_10.a right join t_100 on t_
 ----
 Memo
 ├── root group: #5
-├── estimated memory: 7080 bytes
+├── estimated memory: 7200 bytes
 ├── Group #0
 │   ├── Best properties
 │   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
@@ -180,7 +185,7 @@ select * from t_1000 full join t_10 on t_1000.a = t_10.a full join t_100 on t_10
 ----
 Memo
 ├── root group: #5
-├── estimated memory: 7080 bytes
+├── estimated memory: 7200 bytes
 ├── Group #0
 │   ├── Best properties
 │   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
@@ -230,7 +235,7 @@ select * from t_10, t_100, t_1000
 ----
 Memo
 ├── root group: #5
-├── estimated memory: 6136 bytes
+├── estimated memory: 6240 bytes
 ├── Group #0
 │   ├── Best properties
 │   │   ├── { dist: Any }: expr: #0, cost: 10.000, children: []
diff --git a/tests/sqllogictests/suites/mode/cluster/memo/mix_property.test b/tests/sqllogictests/suites/mode/cluster/memo/mix_property.test
index 4bdc095fa684..14d8473afbb1 100644
--- a/tests/sqllogictests/suites/mode/cluster/memo/mix_property.test
+++ b/tests/sqllogictests/suites/mode/cluster/memo/mix_property.test
@@ -7,6 +7,9 @@ create database mix_property
 statement ok
 use mix_property
 
+statement ok
+set max_threads = 1
+
 statement ok
 create table t_10(a int) as select * from numbers(10)
 
@@ -26,22 +29,22 @@ limit 10
 ----
 Memo
 ├── root group: #10
-├── estimated memory: 10856 bytes
+├── estimated memory: 11520 bytes
 ├── Group #0
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
-│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
+│   │   ├── { dist: Hash(t_1000.a (#0)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#0]
+│   ├── #1 Exchange: (Hash(t_1000.a (#0)::Int32 NULL)) [#0]
 │   └── #2 Exchange: (Merge) [#0]
 ├── Group #1
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 1000.000, children: []
-│   │   ├── { dist: Hash(t_1000.a (#0)::Int32 NULL) }: expr: #1, cost: 52000.000, children: [{ dist: Any }]
-│   │   └── { dist: Serial }: expr: #2, cost: 351000.000, children: [{ dist: Any }]
+│   │   ├── { dist: Any }: expr: #0, cost: 100.000, children: []
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #1, cost: 5200.000, children: [{ dist: Any }]
+│   │   └── { dist: Serial }: expr: #2, cost: 35100.000, children: [{ dist: Any }]
 │   ├── #0 Scan []
-│   ├── #1 Exchange: (Hash(t_1000.a (#0)::Int32 NULL)) [#1]
+│   ├── #1 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#1]
 │   └── #2 Exchange: (Merge) [#1]
 ├── Group #2
 │   ├── Best properties
@@ -55,14 +58,16 @@ Memo
 │   └── #3 Exchange: (Merge) [#2]
 ├── Group #3
 │   ├── Best properties
-│   │   ├── { dist: Any }: expr: #0, cost: 3110.000, children: [{ dist: Any }, { dist: Broadcast }]
-│   │   ├── { dist: Broadcast }: expr: #1, cost: 4110.000, children: [{ dist: Any }]
-│   │   ├── { dist: Hash(t_1000.a (#0)::Int32 NULL) }: expr: #0, cost: 54110.000, children: [{ dist: Hash(t_1000.a (#0)::Int32 NULL) }, { dist: Hash(t_10.a (#2)::Int32 NULL) }]
-│   │   └── { dist: Serial }: expr: #0, cost: 355610.000, children: [{ dist: Serial }, { dist: Serial }]
+│   │   ├── { dist: Any }: expr: #0, cost: 1310.000, children: [{ dist: Any }, { dist: Broadcast }]
+│   │   ├── { dist: Broadcast }: expr: #1, cost: 2310.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_10.a (#2)::Int32 NULL) }: expr: #3, cost: 1820.000, children: [{ dist: Any }]
+│   │   ├── { dist: Hash(t_100.a (#1)::Int32 NULL) }: expr: #0, cost: 6410.000, children: [{ dist: Hash(t_100.a (#1)::Int32 NULL) }, { dist: Hash(t_10.a (#2)::Int32 NULL) }]
+│   │   └── { dist: Serial }: expr: #0, cost: 38810.000, children: [{ dist: Serial }, { dist: Serial }]
 │   ├── #0 Join [#1, #2]
 │   ├── #1 Exchange: (Broadcast) [#3]
 │   ├── #2 Exchange: (Merge) [#3]
-│   └── #3 Exchange: (Hash(t_1000.a (#0)::Int32 NULL)) [#3]
+│   ├── #3 Exchange: (Hash(t_10.a (#2)::Int32 NULL)) [#3]
+│   └── #4 Exchange: (Hash(t_100.a (#1)::Int32 NULL)) [#3]
 ├── Group #4
 │   ├── Best properties
 │   │   └── { dist: Any }: expr: #0, cost: 4410.000, children: [{ dist: Any }, { dist: Broadcast }]
diff --git a/tests/sqllogictests/suites/mode/cluster/merge_into_non_equal_distributed.test b/tests/sqllogictests/suites/mode/cluster/merge_into_non_equal_distributed.test
new file mode 100644
index 000000000000..2d8c9c1af6c2
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/cluster/merge_into_non_equal_distributed.test
@@ -0,0 +1,103 @@
+statement ok
+set enable_experimental_merge_into = 1;
+
+statement ok
+set enable_distributed_merge_into = 1;
+
+statement ok
+create table t1(a int);
+
+statement ok
+create table t2(a int);
+
+statement ok
+insert into t1 values(1),(2),(3),(4),(5);
+
+statement ok
+insert into t1 values(6),(7),(8),(9),(10);
+
+statement ok
+insert into t1 values(11),(12),(13),(14),(15);
+
+statement ok
+insert into t2 values(8);
+
+query t
+select * from t1 order by a;
+----
+1
+2
+3
+4
+5
+6
+7
+8
+9
+10
+11
+12
+13
+14
+15
+
+query T
+select * from t2 order by a;
+----
+8
+
+## check there is no add row_number.
+query T
+explain merge into t1 using t2 on t1.a < t2.a when matched then update * when not matched then insert *;
+----
+MergeInto:
+target_table: default.default.t1
+├── distributed: false
+├── target_build_optimization: false
+├── matched update: [condition: None,update set a = a (#0)]
+├── unmatched insert: [condition: None,insert into (a) values(CAST(a (#0) AS Int32 NULL))]
+└── HashJoin: RIGHT OUTER
+    ├── equi conditions: []
+    ├── non-equi conditions: [lt(t1.a (#1), t2.a (#0))]
+    ├── Exchange(Merge)
+    │   └── LogicalGet
+    │       ├── table: default.default.t1
+    │       ├── filters: []
+    │       ├── order by: []
+    │       └── limit: NONE
+    └── Exchange(Merge)
+        └── LogicalGet
+            ├── table: default.default.t2
+            ├── filters: []
+            ├── order by: []
+            └── limit: NONE
+
+query TT
+merge into t1 using t2 on t1.a < t2.a when matched then update * when not matched then insert *;
+----
+0 7
+
+query T
+select * from t1 order by a;
+----
+8
+8
+8
+8
+8
+8
+8
+8
+9
+10
+11
+12
+13
+14
+15
+
+statement ok
+set enable_experimental_merge_into = 0;
+
+statement ok
+set enable_distributed_merge_into = 0;
\ No newline at end of file
diff --git a/tests/sqllogictests/suites/mode/standalone/ee/explain_agg_index.test b/tests/sqllogictests/suites/mode/standalone/ee/explain_agg_index.test
index 95986f444c75..20774352bea3 100644
--- a/tests/sqllogictests/suites/mode/standalone/ee/explain_agg_index.test
+++ b/tests/sqllogictests/suites/mode/standalone/ee/explain_agg_index.test
@@ -154,67 +154,61 @@ EvalScalar
                 ├── partitions scanned: 0
                 ├── push downs: [filters: [t1.b (#1) > 5], limit: NONE]
                 ├── aggregating index: [SELECT b, SUM(a) FROM test_index_db.t1 WHERE (b > 3) GROUP BY b]
-                ├── rewritten query: [selection: [index_col_0 (#0), index_col_1 (#1)], filter: index_col_0 (#0) > to_int32(5)]
+                ├── rewritten query: [selection: [index_col_0 (#0), index_col_1 (#1)], filter: index_col_0 (#0) > 5]
                 └── estimated rows: 0.00
 
 query T
 EXPLAIN SELECT t1.b, SUM(a) FROM t1 GROUP BY t1.b HAVING SUM(a)=(SELECT SUM(a) FROM t1 t WHERE t1.b=t.b and t.b > 3)
 ----
-Filter
+HashJoin
 ├── output columns: [SUM(a) (#2), t1.b (#1)]
-├── filters: [is_true(SUM(a) (#2) = scalar_subquery_5 (#5))]
+├── join type: INNER
+├── build keys: [scalar_subquery_5 (#5), b (#4)]
+├── probe keys: [SUM(a) (#2), b (#1)]
+├── filters: []
 ├── estimated rows: 0.00
-└── HashJoin
-    ├── output columns: [SUM(a) (#2), t1.b (#1), SUM(a) (#5)]
-    ├── join type: LEFT SINGLE
-    ├── build keys: [b (#4)]
-    ├── probe keys: [CAST(b (#1) AS Int32 NULL)]
-    ├── filters: []
+├── AggregateFinal(Build)
+│   ├── output columns: [SUM(a) (#5), t.b (#4)]
+│   ├── group by: [b]
+│   ├── aggregate functions: [sum(a)]
+│   ├── estimated rows: 0.00
+│   └── AggregatePartial
+│       ├── group by: [b]
+│       ├── aggregate functions: [sum(a)]
+│       ├── estimated rows: 0.00
+│       └── Filter
+│           ├── output columns: [t.a (#3), t.b (#4)]
+│           ├── filters: [b (#4) > 3]
+│           ├── estimated rows: 0.00
+│           └── TableScan
+│               ├── table: default.test_index_db.t1
+│               ├── output columns: [a (#3), b (#4)]
+│               ├── read rows: 0
+│               ├── read bytes: 0
+│               ├── partitions total: 0
+│               ├── partitions scanned: 0
+│               ├── push downs: [filters: [t1.b (#4) > 3], limit: NONE]
+│               ├── aggregating index: [SELECT b, SUM(a) FROM test_index_db.t1 WHERE b > 3 GROUP BY b]
+│               ├── rewritten query: [selection: [index_col_0 (#0), index_col_1 (#1)]]
+│               └── estimated rows: 0.00
+└── AggregateFinal(Probe)
+    ├── output columns: [SUM(a) (#2), t1.b (#1)]
+    ├── group by: [b]
+    ├── aggregate functions: [sum(a)]
     ├── estimated rows: 0.00
-    ├── AggregateFinal(Build)
-    │   ├── output columns: [SUM(a) (#5), t.b (#4)]
-    │   ├── group by: [b]
-    │   ├── aggregate functions: [sum(a)]
-    │   ├── estimated rows: 0.00
-    │   └── AggregatePartial
-    │       ├── output columns: [SUM(a) (#5), #_group_by_key]
-    │       ├── group by: [b]
-    │       ├── aggregate functions: [sum(a)]
-    │       ├── estimated rows: 0.00
-    │       └── Filter
-    │           ├── output columns: [t.a (#3), t.b (#4)]
-    │           ├── filters: [t.b (#4) > 3]
-    │           ├── estimated rows: 0.00
-    │           └── TableScan
-    │               ├── table: default.test_index_db.t1
-    │               ├── output columns: [a (#3), b (#4)]
-    │               ├── read rows: 0
-    │               ├── read bytes: 0
-    │               ├── partitions total: 0
-    │               ├── partitions scanned: 0
-    │               ├── push downs: [filters: [t1.b (#4) > 3], limit: NONE]
-    │               ├── aggregating index: [SELECT b, SUM(a) FROM test_index_db.t1 WHERE (b > 3) GROUP BY b]
-    │               ├── rewritten query: [selection: [index_col_0 (#0), index_col_1 (#1)]]
-    │               └── estimated rows: 0.00
-    └── AggregateFinal(Probe)
-        ├── output columns: [SUM(a) (#2), t1.b (#1)]
+    └── AggregatePartial
         ├── group by: [b]
         ├── aggregate functions: [sum(a)]
         ├── estimated rows: 0.00
-        └── AggregatePartial
-            ├── output columns: [SUM(a) (#2), #_group_by_key]
-            ├── group by: [b]
-            ├── aggregate functions: [sum(a)]
-            ├── estimated rows: 0.00
-            └── TableScan
-                ├── table: default.test_index_db.t1
-                ├── output columns: [a (#0), b (#1)]
-                ├── read rows: 0
-                ├── read bytes: 0
-                ├── partitions total: 0
-                ├── partitions scanned: 0
-                ├── push downs: [filters: [], limit: NONE]
-                └── estimated rows: 0.00
+        └── TableScan
+            ├── table: default.test_index_db.t1
+            ├── output columns: [a (#0), b (#1)]
+            ├── read rows: 0
+            ├── read bytes: 0
+            ├── partitions total: 0
+            ├── partitions scanned: 0
+            ├── push downs: [filters: [], limit: NONE]
+            └── estimated rows: 0.00
 
 # Disable aggregating index scan
 statement ok
@@ -509,7 +503,7 @@ Sort
             ├── estimated rows: 0.00
             └── Filter
                 ├── output columns: [onebrc.station_name (#0), onebrc.measurement (#1)]
-                ├── filters: [is_true(onebrc.station_name (#0) = 'Beijing'), is_true(onebrc.measurement (#1) > 0), is_true(onebrc.measurement (#1) = 1 OR onebrc.measurement (#1) = 2)]
+                ├── filters: [is_true(onebrc.measurement (#1) > 0), is_true(onebrc.station_name (#0) = 'Beijing'), is_true(onebrc.measurement (#1) = 1 OR onebrc.measurement (#1) = 2)]
                 ├── estimated rows: 0.00
                 └── TableScan
                     ├── table: default.test_index_db.onebrc
@@ -518,7 +512,7 @@ Sort
                     ├── read bytes: 0
                     ├── partitions total: 0
                     ├── partitions scanned: 0
-                    ├── push downs: [filters: [and_filters(and_filters(onebrc.station_name (#0) = 'Beijing', onebrc.measurement (#1) > 0), onebrc.measurement (#1) = 1 OR onebrc.measurement (#1) = 2)], limit: NONE]
+                    ├── push downs: [filters: [and_filters(and_filters(onebrc.measurement (#1) > 0, onebrc.station_name (#0) = 'Beijing'), onebrc.measurement (#1) = 1 OR onebrc.measurement (#1) = 2)], limit: NONE]
                     ├── aggregating index: [SELECT station_name, measurement, COUNT(), COUNT(measurement), MAX(measurement), MIN(measurement), SUM(measurement) FROM test_index_db.onebrc GROUP BY station_name, measurement]
                     ├── rewritten query: [selection: [index_col_0 (#0), index_col_1 (#1), index_col_5 (#5), index_col_6 (#6), index_col_3 (#3), index_col_4 (#4)], filter: is_true(CAST(onebrc.station_name (#0) AS Boolean NULL) AND CAST('Beijing' AS Boolean NULL) AND (onebrc.measurement (#1) = CAST(1 AS Float64 NULL) OR onebrc.measurement (#1) = CAST(2 AS Float64 NULL)) AND index_col_1 (#1) > CAST(0 AS Float64 NULL))]
                     └── estimated rows: 0.00
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/09_0039_target_build_merge_into_standalone.test b/tests/sqllogictests/suites/mode/standalone/explain/09_0039_target_build_merge_into_standalone.test
index e9cdadf2b9a9..b877d7537827 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/09_0039_target_build_merge_into_standalone.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/09_0039_target_build_merge_into_standalone.test
@@ -1,6 +1,12 @@
 statement ok
 set enable_experimental_merge_into = 1;
 
+statement ok
+drop table if exists target_build_optimization;
+
+statement ok
+drop table if exists source_optimization;
+
 ## Target Build Optimization Test
 statement ok
 create table target_build_optimization(a int,b string,c string);
@@ -150,7 +156,7 @@ select * from target_build_optimization order by a,b,c;
 11 b11 c_11
 12 b12 c_12
 
-### test with conjunct 
+### test with conjunct
 #### we need to make sure the blocks count and layout, so we should truncate and insert again.
 statement ok
 truncate table target_build_optimization;
@@ -303,4 +309,4 @@ select * from target_build_optimization order by a,b,c;
 12 b12 c_12
 
 statement ok
-set enable_experimental_merge_into = 0;
\ No newline at end of file
+set enable_experimental_merge_into = 0;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/bloom_filter.test b/tests/sqllogictests/suites/mode/standalone/explain/bloom_filter.test
index fdc6c353dec6..26ec22165615 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/bloom_filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/bloom_filter.test
@@ -1,4 +1,7 @@
 # This case depends on explain(standalone mode), thus we put it here
+statement ok
+drop table if exists bloom_test_t;
+
 statement ok
 create table bloom_test_t(c1 int, c2 int)
 
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/eliminate_outer_join.test b/tests/sqllogictests/suites/mode/standalone/explain/eliminate_outer_join.test
index d98e7ac8db2d..d2accfbc4f93 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/eliminate_outer_join.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/eliminate_outer_join.test
@@ -185,71 +185,79 @@ query T
 explain select * from t full join t t1 on t.a = t1.a where t.a is not null
 ----
 HashJoin
-├── output columns: [t1.a (#1), t.a (#0)]
-├── join type: RIGHT OUTER
-├── build keys: [t.a (#0)]
-├── probe keys: [t1.a (#1)]
+├── output columns: [t.a (#0), t1.a (#1)]
+├── join type: LEFT OUTER
+├── build keys: [t1.a (#1)]
+├── probe keys: [t.a (#0)]
 ├── filters: []
 ├── estimated rows: 2.00
 ├── Filter(Build)
-│   ├── output columns: [t.a (#0)]
-│   ├── filters: [is_not_null(t.a (#0))]
+│   ├── output columns: [t1.a (#1)]
+│   ├── filters: [is_not_null(t1.a (#1))]
 │   ├── estimated rows: 2.00
 │   └── TableScan
 │       ├── table: default.eliminate_outer_join.t
-│       ├── output columns: [a (#0)]
+│       ├── output columns: [a (#1)]
 │       ├── read rows: 10
 │       ├── read bytes: 74
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-│       ├── push downs: [filters: [is_not_null(t.a (#0))], limit: NONE]
+│       ├── push downs: [filters: [is_not_null(t.a (#1))], limit: NONE]
 │       └── estimated rows: 10.00
-└── TableScan(Probe)
-    ├── table: default.eliminate_outer_join.t
-    ├── output columns: [a (#1)]
-    ├── read rows: 10
-    ├── read bytes: 74
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+└── Filter(Probe)
+    ├── output columns: [t.a (#0)]
+    ├── filters: [is_not_null(t.a (#0))]
+    ├── estimated rows: 2.00
+    └── TableScan
+        ├── table: default.eliminate_outer_join.t
+        ├── output columns: [a (#0)]
+        ├── read rows: 10
+        ├── read bytes: 74
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_not_null(t.a (#0))], limit: NONE]
+        └── estimated rows: 10.00
 
 query T
 explain select * from t full join t t1 on t.a = t1.a where t1.a is not null
 ----
 HashJoin
-├── output columns: [t.a (#0), t1.a (#1)]
-├── join type: RIGHT OUTER
-├── build keys: [t1.a (#1)]
-├── probe keys: [t.a (#0)]
+├── output columns: [t1.a (#1), t.a (#0)]
+├── join type: LEFT OUTER
+├── build keys: [t.a (#0)]
+├── probe keys: [t1.a (#1)]
 ├── filters: []
 ├── estimated rows: 2.00
 ├── Filter(Build)
-│   ├── output columns: [t1.a (#1)]
-│   ├── filters: [is_not_null(t1.a (#1))]
+│   ├── output columns: [t.a (#0)]
+│   ├── filters: [is_not_null(t.a (#0))]
 │   ├── estimated rows: 2.00
 │   └── TableScan
 │       ├── table: default.eliminate_outer_join.t
-│       ├── output columns: [a (#1)]
+│       ├── output columns: [a (#0)]
 │       ├── read rows: 10
 │       ├── read bytes: 74
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-│       ├── push downs: [filters: [is_not_null(t.a (#1))], limit: NONE]
+│       ├── push downs: [filters: [is_not_null(t.a (#0))], limit: NONE]
 │       └── estimated rows: 10.00
-└── TableScan(Probe)
-    ├── table: default.eliminate_outer_join.t
-    ├── output columns: [a (#0)]
-    ├── read rows: 10
-    ├── read bytes: 74
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#1)]
+    ├── filters: [is_not_null(t1.a (#1))]
+    ├── estimated rows: 2.00
+    └── TableScan
+        ├── table: default.eliminate_outer_join.t
+        ├── output columns: [a (#1)]
+        ├── read rows: 10
+        ├── read bytes: 74
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_not_null(t.a (#1))], limit: NONE]
+        └── estimated rows: 10.00
 
 query T
 explain select * from t left join t t1 on t.a = t1.a where t1.a is not null and t.a is not null
@@ -606,37 +614,33 @@ Filter
 query T
 explain select * from t left join t t1 on t.a = t1.a where t1.a <= 1 or (t.a > 1 and t.a < 2)
 ----
-Filter
+HashJoin
 ├── output columns: [t.a (#0), t1.a (#1)]
-├── filters: [is_true(t1.a (#1) <= 1 OR t.a (#0) > 1 AND t.a (#0) < 2)]
-├── estimated rows: 4.05
-└── HashJoin
-    ├── output columns: [t.a (#0), t1.a (#1)]
-    ├── join type: LEFT OUTER
-    ├── build keys: [t1.a (#1)]
-    ├── probe keys: [t.a (#0)]
-    ├── filters: []
-    ├── estimated rows: 10.00
-    ├── TableScan(Build)
-    │   ├── table: default.eliminate_outer_join.t
-    │   ├── output columns: [a (#1)]
-    │   ├── read rows: 10
-    │   ├── read bytes: 74
-    │   ├── partitions total: 1
-    │   ├── partitions scanned: 1
-    │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    │   ├── push downs: [filters: [], limit: NONE]
-    │   └── estimated rows: 10.00
-    └── TableScan(Probe)
-        ├── table: default.eliminate_outer_join.t
-        ├── output columns: [a (#0)]
-        ├── read rows: 10
-        ├── read bytes: 74
-        ├── partitions total: 1
-        ├── partitions scanned: 1
-        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-        ├── push downs: [filters: [], limit: NONE]
-        └── estimated rows: 10.00
+├── join type: INNER
+├── build keys: [t1.a (#1)]
+├── probe keys: [t.a (#0)]
+├── filters: [t1.a (#1) <= 1 OR t.a (#0) > 1 AND t.a (#0) < 2]
+├── estimated rows: 10.00
+├── TableScan(Build)
+│   ├── table: default.eliminate_outer_join.t
+│   ├── output columns: [a (#1)]
+│   ├── read rows: 10
+│   ├── read bytes: 74
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 10.00
+└── TableScan(Probe)
+    ├── table: default.eliminate_outer_join.t
+    ├── output columns: [a (#0)]
+    ├── read rows: 10
+    ├── read bytes: 74
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 10.00
 
 statement ok
 create table time_table(a timestamp)
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/explain.test b/tests/sqllogictests/suites/mode/standalone/explain/explain.test
index f2f0f7b11ccf..8bbd4b423df2 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/explain.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/explain.test
@@ -703,54 +703,50 @@ create table t3 as select number as a, number as b from numbers(10)
 query T
 explain select * from t1,t2, t3 where (t1.a > 1 and t2.a > 2) or (t1.b < 3 and t2.b < 4) or t3.a = 2
 ----
-Filter
+HashJoin
 ├── output columns: [t3.a (#4), t3.b (#5), t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
+├── join type: INNER
+├── build keys: []
+├── probe keys: []
 ├── filters: [t1.a (#0) > 1 AND t2.a (#2) > 2 OR t1.b (#1) < 3 AND t2.b (#3) < 4 OR t3.a (#4) = 2]
-├── estimated rows: 21.20
-└── HashJoin
-    ├── output columns: [t3.a (#4), t3.b (#5), t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
-    ├── join type: CROSS
-    ├── build keys: []
-    ├── probe keys: []
-    ├── filters: []
-    ├── estimated rows: 50.00
-    ├── HashJoin(Build)
-    │   ├── output columns: [t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
-    │   ├── join type: CROSS
-    │   ├── build keys: []
-    │   ├── probe keys: []
-    │   ├── filters: []
-    │   ├── estimated rows: 5.00
-    │   ├── TableScan(Build)
-    │   │   ├── table: default.default.t1
-    │   │   ├── output columns: [a (#0), b (#1)]
-    │   │   ├── read rows: 1
-    │   │   ├── read bytes: 78
-    │   │   ├── partitions total: 1
-    │   │   ├── partitions scanned: 1
-    │   │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    │   │   ├── push downs: [filters: [], limit: NONE]
-    │   │   └── estimated rows: 1.00
-    │   └── TableScan(Probe)
-    │       ├── table: default.default.t2
-    │       ├── output columns: [a (#2), b (#3)]
-    │       ├── read rows: 5
-    │       ├── read bytes: 108
-    │       ├── partitions total: 1
-    │       ├── partitions scanned: 1
-    │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    │       ├── push downs: [filters: [], limit: NONE]
-    │       └── estimated rows: 5.00
-    └── TableScan(Probe)
-        ├── table: default.default.t3
-        ├── output columns: [a (#4), b (#5)]
-        ├── read rows: 10
-        ├── read bytes: 130
-        ├── partitions total: 1
-        ├── partitions scanned: 1
-        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-        ├── push downs: [filters: [], limit: NONE]
-        └── estimated rows: 10.00
+├── estimated rows: 50.00
+├── HashJoin(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
+│   ├── join type: CROSS
+│   ├── build keys: []
+│   ├── probe keys: []
+│   ├── filters: []
+│   ├── estimated rows: 5.00
+│   ├── TableScan(Build)
+│   │   ├── table: default.default.t1
+│   │   ├── output columns: [a (#0), b (#1)]
+│   │   ├── read rows: 1
+│   │   ├── read bytes: 78
+│   │   ├── partitions total: 1
+│   │   ├── partitions scanned: 1
+│   │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   │   ├── push downs: [filters: [], limit: NONE]
+│   │   └── estimated rows: 1.00
+│   └── TableScan(Probe)
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 5
+│       ├── read bytes: 108
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [], limit: NONE]
+│       └── estimated rows: 5.00
+└── TableScan(Probe)
+    ├── table: default.default.t3
+    ├── output columns: [a (#4), b (#5)]
+    ├── read rows: 10
+    ├── read bytes: 130
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 10.00
 
 query T
 explain select * from t1,t2, t3 where ((t1.a > 1 and t2.a > 2) or (t1.b < 3 and t2.b < 4)) and t3.a > 1
@@ -1132,7 +1128,7 @@ Limit
     ├── estimated rows: 0.08
     ├── Filter(Build)
     │   ├── output columns: [a.id (#0), a.c1 (#1)]
-    │   ├── filters: [is_true(CAST(a.c1 (#1) AS Int64 NULL) >= 1683648000), is_true(CAST(a.c1 (#1) AS Int64 NULL) <= 1683734400)]
+    │   ├── filters: [is_true(a.c1 (#1) >= 1683648000), is_true(a.c1 (#1) <= 1683734400)]
     │   ├── estimated rows: 0.20
     │   └── TableScan
     │       ├── table: default.default.a
@@ -1142,11 +1138,11 @@ Limit
     │       ├── partitions total: 1
     │       ├── partitions scanned: 1
     │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    │       ├── push downs: [filters: [and_filters(CAST(a.c1 (#1) AS Int64 NULL) >= 1683648000, CAST(a.c1 (#1) AS Int64 NULL) <= 1683734400)], limit: NONE]
+    │       ├── push downs: [filters: [and_filters(a.c1 (#1) >= 1683648000, a.c1 (#1) <= 1683734400)], limit: NONE]
     │       └── estimated rows: 1.00
     └── Filter(Probe)
         ├── output columns: [b.id (#2), b.c1 (#3)]
-        ├── filters: [is_true(CAST(b.c1 (#3) AS Int64 NULL) >= 1683648000), is_true(CAST(b.c1 (#3) AS Int64 NULL) <= 1683734400)]
+        ├── filters: [is_true(b.c1 (#3) >= 1683648000), is_true(b.c1 (#3) <= 1683734400)]
         ├── estimated rows: 0.40
         └── TableScan
             ├── table: default.default.b
@@ -1156,7 +1152,7 @@ Limit
             ├── partitions total: 2
             ├── partitions scanned: 2
             ├── pruning stats: [segments: <range pruning: 2 to 2>, blocks: <range pruning: 2 to 2, bloom pruning: 0 to 0>]
-            ├── push downs: [filters: [and_filters(CAST(b.c1 (#3) AS Int64 NULL) >= 1683648000, CAST(b.c1 (#3) AS Int64 NULL) <= 1683734400)], limit: NONE]
+            ├── push downs: [filters: [and_filters(b.c1 (#3) >= 1683648000, b.c1 (#3) <= 1683734400)], limit: NONE]
             └── estimated rows: 2.00
 
 statement ok
@@ -1173,49 +1169,49 @@ from numbers(10) where number > 5
 ) b on  a.number=b.number order by a.number) where pt = register_at;
 ----
 Sort
-├── output columns: [register_at (#3), numbers.number (#0), pt (#1)]
+├── output columns: [numbers.number (#0), pt (#1), register_at (#3)]
 ├── sort keys: [number ASC NULLS LAST]
 ├── estimated rows: 0.00
 └── HashJoin
-    ├── output columns: [register_at (#3), numbers.number (#0), pt (#1)]
+    ├── output columns: [numbers.number (#0), pt (#1), register_at (#3)]
     ├── join type: INNER
-    ├── build keys: [a.number (#0), a.pt (#1)]
-    ├── probe keys: [b.number (#2), b.register_at (#3)]
+    ├── build keys: [b.register_at (#3), b.number (#2)]
+    ├── probe keys: [a.pt (#1), a.number (#0)]
     ├── filters: []
     ├── estimated rows: 0.00
     ├── EvalScalar(Build)
-    │   ├── output columns: [numbers.number (#0), pt (#1)]
-    │   ├── expressions: [to_yyyymmdd(to_timestamp(to_int64(numbers.number (#0))))]
+    │   ├── output columns: [numbers.number (#2), register_at (#3)]
+    │   ├── expressions: [to_yyyymmdd(to_timestamp(to_int64(numbers.number (#2))))]
     │   ├── estimated rows: 0.00
     │   └── Filter
-    │       ├── output columns: [numbers.number (#0)]
-    │       ├── filters: [numbers.number (#0) > 5]
+    │       ├── output columns: [numbers.number (#2)]
+    │       ├── filters: [numbers.number (#2) > 5]
     │       ├── estimated rows: 0.00
     │       └── TableScan
     │           ├── table: default.system.numbers
-    │           ├── output columns: [number (#0)]
+    │           ├── output columns: [number (#2)]
     │           ├── read rows: 10
     │           ├── read bytes: 80
     │           ├── partitions total: 1
     │           ├── partitions scanned: 1
-    │           ├── push downs: [filters: [numbers.number (#0) > 5], limit: NONE]
+    │           ├── push downs: [filters: [numbers.number (#2) > 5], limit: NONE]
     │           └── estimated rows: 10.00
     └── EvalScalar(Probe)
-        ├── output columns: [numbers.number (#2), register_at (#3)]
-        ├── expressions: [to_yyyymmdd(to_timestamp(to_int64(numbers.number (#2))))]
+        ├── output columns: [numbers.number (#0), pt (#1)]
+        ├── expressions: [to_yyyymmdd(to_timestamp(to_int64(numbers.number (#0))))]
         ├── estimated rows: 0.00
         └── Filter
-            ├── output columns: [numbers.number (#2)]
-            ├── filters: [numbers.number (#2) > 5]
+            ├── output columns: [numbers.number (#0)]
+            ├── filters: [numbers.number (#0) > 5]
             ├── estimated rows: 0.00
             └── TableScan
                 ├── table: default.system.numbers
-                ├── output columns: [number (#2)]
+                ├── output columns: [number (#0)]
                 ├── read rows: 10
                 ├── read bytes: 80
                 ├── partitions total: 1
                 ├── partitions scanned: 1
-                ├── push downs: [filters: [numbers.number (#2) > 5], limit: NONE]
+                ├── push downs: [filters: [numbers.number (#0) > 5], limit: NONE]
                 └── estimated rows: 10.00
 
 
@@ -1244,17 +1240,17 @@ explain select * from a where a.id = (select id from b where a.id = b.id);
 Filter
 ├── output columns: [a.id (#0), a.c1 (#1)]
 ├── filters: [is_true(a.id (#0) = scalar_subquery_2 (#2))]
-├── estimated rows: 0.60
+├── estimated rows: 0.08
 └── HashJoin
     ├── output columns: [a.id (#0), a.c1 (#1), b.id (#2)]
-    ├── join type: LEFT SINGLE
+    ├── join type: INNER
     ├── build keys: [id (#2)]
     ├── probe keys: [id (#0)]
     ├── filters: []
-    ├── estimated rows: 3.00
+    ├── estimated rows: 0.40
     ├── Filter(Build)
     │   ├── output columns: [b.id (#2)]
-    │   ├── filters: [is_true(id (#2) = b.id (#2))]
+    │   ├── filters: [is_true(b.id (#2) = b.id (#2))]
     │   ├── estimated rows: 0.40
     │   └── TableScan
     │       ├── table: default.default.b
@@ -1332,34 +1328,34 @@ Filter
 ├── estimated rows: 0.00
 └── HashJoin
     ├── output columns: [t1.a (#0), t1.b (#1), marker (#3)]
-    ├── join type: RIGHT MARK
-    ├── build keys: [CAST(subquery_2 (#2) AS Int32 NULL)]
-    ├── probe keys: [t1.a (#0)]
+    ├── join type: LEFT MARK
+    ├── build keys: [t1.a (#0)]
+    ├── probe keys: [CAST(subquery_2 (#2) AS Int32 NULL)]
     ├── filters: []
     ├── estimated rows: 3.00
-    ├── AggregateFinal(Build)
-    │   ├── output columns: [col0 (#2)]
-    │   ├── group by: [col0]
-    │   ├── aggregate functions: []
-    │   ├── estimated rows: 1301.00
-    │   └── AggregatePartial
-    │       ├── output columns: [#_group_by_key]
-    │       ├── group by: [col0]
-    │       ├── aggregate functions: []
-    │       ├── estimated rows: 1301.00
-    │       └── ConstantTableScan
-    │           ├── output columns: [col0 (#2)]
-    │           └── column 0: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 360, 361, 362, 363, 364, 365, 366, 367, 368, 369, 370, 371, 372, 373, 374, 375, 376, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415, 416, 417, 418, 419, 420, 421, 422, 423, 424, 425, 426, 427, 428, 429, 430, 431, 432, 433, 434, 435, 436, 437, 438, 439, 440, 441, 442, 443, 444, 445, 446, 447, 448, 449, 450, 451, 452, 453, 454, 455, 456, 457, 458, 459, 460, 461, 462, 463, 464, 465, 466, 467, 468, 469, 470, 471, 472, 473, 474, 475, 476, 477, 478, 479, 480, 481, 482, 483, 484, 485, 486, 487, 488, 489, 490, 491, 492, 493, 494, 495, 496, 497, 498, 499, 500, 501, 502, 503, 504, 505, 506, 507, 508, 509, 510, 511, 512, 513, 514, 515, 516, 517, 518, 519, 520, 521, 522, 523, 524, 525, 526, 527, 528, 529, 530, 531, 532, 533, 534, 535, 536, 537, 538, 539, 540, 541, 542, 543, 544, 545, 546, 547, 548, 549, 550, 551, 552, 553, 554, 555, 556, 557, 558, 559, 560, 561, 562, 563, 564, 565, 566, 567, 568, 569, 570, 571, 572, 573, 574, 575, 576, 577, 578, 579, 580, 581, 582, 583, 584, 585, 586, 587, 588, 589, 590, 591, 592, 593, 594, 595, 596, 597, 598, 599, 600, 601, 602, 603, 604, 605, 606, 607, 608, 609, 610, 611, 612, 613, 614, 615, 616, 617, 618, 619, 620, 621, 622, 623, 624, 625, 626, 627, 628, 629, 630, 631, 632, 633, 634, 635, 636, 637, 638, 639, 640, 641, 642, 643, 644, 645, 646, 647, 648, 649, 650, 651, 652, 653, 654, 655, 656, 657, 658, 659, 660, 661, 662, 663, 664, 665, 666, 667, 668, 669, 670, 671, 672, 673, 674, 675, 676, 677, 678, 679, 680, 681, 682, 683, 684, 685, 686, 687, 688, 689, 690, 691, 692, 693, 694, 695, 696, 697, 698, 699, 700, 701, 702, 703, 704, 705, 706, 707, 708, 709, 710, 711, 712, 713, 714, 715, 716, 717, 718, 719, 720, 721, 722, 723, 724, 725, 726, 727, 728, 729, 730, 731, 732, 733, 734, 735, 736, 737, 738, 739, 740, 741, 742, 743, 744, 745, 746, 747, 748, 749, 750, 751, 752, 753, 754, 755, 756, 757, 758, 759, 760, 761, 762, 763, 764, 765, 766, 767, 768, 769, 770, 771, 772, 773, 774, 775, 776, 777, 778, 779, 780, 781, 782, 783, 784, 785, 786, 787, 788, 789, 790, 791, 792, 793, 794, 795, 796, 797, 798, 799, 800, 801, 802, 803, 804, 805, 806, 807, 808, 809, 810, 811, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 825, 826, 827, 828, 829, 830, 831, 832, 833, 834, 835, 836, 837, 838, 839, 840, 841, 842, 843, 844, 845, 846, 847, 848, 849, 850, 851, 852, 853, 854, 855, 856, 857, 858, 859, 860, 861, 862, 863, 864, 865, 866, 867, 868, 869, 870, 871, 872, 873, 874, 875, 876, 877, 878, 879, 880, 881, 882, 883, 884, 885, 886, 887, 888, 889, 890, 891, 892, 893, 894, 895, 896, 897, 898, 899, 900, 901, 902, 903, 904, 905, 906, 907, 908, 909, 910, 911, 912, 913, 914, 915, 916, 917, 918, 919, 920, 921, 922, 923, 924, 925, 926, 927, 928, 929, 930, 931, 932, 933, 934, 935, 936, 937, 938, 939, 940, 941, 942, 943, 944, 945, 946, 947, 948, 949, 950, 951, 952, 953, 954, 955, 956, 957, 958, 959, 960, 961, 962, 963, 964, 965, 966, 967, 968, 969, 970, 971, 972, 973, 974, 975, 976, 977, 978, 979, 980, 981, 982, 983, 984, 985, 986, 987, 988, 989, 990, 991, 992, 993, 994, 995, 996, 997, 998, 999, 1000, 1001, 1002, 1003, 1004, 1005, 1006, 1007, 1008, 1009, 1010, 1011, 1012, 1013, 1014, 1015, 1016, 1017, 1018, 1019, 1020, 1021, 1022, 1023, 1024, 1025, 1026, 1027, 1028, 1029, 1030, 1031, 1032, 1033, 1034, 1035, 1036, 1037, 1038, 1039, 1040, 1041, 1042, 1043, 1044, 1045, 1046, 1047, 1048, 1049, 1050, 1051, 1052, 1053, 1054, 1055, 1056, 1057, 1058, 1059, 1060, 1061, 1062, 1063, 1064, 1065, 1066, 1067, 1068, 1069, 1070, 1071, 1072, 1073, 1074, 1075, 1076, 1077, 1078, 1079, 1080, 1081, 1082, 1083, 1084, 1085, 1086, 1087, 1088, 1089, 1090, 1091, 1092, 1093, 1094, 1095, 1096, 1097, 1098, 1099, 1100, 1101, 1102, 1103, 1104, 1105, 1106, 1107, 1108, 1109, 1110, 1111, 1112, 1113, 1114, 1115, 1116, 1117, 1118, 1119, 1120, 1121, 1122, 1123, 1124, 1125, 1126, 1127, 1128, 1129, 1130, 1131, 1132, 1133, 1134, 1135, 1136, 1137, 1138, 1139, 1140, 1141, 1142, 1143, 1144, 1145, 1146, 1147, 1148, 1149, 1150, 1151, 1152, 1153, 1154, 1155, 1156, 1157, 1158, 1159, 1160, 1161, 1162, 1163, 1164, 1165, 1166, 1167, 1168, 1169, 1170, 1171, 1172, 1173, 1174, 1175, 1176, 1177, 1178, 1179, 1180, 1181, 1182, 1183, 1184, 1185, 1186, 1187, 1188, 1189, 1190, 1191, 1192, 1193, 1194, 1195, 1196, 1197, 1198, 1199, 1200, 1201, 1202, 1203, 1204, 1205, 1206, 1207, 1208, 1209, 1210, 1211, 1212, 1213, 1214, 1215, 1216, 1217, 1218, 1219, 1220, 1221, 1222, 1223, 1224, 1225, 1226, 1227, 1228, 1229, 1230, 1231, 1232, 1233, 1234, 1235, 1236, 1237, 1238, 1239, 1240, 1241, 1242, 1243, 1244, 1245, 1246, 1247, 1248, 1249, 1250, 1251, 1252, 1253, 1254, 1255, 1256, 1257, 1258, 1259, 1260, 1261, 1262, 1263, 1264, 1265, 1266, 1267, 1268, 1269, 1270, 1271, 1272, 1273, 1274, 1275, 1276, 1277, 1278, 1279, 1280, 1281, 1282, 1283, 1284, 1285, 1286, 1287, 1288, 1289, 1290, 1291, 1292, 1293, 1294, 1295, 1296, 1297, 1298, 1299, 1300]
-    └── TableScan(Probe)
-        ├── table: default.default.t1
-        ├── output columns: [a (#0), b (#1)]
-        ├── read rows: 3
-        ├── read bytes: 90
-        ├── partitions total: 1
-        ├── partitions scanned: 1
-        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-        ├── push downs: [filters: [], limit: NONE]
-        └── estimated rows: 3.00
+    ├── TableScan(Build)
+    │   ├── table: default.default.t1
+    │   ├── output columns: [a (#0), b (#1)]
+    │   ├── read rows: 3
+    │   ├── read bytes: 90
+    │   ├── partitions total: 1
+    │   ├── partitions scanned: 1
+    │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    │   ├── push downs: [filters: [], limit: NONE]
+    │   └── estimated rows: 3.00
+    └── AggregateFinal(Probe)
+        ├── output columns: [col0 (#2)]
+        ├── group by: [col0]
+        ├── aggregate functions: []
+        ├── estimated rows: 1301.00
+        └── AggregatePartial
+            ├── output columns: [#_group_by_key]
+            ├── group by: [col0]
+            ├── aggregate functions: []
+            ├── estimated rows: 1301.00
+            └── ConstantTableScan
+                ├── output columns: [col0 (#2)]
+                └── column 0: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 245, 246, 247, 248, 249, 250, 251, 252, 253, 254, 255, 256, 257, 258, 259, 260, 261, 262, 263, 264, 265, 266, 267, 268, 269, 270, 271, 272, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 283, 284, 285, 286, 287, 288, 289, 290, 291, 292, 293, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309, 310, 311, 312, 313, 314, 315, 316, 317, 318, 319, 320, 321, 322, 323, 324, 325, 326, 327, 328, 329, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 350, 351, 352, 353, 354, 355, 356, 357, 358, 359, 360, 361, 362, 363, 364, 365, 366, 367, 368, 369, 370, 371, 372, 373, 374, 375, 376, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400, 401, 402, 403, 404, 405, 406, 407, 408, 409, 410, 411, 412, 413, 414, 415, 416, 417, 418, 419, 420, 421, 422, 423, 424, 425, 426, 427, 428, 429, 430, 431, 432, 433, 434, 435, 436, 437, 438, 439, 440, 441, 442, 443, 444, 445, 446, 447, 448, 449, 450, 451, 452, 453, 454, 455, 456, 457, 458, 459, 460, 461, 462, 463, 464, 465, 466, 467, 468, 469, 470, 471, 472, 473, 474, 475, 476, 477, 478, 479, 480, 481, 482, 483, 484, 485, 486, 487, 488, 489, 490, 491, 492, 493, 494, 495, 496, 497, 498, 499, 500, 501, 502, 503, 504, 505, 506, 507, 508, 509, 510, 511, 512, 513, 514, 515, 516, 517, 518, 519, 520, 521, 522, 523, 524, 525, 526, 527, 528, 529, 530, 531, 532, 533, 534, 535, 536, 537, 538, 539, 540, 541, 542, 543, 544, 545, 546, 547, 548, 549, 550, 551, 552, 553, 554, 555, 556, 557, 558, 559, 560, 561, 562, 563, 564, 565, 566, 567, 568, 569, 570, 571, 572, 573, 574, 575, 576, 577, 578, 579, 580, 581, 582, 583, 584, 585, 586, 587, 588, 589, 590, 591, 592, 593, 594, 595, 596, 597, 598, 599, 600, 601, 602, 603, 604, 605, 606, 607, 608, 609, 610, 611, 612, 613, 614, 615, 616, 617, 618, 619, 620, 621, 622, 623, 624, 625, 626, 627, 628, 629, 630, 631, 632, 633, 634, 635, 636, 637, 638, 639, 640, 641, 642, 643, 644, 645, 646, 647, 648, 649, 650, 651, 652, 653, 654, 655, 656, 657, 658, 659, 660, 661, 662, 663, 664, 665, 666, 667, 668, 669, 670, 671, 672, 673, 674, 675, 676, 677, 678, 679, 680, 681, 682, 683, 684, 685, 686, 687, 688, 689, 690, 691, 692, 693, 694, 695, 696, 697, 698, 699, 700, 701, 702, 703, 704, 705, 706, 707, 708, 709, 710, 711, 712, 713, 714, 715, 716, 717, 718, 719, 720, 721, 722, 723, 724, 725, 726, 727, 728, 729, 730, 731, 732, 733, 734, 735, 736, 737, 738, 739, 740, 741, 742, 743, 744, 745, 746, 747, 748, 749, 750, 751, 752, 753, 754, 755, 756, 757, 758, 759, 760, 761, 762, 763, 764, 765, 766, 767, 768, 769, 770, 771, 772, 773, 774, 775, 776, 777, 778, 779, 780, 781, 782, 783, 784, 785, 786, 787, 788, 789, 790, 791, 792, 793, 794, 795, 796, 797, 798, 799, 800, 801, 802, 803, 804, 805, 806, 807, 808, 809, 810, 811, 812, 813, 814, 815, 816, 817, 818, 819, 820, 821, 822, 823, 824, 825, 826, 827, 828, 829, 830, 831, 832, 833, 834, 835, 836, 837, 838, 839, 840, 841, 842, 843, 844, 845, 846, 847, 848, 849, 850, 851, 852, 853, 854, 855, 856, 857, 858, 859, 860, 861, 862, 863, 864, 865, 866, 867, 868, 869, 870, 871, 872, 873, 874, 875, 876, 877, 878, 879, 880, 881, 882, 883, 884, 885, 886, 887, 888, 889, 890, 891, 892, 893, 894, 895, 896, 897, 898, 899, 900, 901, 902, 903, 904, 905, 906, 907, 908, 909, 910, 911, 912, 913, 914, 915, 916, 917, 918, 919, 920, 921, 922, 923, 924, 925, 926, 927, 928, 929, 930, 931, 932, 933, 934, 935, 936, 937, 938, 939, 940, 941, 942, 943, 944, 945, 946, 947, 948, 949, 950, 951, 952, 953, 954, 955, 956, 957, 958, 959, 960, 961, 962, 963, 964, 965, 966, 967, 968, 969, 970, 971, 972, 973, 974, 975, 976, 977, 978, 979, 980, 981, 982, 983, 984, 985, 986, 987, 988, 989, 990, 991, 992, 993, 994, 995, 996, 997, 998, 999, 1000, 1001, 1002, 1003, 1004, 1005, 1006, 1007, 1008, 1009, 1010, 1011, 1012, 1013, 1014, 1015, 1016, 1017, 1018, 1019, 1020, 1021, 1022, 1023, 1024, 1025, 1026, 1027, 1028, 1029, 1030, 1031, 1032, 1033, 1034, 1035, 1036, 1037, 1038, 1039, 1040, 1041, 1042, 1043, 1044, 1045, 1046, 1047, 1048, 1049, 1050, 1051, 1052, 1053, 1054, 1055, 1056, 1057, 1058, 1059, 1060, 1061, 1062, 1063, 1064, 1065, 1066, 1067, 1068, 1069, 1070, 1071, 1072, 1073, 1074, 1075, 1076, 1077, 1078, 1079, 1080, 1081, 1082, 1083, 1084, 1085, 1086, 1087, 1088, 1089, 1090, 1091, 1092, 1093, 1094, 1095, 1096, 1097, 1098, 1099, 1100, 1101, 1102, 1103, 1104, 1105, 1106, 1107, 1108, 1109, 1110, 1111, 1112, 1113, 1114, 1115, 1116, 1117, 1118, 1119, 1120, 1121, 1122, 1123, 1124, 1125, 1126, 1127, 1128, 1129, 1130, 1131, 1132, 1133, 1134, 1135, 1136, 1137, 1138, 1139, 1140, 1141, 1142, 1143, 1144, 1145, 1146, 1147, 1148, 1149, 1150, 1151, 1152, 1153, 1154, 1155, 1156, 1157, 1158, 1159, 1160, 1161, 1162, 1163, 1164, 1165, 1166, 1167, 1168, 1169, 1170, 1171, 1172, 1173, 1174, 1175, 1176, 1177, 1178, 1179, 1180, 1181, 1182, 1183, 1184, 1185, 1186, 1187, 1188, 1189, 1190, 1191, 1192, 1193, 1194, 1195, 1196, 1197, 1198, 1199, 1200, 1201, 1202, 1203, 1204, 1205, 1206, 1207, 1208, 1209, 1210, 1211, 1212, 1213, 1214, 1215, 1216, 1217, 1218, 1219, 1220, 1221, 1222, 1223, 1224, 1225, 1226, 1227, 1228, 1229, 1230, 1231, 1232, 1233, 1234, 1235, 1236, 1237, 1238, 1239, 1240, 1241, 1242, 1243, 1244, 1245, 1246, 1247, 1248, 1249, 1250, 1251, 1252, 1253, 1254, 1255, 1256, 1257, 1258, 1259, 1260, 1261, 1262, 1263, 1264, 1265, 1266, 1267, 1268, 1269, 1270, 1271, 1272, 1273, 1274, 1275, 1276, 1277, 1278, 1279, 1280, 1281, 1282, 1283, 1284, 1285, 1286, 1287, 1288, 1289, 1290, 1291, 1292, 1293, 1294, 1295, 1296, 1297, 1298, 1299, 1300]
 
 statement ok
 drop table t1;
@@ -1489,23 +1485,23 @@ FROM numbers(500);
 query T
 explain join SELECT customer_name, segment, (SELECT SUM(net_paid) FROM sales WHERE customer_id IN (SELECT customer_id FROM customers WHERE segment = c.segment AND active = true)) FROM customers c WHERE c.customer_id IN (SELECT customer_id FROM sales WHERE net_paid > 100) LIMIT 10;
 ----
-HashJoin: LEFT SINGLE
+HashJoin: LEFT OUTER
 ├── Build
 │   └── HashJoin: RIGHT SEMI
 │       ├── Build
 │       │   └── HashJoin: CROSS
 │       │       ├── Build
-│       │       │   └── Scan: default.default.customers (read rows: 100)
+│       │       │   └── Scan: default.default.customers (#0) (read rows: 100)
 │       │       └── Probe
-│       │           └── Scan: default.default.sales (read rows: 500)
+│       │           └── Scan: default.default.sales (#1) (read rows: 500)
 │       └── Probe
-│           └── Scan: default.default.customers (read rows: 100)
+│           └── Scan: default.default.customers (#2) (read rows: 100)
 └── Probe
     └── HashJoin: LEFT SEMI
         ├── Build
-        │   └── Scan: default.default.sales (read rows: 0)
+        │   └── Scan: default.default.sales (#3) (read rows: 0)
         └── Probe
-            └── Scan: default.default.customers (read rows: 100)
+            └── Scan: default.default.customers (#0) (read rows: 100)
 
 
 query T
@@ -1513,21 +1509,21 @@ explain join SELECT c.customer_name FROM customers c WHERE NOT EXISTS ( SELECT c
 ----
 HashJoin: RIGHT MARK
 ├── Build
-│   └── HashJoin: RIGHT MARK
+│   └── HashJoin: LEFT MARK
 │       ├── Build
-│       │   └── HashJoin: INNER
+│       │   └── HashJoin: CROSS
 │       │       ├── Build
-│       │       │   └── Scan: default.default.products (read rows: 10)
+│       │       │   └── Scan: default.default.customers (#0) (read rows: 100)
 │       │       └── Probe
-│       │           └── Scan: default.default.sales (read rows: 500)
+│       │           └── Scan: default.default.products (#1) (read rows: 10)
 │       └── Probe
-│           └── HashJoin: CROSS
+│           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.default.customers (read rows: 100)
+│               │   └── Scan: default.default.products (#3) (read rows: 10)
 │               └── Probe
-│                   └── Scan: default.default.products (read rows: 10)
+│                   └── Scan: default.default.sales (#2) (read rows: 500)
 └── Probe
-    └── Scan: default.default.customers (read rows: 100)
+    └── Scan: default.default.customers (#0) (read rows: 100)
 
 statement ok
 drop table customers;
@@ -1608,7 +1604,7 @@ Filter
 ├── estimated rows: 0.00
 └── HashJoin
     ├── output columns: [t2.a (#0), t2.b (#1), t2.c (#2), t1.c (#5)]
-    ├── join type: LEFT SINGLE
+    ├── join type: INNER
     ├── build keys: [a (#3)]
     ├── probe keys: [a (#0)]
     ├── filters: []
@@ -1633,7 +1629,7 @@ Filter
     │               ├── estimated rows: 0.00
     │               └── Filter
     │                   ├── output columns: [t1.a (#3), t1.c (#5)]
-    │                   ├── filters: [is_true(t1.a (#3) = a (#3))]
+    │                   ├── filters: [is_true(t1.a (#3) = t1.a (#3))]
     │                   ├── estimated rows: 0.00
     │                   └── TableScan
     │                       ├── table: default.default.t1
@@ -1663,7 +1659,7 @@ Filter
 ├── estimated rows: 0.00
 └── HashJoin
     ├── output columns: [t2.a (#0), t2.b (#1), t2.c (#2), t1.c (#5)]
-    ├── join type: LEFT SINGLE
+    ├── join type: INNER
     ├── build keys: [a (#3)]
     ├── probe keys: [a (#0)]
     ├── filters: []
@@ -1688,7 +1684,7 @@ Filter
     │               ├── estimated rows: 0.00
     │               └── Filter
     │                   ├── output columns: [t1.a (#3), t1.c (#5)]
-    │                   ├── filters: [is_true(t1.a (#3) = a (#3))]
+    │                   ├── filters: [is_true(a (#3) = a (#3))]
     │                   ├── estimated rows: 0.00
     │                   └── TableScan
     │                       ├── table: default.default.t1
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/filter.test b/tests/sqllogictests/suites/mode/standalone/explain/filter.test
index f4222c31218c..1a1009ba94a7 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/filter.test
@@ -79,7 +79,7 @@ HashJoin
 │   ├── estimated rows: 0.00
 │   └── Filter
 │       ├── output columns: [t.a (#0)]
-│       ├── filters: [false, is_true(t.a (#0) > 1)]
+│       ├── filters: [is_true(t.a (#0) > 1), false]
 │       ├── estimated rows: 0.00
 │       └── TableScan
 │           ├── table: default.default.t
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/infer_filter.test b/tests/sqllogictests/suites/mode/standalone/explain/infer_filter.test
index 3e3a1cf71bf2..4ee8635e7c10 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/infer_filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/infer_filter.test
@@ -581,7 +581,7 @@ explain select * from t1 where 10000 > a and a > 10 and 100 >= a;
 ----
 Filter
 ├── output columns: [t1.a (#0), t1.b (#1)]
-├── filters: [t1.a (#0) > 10, t1.a (#0) <= 100]
+├── filters: [t1.a (#0) <= 100, t1.a (#0) > 10]
 ├── estimated rows: 0.00
 └── TableScan
     ├── table: default.default.t1
@@ -590,7 +590,7 @@ Filter
     ├── read bytes: 0
     ├── partitions total: 0
     ├── partitions scanned: 0
-    ├── push downs: [filters: [and_filters(t1.a (#0) > 10, t1.a (#0) <= 100)], limit: NONE]
+    ├── push downs: [filters: [and_filters(t1.a (#0) <= 100, t1.a (#0) > 10)], limit: NONE]
     └── estimated rows: 0.00
 
 # t1.a > 10 and t2.a > 10
@@ -636,58 +636,58 @@ query T
 explain select * from t1, t2, t3 where t1.a = t2.a and t1.a = t3.a and t1.a > 5 and t3.a < 10;
 ----
 HashJoin
-├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2), t3.b (#5), t3.a (#4)]
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
 ├── join type: INNER
 ├── build keys: [t3.a (#4)]
 ├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 0.00
-├── Filter(Build)
-│   ├── output columns: [t3.a (#4), t3.b (#5)]
-│   ├── filters: [t3.a (#4) > 5, t3.a (#4) < 10]
+├── HashJoin(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
+│   ├── join type: INNER
+│   ├── build keys: [t3.a (#4)]
+│   ├── probe keys: [t2.a (#2)]
+│   ├── filters: []
 │   ├── estimated rows: 0.00
-│   └── TableScan
-│       ├── table: default.default.t3
-│       ├── output columns: [a (#4), b (#5)]
-│       ├── read rows: 0
-│       ├── read bytes: 0
-│       ├── partitions total: 0
-│       ├── partitions scanned: 0
-│       ├── push downs: [filters: [and_filters(t3.a (#4) > 5, t3.a (#4) < 10)], limit: NONE]
-│       └── estimated rows: 0.00
-└── HashJoin(Probe)
-    ├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
-    ├── join type: INNER
-    ├── build keys: [t2.a (#2)]
-    ├── probe keys: [t1.a (#0)]
-    ├── filters: []
+│   ├── Filter(Build)
+│   │   ├── output columns: [t3.a (#4), t3.b (#5)]
+│   │   ├── filters: [t3.a (#4) > 5, t3.a (#4) < 10]
+│   │   ├── estimated rows: 0.00
+│   │   └── TableScan
+│   │       ├── table: default.default.t3
+│   │       ├── output columns: [a (#4), b (#5)]
+│   │       ├── read rows: 0
+│   │       ├── read bytes: 0
+│   │       ├── partitions total: 0
+│   │       ├── partitions scanned: 0
+│   │       ├── push downs: [filters: [and_filters(t3.a (#4) > 5, t3.a (#4) < 10)], limit: NONE]
+│   │       └── estimated rows: 0.00
+│   └── Filter(Probe)
+│       ├── output columns: [t2.a (#2), t2.b (#3)]
+│       ├── filters: [t2.a (#2) > 5, t2.a (#2) < 10]
+│       ├── estimated rows: 0.00
+│       └── TableScan
+│           ├── table: default.default.t2
+│           ├── output columns: [a (#2), b (#3)]
+│           ├── read rows: 0
+│           ├── read bytes: 0
+│           ├── partitions total: 0
+│           ├── partitions scanned: 0
+│           ├── push downs: [filters: [and_filters(t2.a (#2) > 5, t2.a (#2) < 10)], limit: NONE]
+│           └── estimated rows: 0.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [t1.a (#0) > 5, t1.a (#0) < 10]
     ├── estimated rows: 0.00
-    ├── Filter(Build)
-    │   ├── output columns: [t2.a (#2), t2.b (#3)]
-    │   ├── filters: [t2.a (#2) > 5, t2.a (#2) < 10]
-    │   ├── estimated rows: 0.00
-    │   └── TableScan
-    │       ├── table: default.default.t2
-    │       ├── output columns: [a (#2), b (#3)]
-    │       ├── read rows: 0
-    │       ├── read bytes: 0
-    │       ├── partitions total: 0
-    │       ├── partitions scanned: 0
-    │       ├── push downs: [filters: [and_filters(t2.a (#2) > 5, t2.a (#2) < 10)], limit: NONE]
-    │       └── estimated rows: 0.00
-    └── Filter(Probe)
-        ├── output columns: [t1.a (#0), t1.b (#1)]
-        ├── filters: [t1.a (#0) > 5, t1.a (#0) < 10]
-        ├── estimated rows: 0.00
-        └── TableScan
-            ├── table: default.default.t1
-            ├── output columns: [a (#0), b (#1)]
-            ├── read rows: 0
-            ├── read bytes: 0
-            ├── partitions total: 0
-            ├── partitions scanned: 0
-            ├── push downs: [filters: [and_filters(t1.a (#0) > 5, t1.a (#0) < 10)], limit: NONE]
-            └── estimated rows: 0.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 0
+        ├── read bytes: 0
+        ├── partitions total: 0
+        ├── partitions scanned: 0
+        ├── push downs: [filters: [and_filters(t1.a (#0) > 5, t1.a (#0) < 10)], limit: NONE]
+        └── estimated rows: 0.00
 
 # t1.a > 5 and t2.a > 10
 query T
@@ -855,6 +855,53 @@ HashJoin
             ├── push downs: [filters: [false], limit: NONE]
             └── estimated rows: 0.00
 
+# t1.a = t2.a, t1.a = t3.a => t2.a = t3.a
+query T
+explain select * from t1, t2, t3 where t1.a = t2.a and t1.a = t3.a;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
+├── join type: INNER
+├── build keys: [t3.a (#4)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.00
+├── HashJoin(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
+│   ├── join type: INNER
+│   ├── build keys: [t3.a (#4)]
+│   ├── probe keys: [t2.a (#2)]
+│   ├── filters: []
+│   ├── estimated rows: 0.00
+│   ├── TableScan(Build)
+│   │   ├── table: default.default.t3
+│   │   ├── output columns: [a (#4), b (#5)]
+│   │   ├── read rows: 0
+│   │   ├── read bytes: 0
+│   │   ├── partitions total: 0
+│   │   ├── partitions scanned: 0
+│   │   ├── push downs: [filters: [], limit: NONE]
+│   │   └── estimated rows: 0.00
+│   └── TableScan(Probe)
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 0.00
+
+
 statement ok
 drop table if exists t1;
 
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/join.test b/tests/sqllogictests/suites/mode/standalone/explain/join.test
index 596d2be32efc..4173c36e02bb 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/join.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/join.test
@@ -67,16 +67,20 @@ HashJoin
 │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │   ├── push downs: [filters: [], limit: NONE]
 │   └── estimated rows: 1.00
-└── TableScan(Probe)
-    ├── table: default.default.t1
-    ├── output columns: [number (#1)]
-    ├── read rows: 10
-    ├── read bytes: 65
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+└── Filter(Probe)
+    ├── output columns: [t1.number (#1)]
+    ├── filters: [t1.number (#1) = t1.number (#1) + 1]
+    ├── estimated rows: 2.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [number (#1)]
+        ├── read rows: 10
+        ├── read bytes: 65
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [t1.number (#1) = t1.number (#1) + 1], limit: NONE]
+        └── estimated rows: 10.00
 
 query T
 explain select t.number from t, t1 where t.number > 1 and 1 < t1.number
@@ -87,7 +91,7 @@ HashJoin
 ├── build keys: []
 ├── probe keys: []
 ├── filters: []
-├── estimated rows: 0.18
+├── estimated rows: 1.64
 ├── Filter(Build)
 │   ├── output columns: [t.number (#0)]
 │   ├── filters: [t.number (#0) > 1]
@@ -104,8 +108,8 @@ HashJoin
 │       └── estimated rows: 1.00
 └── Filter(Probe)
     ├── output columns: []
-    ├── filters: [1 < t1.number (#1)]
-    ├── estimated rows: 0.91
+    ├── filters: [t1.number (#1) > 1]
+    ├── estimated rows: 8.18
     └── TableScan
         ├── table: default.default.t1
         ├── output columns: [number (#1)]
@@ -114,7 +118,7 @@ HashJoin
         ├── partitions total: 1
         ├── partitions scanned: 1
         ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-        ├── push downs: [filters: [1 < t1.number (#1)], limit: NONE]
+        ├── push downs: [filters: [t1.number (#1) > 1], limit: NONE]
         └── estimated rows: 10.00
 
 query T
@@ -383,36 +387,40 @@ query T
 explain select * from onecolumn as a right join twocolumn as b on a.x = b.x where b.x > 42 and b.x < 45
 ----
 HashJoin
-├── output columns: [a.x (#0), b.x (#1), b.y (#2)]
-├── join type: RIGHT OUTER
-├── build keys: [b.x (#1)]
-├── probe keys: [a.x (#0)]
+├── output columns: [b.x (#1), b.y (#2), a.x (#0)]
+├── join type: LEFT OUTER
+├── build keys: [a.x (#0)]
+├── probe keys: [b.x (#1)]
 ├── filters: []
 ├── estimated rows: 3.20
 ├── Filter(Build)
-│   ├── output columns: [b.x (#1), b.y (#2)]
-│   ├── filters: [is_true(b.x (#1) > 42), is_true(b.x (#1) < 45)]
-│   ├── estimated rows: 3.20
+│   ├── output columns: [a.x (#0)]
+│   ├── filters: [is_true(a.x (#0) > 42), is_true(a.x (#0) < 45)]
+│   ├── estimated rows: 3.00
 │   └── TableScan
-│       ├── table: default.default.twocolumn
-│       ├── output columns: [x (#1), y (#2)]
+│       ├── table: default.default.onecolumn
+│       ├── output columns: [x (#0)]
 │       ├── read rows: 4
-│       ├── read bytes: 94
+│       ├── read bytes: 45
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-│       ├── push downs: [filters: [and_filters(twocolumn.x (#1) > 42, twocolumn.x (#1) < 45)], limit: NONE]
+│       ├── push downs: [filters: [and_filters(onecolumn.x (#0) > 42, onecolumn.x (#0) < 45)], limit: NONE]
 │       └── estimated rows: 4.00
-└── TableScan(Probe)
-    ├── table: default.default.onecolumn
-    ├── output columns: [x (#0)]
-    ├── read rows: 4
-    ├── read bytes: 45
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 4.00
+└── Filter(Probe)
+    ├── output columns: [b.x (#1), b.y (#2)]
+    ├── filters: [is_true(b.x (#1) > 42), is_true(b.x (#1) < 45)]
+    ├── estimated rows: 3.20
+    └── TableScan
+        ├── table: default.default.twocolumn
+        ├── output columns: [x (#1), y (#2)]
+        ├── read rows: 4
+        ├── read bytes: 94
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [and_filters(twocolumn.x (#1) > 42, twocolumn.x (#1) < 45)], limit: NONE]
+        └── estimated rows: 4.00
 
 statement ok
 drop table t
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/chain.test b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/chain.test
index e0fec1332775..dceb3bf47c8e 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/chain.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/chain.test
@@ -29,9 +29,9 @@ query T
 explain select * from t, t1, t2 where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#2), t.a (#0), t1.a (#1)]
+├── output columns: [t2.a (#2), t1.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t1.a (#1)]
+├── build keys: [t.a (#0)]
 ├── probe keys: [t2.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -77,17 +77,17 @@ query T
 explain select * from t, t2, t1 where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#1), t.a (#0), t1.a (#2)]
+├── output columns: [t1.a (#2), t2.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t1.a (#2)]
-├── probe keys: [t2.a (#1)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t1.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t1.a (#2), t.a (#0)]
+│   ├── output columns: [t2.a (#1), t.a (#0)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#0)]
-│   ├── probe keys: [t1.a (#2)]
+│   ├── probe keys: [t2.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -101,41 +101,41 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t1
-│       ├── output columns: [a (#2)]
-│       ├── read rows: 10
-│       ├── read bytes: 65
+│       ├── table: default.join_reorder.t2
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 100
+│       ├── read bytes: 172
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 10.00
+│       └── estimated rows: 100.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t2
-    ├── output columns: [a (#1)]
-    ├── read rows: 100
-    ├── read bytes: 172
+    ├── table: default.join_reorder.t1
+    ├── output columns: [a (#2)]
+    ├── read rows: 10
+    ├── read bytes: 65
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 100.00
+    └── estimated rows: 10.00
 
 query T
 explain select * from t1, t, t2 where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#2), t.a (#1), t1.a (#0)]
+├── output columns: [t1.a (#0), t.a (#1), t2.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#0)]
-├── probe keys: [t2.a (#2)]
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t1.a (#0), t.a (#1)]
+│   ├── output columns: [t2.a (#2), t.a (#1)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#1)]
-│   ├── probe keys: [t1.a (#0)]
+│   ├── probe keys: [t2.a (#2)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -149,41 +149,41 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t1
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 10
-│       ├── read bytes: 65
+│       ├── table: default.join_reorder.t2
+│       ├── output columns: [a (#2)]
+│       ├── read rows: 100
+│       ├── read bytes: 172
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 10.00
+│       └── estimated rows: 100.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t2
-    ├── output columns: [a (#2)]
-    ├── read rows: 100
-    ├── read bytes: 172
+    ├── table: default.join_reorder.t1
+    ├── output columns: [a (#0)]
+    ├── read rows: 10
+    ├── read bytes: 65
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 100.00
+    └── estimated rows: 10.00
 
 query T
 explain select * from t1, t2, t where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#1), t.a (#2), t1.a (#0)]
+├── output columns: [t1.a (#0), t2.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#0)]
-├── probe keys: [t2.a (#1)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t1.a (#0), t.a (#2)]
+│   ├── output columns: [t2.a (#1), t.a (#2)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#2)]
-│   ├── probe keys: [t1.a (#0)]
+│   ├── probe keys: [t2.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -197,33 +197,33 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t1
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 10
-│       ├── read bytes: 65
+│       ├── table: default.join_reorder.t2
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 100
+│       ├── read bytes: 172
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 10.00
+│       └── estimated rows: 100.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t2
-    ├── output columns: [a (#1)]
-    ├── read rows: 100
-    ├── read bytes: 172
+    ├── table: default.join_reorder.t1
+    ├── output columns: [a (#0)]
+    ├── read rows: 10
+    ├── read bytes: 65
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 100.00
+    └── estimated rows: 10.00
 
 query T
 explain select * from t2, t1, t where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#0), t.a (#2), t1.a (#1)]
+├── output columns: [t2.a (#0), t1.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#1)]
+├── build keys: [t.a (#2)]
 ├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -504,9 +504,9 @@ explain join select * from t right anti join t1 on t1.a = t.a
 ----
 HashJoin: LEFT ANTI
 ├── Build
-│   └── Scan: default.join_reorder.t (read rows: 1)
+│   └── Scan: default.join_reorder.t (#0) (read rows: 1)
 └── Probe
-    └── Scan: default.join_reorder.t1 (read rows: 10)
+    └── Scan: default.join_reorder.t1 (#1) (read rows: 10)
 
 statement ok
 drop database join_reorder
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/cycles.test b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/cycles.test
index 58aada62528b..b4b5e7651757 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/cycles.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/cycles.test
@@ -22,8 +22,8 @@ explain select * from t, t1, t2 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t2.a (#2), t1.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t.a (#0), t1.a (#1)]
-├── probe keys: [t2.a (#2), t2.a (#2)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t2.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -70,8 +70,8 @@ explain select * from t, t2, t1 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t1.a (#2), t2.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t.a (#0), t2.a (#1)]
-├── probe keys: [t1.a (#2), t1.a (#2)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t1.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -116,10 +116,10 @@ query T
 explain select * from t1, t, t2 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 ----
 HashJoin
-├── output columns: [t1.a (#0), t2.a (#2), t.a (#1)]
+├── output columns: [t1.a (#0), t.a (#1), t2.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#2), t.a (#1)]
-├── probe keys: [t1.a (#0), t1.a (#0)]
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -166,8 +166,8 @@ explain select * from t1, t2, t where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t1.a (#0), t2.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t.a (#2), t2.a (#1)]
-├── probe keys: [t1.a (#0), t1.a (#0)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -214,8 +214,8 @@ explain select * from t2, t1, t where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t2.a (#0), t1.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t.a (#2), t1.a (#1)]
-├── probe keys: [t2.a (#0), t2.a (#0)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -260,10 +260,10 @@ query T
 explain select * from t2, t, t1 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 ----
 HashJoin
-├── output columns: [t2.a (#0), t1.a (#2), t.a (#1)]
+├── output columns: [t2.a (#0), t.a (#1), t1.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#2), t.a (#1)]
-├── probe keys: [t2.a (#0), t2.a (#0)]
+├── build keys: [t1.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/mark.test b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/mark.test
index 86ffbb800d07..8740e9732bf1 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/mark.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/mark.test
@@ -55,3 +55,110 @@ HashJoin
     ├── partitions scanned: 1
     ├── push downs: [filters: [], limit: NONE]
     └── estimated rows: 10000.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(c int, d int);
+
+statement ok
+insert into t1 values(1, 2), (2, 3), (3, 4);
+
+statement ok
+insert into t2 values(1, 2), (2, 3), (3, 4), (4, 5);
+
+query T
+explain SELECT
+    (SELECT IF(EXISTS(SELECT 1 FROM t1 WHERE t1.a = t2.c), '1', '0'))
+FROM
+    t2;
+----
+EvalScalar
+├── output columns: [(select if(exists (select 1 from t1 where (t1.a = t2.c)), '1', '0')) (#6)]
+├── expressions: [scalar_subquery_5 (#5)]
+├── estimated rows: 4.00
+└── HashJoin
+    ├── output columns: [if(exists (select 1 from t1 where (t1.a = t2.c)), '1', '0') (#5)]
+    ├── join type: LEFT SINGLE
+    ├── build keys: [c (#8)]
+    ├── probe keys: [c (#0)]
+    ├── filters: []
+    ├── estimated rows: 4.00
+    ├── EvalScalar(Build)
+    │   ├── output columns: [c (#8), if(exists (select 1 from t1 where (t1.a = t2.c)), '1', '0') (#5)]
+    │   ├── expressions: [if(7 (#7), '1', '0')]
+    │   ├── estimated rows: 0.00
+    │   └── HashJoin
+    │       ├── output columns: [c (#8), marker (#7)]
+    │       ├── join type: LEFT MARK
+    │       ├── build keys: [c (#8)]
+    │       ├── probe keys: [a (#2)]
+    │       ├── filters: []
+    │       ├── estimated rows: 0.00
+    │       ├── HashJoin(Build)
+    │       │   ├── output columns: [c (#8)]
+    │       │   ├── join type: CROSS
+    │       │   ├── build keys: []
+    │       │   ├── probe keys: []
+    │       │   ├── filters: []
+    │       │   ├── estimated rows: 0.00
+    │       │   ├── AggregateFinal(Build)
+    │       │   │   ├── output columns: [c (#8)]
+    │       │   │   ├── group by: [c]
+    │       │   │   ├── aggregate functions: []
+    │       │   │   ├── estimated rows: 0.00
+    │       │   │   └── AggregatePartial
+    │       │   │       ├── output columns: [#_group_by_key]
+    │       │   │       ├── group by: [c]
+    │       │   │       ├── aggregate functions: []
+    │       │   │       ├── estimated rows: 0.00
+    │       │   │       └── TableScan
+    │       │   │           ├── table: default.default.t2
+    │       │   │           ├── output columns: [c (#8)]
+    │       │   │           ├── read rows: 4
+    │       │   │           ├── read bytes: 49
+    │       │   │           ├── partitions total: 1
+    │       │   │           ├── partitions scanned: 1
+    │       │   │           ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    │       │   │           ├── push downs: [filters: [], limit: NONE]
+    │       │   │           └── estimated rows: 0.00
+    │       │   └── DummyTableScan(Probe)
+    │       └── Filter(Probe)
+    │           ├── output columns: [t1.a (#2)]
+    │           ├── filters: [is_true(a (#2) = a (#2))]
+    │           ├── estimated rows: 0.60
+    │           └── TableScan
+    │               ├── table: default.default.t1
+    │               ├── output columns: [a (#2)]
+    │               ├── read rows: 3
+    │               ├── read bytes: 45
+    │               ├── partitions total: 1
+    │               ├── partitions scanned: 1
+    │               ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    │               ├── push downs: [filters: [is_true(t1.a (#2) = t1.a (#2))], limit: NONE]
+    │               └── estimated rows: 3.00
+    └── TableScan(Probe)
+        ├── table: default.default.t2
+        ├── output columns: [c (#0)]
+        ├── read rows: 4
+        ├── read bytes: 49
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 4.00
+
+
+
+statement ok
+drop table t1;
+
+statement ok
+drop table t2
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/star.test b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/star.test
index c21e4714f6b1..796e7576ef57 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/star.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/join_reorder/star.test
@@ -20,17 +20,17 @@ query T
 explain select * from t, t1, t2 where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#1), t.a (#0), t2.a (#2)]
+├── output columns: [t2.a (#2), t1.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t2.a (#2)]
-├── probe keys: [t1.a (#1)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t2.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t2.a (#2), t.a (#0)]
+│   ├── output columns: [t1.a (#1), t.a (#0)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#0)]
-│   ├── probe keys: [t2.a (#2)]
+│   ├── probe keys: [t1.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -44,33 +44,33 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t2
-│       ├── output columns: [a (#2)]
-│       ├── read rows: 100
-│       ├── read bytes: 172
+│       ├── table: default.join_reorder.t1
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 10
+│       ├── read bytes: 65
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 100.00
+│       └── estimated rows: 10.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t1
-    ├── output columns: [a (#1)]
-    ├── read rows: 10
-    ├── read bytes: 65
+    ├── table: default.join_reorder.t2
+    ├── output columns: [a (#2)]
+    ├── read rows: 100
+    ├── read bytes: 172
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    └── estimated rows: 100.00
 
 query T
 explain select * from t, t2, t1 where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#2), t.a (#0), t2.a (#1)]
+├── output columns: [t1.a (#2), t2.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t2.a (#1)]
+├── build keys: [t.a (#0)]
 ├── probe keys: [t1.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -164,9 +164,9 @@ query T
 explain select * from t1, t2, t where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#0), t.a (#2), t2.a (#1)]
+├── output columns: [t1.a (#0), t2.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#1)]
+├── build keys: [t.a (#2)]
 ├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -212,17 +212,17 @@ query T
 explain select * from t2, t1, t where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#1), t.a (#2), t2.a (#0)]
+├── output columns: [t2.a (#0), t1.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#0)]
-├── probe keys: [t1.a (#1)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t2.a (#0), t.a (#2)]
+│   ├── output columns: [t1.a (#1), t.a (#2)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#2)]
-│   ├── probe keys: [t2.a (#0)]
+│   ├── probe keys: [t1.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -236,41 +236,41 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t2
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 100
-│       ├── read bytes: 172
+│       ├── table: default.join_reorder.t1
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 10
+│       ├── read bytes: 65
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 100.00
+│       └── estimated rows: 10.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t1
-    ├── output columns: [a (#1)]
-    ├── read rows: 10
-    ├── read bytes: 65
+    ├── table: default.join_reorder.t2
+    ├── output columns: [a (#0)]
+    ├── read rows: 100
+    ├── read bytes: 172
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    └── estimated rows: 100.00
 
 query T
 explain select * from t2, t, t1 where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#2), t.a (#1), t2.a (#0)]
+├── output columns: [t2.a (#0), t.a (#1), t1.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#0)]
-├── probe keys: [t1.a (#2)]
+├── build keys: [t1.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t2.a (#0), t.a (#1)]
+│   ├── output columns: [t1.a (#2), t.a (#1)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#1)]
-│   ├── probe keys: [t2.a (#0)]
+│   ├── probe keys: [t1.a (#2)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -284,25 +284,25 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t2
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 100
-│       ├── read bytes: 172
+│       ├── table: default.join_reorder.t1
+│       ├── output columns: [a (#2)]
+│       ├── read rows: 10
+│       ├── read bytes: 65
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 100.00
+│       └── estimated rows: 10.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t1
-    ├── output columns: [a (#2)]
-    ├── read rows: 10
-    ├── read bytes: 65
+    ├── table: default.join_reorder.t2
+    ├── output columns: [a (#0)]
+    ├── read rows: 100
+    ├── read bytes: 172
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    └── estimated rows: 100.00
 
 statement ok
 drop database join_reorder
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/lateral.test b/tests/sqllogictests/suites/mode/standalone/explain/lateral.test
index bc0b00c6a2cd..e4466bada019 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/lateral.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/lateral.test
@@ -2,15 +2,15 @@ query T
 explain select * from numbers(10) t(a), lateral(select * from numbers(10) t1(a) where t.a = t1.a) t1
 ----
 HashJoin
-├── output columns: [t1.number (#1), t.number (#0)]
+├── output columns: [t.number (#0), t1.number (#1)]
 ├── join type: INNER
-├── build keys: [number (#0)]
-├── probe keys: [number (#1)]
+├── build keys: [number (#1)]
+├── probe keys: [number (#0)]
 ├── filters: []
 ├── estimated rows: 100.00
 ├── TableScan(Build)
 │   ├── table: default.system.numbers
-│   ├── output columns: [number (#0)]
+│   ├── output columns: [number (#1)]
 │   ├── read rows: 10
 │   ├── read bytes: 80
 │   ├── partitions total: 1
@@ -19,7 +19,7 @@ HashJoin
 │   └── estimated rows: 10.00
 └── TableScan(Probe)
     ├── table: default.system.numbers
-    ├── output columns: [number (#1)]
+    ├── output columns: [number (#0)]
     ├── read rows: 10
     ├── read bytes: 80
     ├── partitions total: 1
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/lazy_read.test b/tests/sqllogictests/suites/mode/standalone/explain/lazy_read.test
index 18d91e281dd9..25ba90fb681b 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/lazy_read.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/lazy_read.test
@@ -97,7 +97,7 @@ RowFetch
         ├── estimated rows: 0.00
         └── Filter
             ├── output columns: [t_11831.uid (#0), t_11831.time (#3), t_11831._row_id (#4)]
-            ├── filters: [is_true(t_11831.uid (#0) = 11), is_true(t_11831.time (#3) >= 1686672000000), is_true(t_11831.time (#3) <= 1686758399000)]
+            ├── filters: [is_true(t_11831.time (#3) >= 1686672000000), is_true(t_11831.time (#3) <= 1686758399000), is_true(t_11831.uid (#0) = 11)]
             ├── estimated rows: 0.00
             └── TableScan
                 ├── table: default.default.t_11831
@@ -106,7 +106,7 @@ RowFetch
                 ├── read bytes: 0
                 ├── partitions total: 0
                 ├── partitions scanned: 0
-                ├── push downs: [filters: [and_filters(and_filters(and_filters(and_filters(t_11831.time (#3) >= 1686672000000, t_11831.time (#3) <= 1686758399000), t_11831.uid (#0) = 11), t_11831.time (#3) >= 1686672000000), t_11831.time (#3) <= 1686758399000)], limit: NONE]
+                ├── push downs: [filters: [and_filters(and_filters(t_11831.time (#3) >= 1686672000000, t_11831.time (#3) <= 1686758399000), t_11831.uid (#0) = 11)], limit: NONE]
                 └── estimated rows: 0.00
 
 statement ok
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/limit.test b/tests/sqllogictests/suites/mode/standalone/explain/limit.test
index 12c1a26d4a41..c33a8410dee6 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/limit.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/limit.test
@@ -97,13 +97,13 @@ Limit
             ├── estimated rows: 0.20
             └── Filter
                 ├── output columns: [t.number (#0)]
-                ├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(CAST(is_not_null(scalar_subquery_4 (#4)) AS Boolean NULL), scalar_subquery_4 (#4), 0))]
+                ├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(true, TRY_CAST(scalar_subquery_4 (#4) AS UInt64 NULL), 0))]
                 ├── estimated rows: 0.20
                 └── HashJoin
                     ├── output columns: [t.number (#0), COUNT(*) (#4)]
-                    ├── join type: LEFT SINGLE
+                    ├── join type: INNER
                     ├── build keys: [number (#2)]
-                    ├── probe keys: [CAST(number (#0) AS UInt64 NULL)]
+                    ├── probe keys: [number (#0)]
                     ├── filters: []
                     ├── estimated rows: 1.00
                     ├── AggregateFinal(Build)
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/outer_to_inner.test b/tests/sqllogictests/suites/mode/standalone/explain/outer_to_inner.test
new file mode 100644
index 000000000000..7d515713b154
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/outer_to_inner.test
@@ -0,0 +1,190 @@
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+# table with string column
+statement ok
+create table t1(a string, b string);
+
+statement ok
+create table t2(a string, b string);
+
+query T
+explain  select * from t1 left join t2 on t1.b = t2.b where t2.a > 'a';
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: INNER
+├── build keys: [t2.b (#3)]
+├── probe keys: [t1.b (#1)]
+├── filters: []
+├── estimated rows: 0.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 'a')]
+│   ├── estimated rows: 0.00
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 'a')], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 0.00
+
+
+statement ok
+drop table t1;
+
+statement ok
+drop table t2;
+
+# table with date column
+statement ok
+create table t1(a date, b date);
+
+statement ok
+create table t2(a date, b date);
+
+query T
+explain  select * from t1 left join t2 on t1.b = t2.b where t2.a > '2022-01-01 02:00:11';
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: INNER
+├── build keys: [t2.b (#3)]
+├── probe keys: [t1.b (#1)]
+├── filters: []
+├── estimated rows: 0.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > '2022-01-01')]
+│   ├── estimated rows: 0.00
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [is_true(t2.a (#2) > '2022-01-01')], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 0.00
+
+statement ok
+drop table t1;
+
+statement ok
+drop table t2;
+
+# table with decimal column
+statement ok
+create table t1(a decimal(4, 2), b decimal(4, 2));
+
+statement ok
+create table t2(a decimal(4, 2), b decimal(4, 2));
+
+query T
+explain  select * from t1 left join t2 on t1.b = t2.b where t2.a > 1.1;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: INNER
+├── build keys: [t2.b (#3)]
+├── probe keys: [t1.b (#1)]
+├── filters: []
+├── estimated rows: 0.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 1.10)]
+│   ├── estimated rows: 0.00
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 1.10)], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 0.00
+
+statement ok
+drop table t1;
+
+statement ok
+drop table t2;
+
+# table with float column
+statement ok
+create table t1(a float, b float);
+
+statement ok
+create table t2(a float, b float);
+
+query T
+explain  select * from t1 left join t2 on t1.b = t2.b where t2.a > 1.1;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: INNER
+├── build keys: [t2.b (#3)]
+├── probe keys: [t1.b (#1)]
+├── filters: []
+├── estimated rows: 0.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 1.1)]
+│   ├── estimated rows: 0.00
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 1.1)], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 0.00
+
+statement ok
+drop table t1;
+
+statement ok
+drop table t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/prune_column.test b/tests/sqllogictests/suites/mode/standalone/explain/prune_column.test
index a8e16b1a1603..d67f409f36af 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/prune_column.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/prune_column.test
@@ -117,90 +117,91 @@ explain select t1.a from (select number + 1 as a, number + 1 as b from numbers(1
 HashJoin
 ├── output columns: [a (#1)]
 ├── join type: INNER
-├── build keys: [CAST(t1.a (#1) AS UInt64 NULL)]
-├── probe keys: [_if_scalar_subquery (#15)]
+├── build keys: [_if_scalar_subquery (#15)]
+├── probe keys: [CAST(t1.a (#1) AS UInt64 NULL)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── EvalScalar(Build)
-│   ├── output columns: [a (#1)]
-│   ├── expressions: [numbers.number (#0) + 1]
+│   ├── output columns: [_if_scalar_subquery (#15)]
+│   ├── expressions: [if(CAST(_count_scalar_subquery (#13) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#14))]
 │   ├── estimated rows: 1.00
-│   └── TableScan
-│       ├── table: default.system.numbers
-│       ├── output columns: [number (#0)]
-│       ├── read rows: 1
-│       ├── read bytes: 8
-│       ├── partitions total: 1
-│       ├── partitions scanned: 1
-│       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 1.00
+│   └── Limit
+│       ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
+│       ├── limit: 1
+│       ├── offset: 0
+│       ├── estimated rows: 1.00
+│       └── AggregateFinal
+│           ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
+│           ├── group by: []
+│           ├── aggregate functions: [count(), any(COUNT(*))]
+│           ├── limit: 1
+│           ├── estimated rows: 1.00
+│           └── AggregatePartial
+│               ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
+│               ├── group by: []
+│               ├── aggregate functions: [count(), any(COUNT(*))]
+│               ├── estimated rows: 1.00
+│               └── AggregateFinal
+│                   ├── output columns: [COUNT(*) (#12)]
+│                   ├── group by: []
+│                   ├── aggregate functions: [count()]
+│                   ├── estimated rows: 1.00
+│                   └── AggregatePartial
+│                       ├── output columns: [COUNT(*) (#12)]
+│                       ├── group by: []
+│                       ├── aggregate functions: [count()]
+│                       ├── estimated rows: 1.00
+│                       └── HashJoin
+│                           ├── output columns: []
+│                           ├── join type: INNER
+│                           ├── build keys: [t2.b (#5)]
+│                           ├── probe keys: [t3.b (#10)]
+│                           ├── filters: []
+│                           ├── estimated rows: 0.20
+│                           ├── EvalScalar(Build)
+│                           │   ├── output columns: [b (#5)]
+│                           │   ├── expressions: [numbers.number (#3) + 1]
+│                           │   ├── estimated rows: 0.20
+│                           │   └── Filter
+│                           │       ├── output columns: [numbers.number (#3)]
+│                           │       ├── filters: [numbers.number (#3) + 1 = 1]
+│                           │       ├── estimated rows: 0.20
+│                           │       └── TableScan
+│                           │           ├── table: default.system.numbers
+│                           │           ├── output columns: [number (#3)]
+│                           │           ├── read rows: 1
+│                           │           ├── read bytes: 8
+│                           │           ├── partitions total: 1
+│                           │           ├── partitions scanned: 1
+│                           │           ├── push downs: [filters: [numbers.number (#3) + 1 = 1], limit: NONE]
+│                           │           └── estimated rows: 1.00
+│                           └── EvalScalar(Probe)
+│                               ├── output columns: [b (#10)]
+│                               ├── expressions: [numbers.number (#8) + 1]
+│                               ├── estimated rows: 1.00
+│                               └── TableScan
+│                                   ├── table: default.system.numbers
+│                                   ├── output columns: [number (#8)]
+│                                   ├── read rows: 1
+│                                   ├── read bytes: 8
+│                                   ├── partitions total: 1
+│                                   ├── partitions scanned: 1
+│                                   ├── push downs: [filters: [], limit: NONE]
+│                                   └── estimated rows: 1.00
 └── EvalScalar(Probe)
-    ├── output columns: [_if_scalar_subquery (#15)]
-    ├── expressions: [if(CAST(_count_scalar_subquery (#13) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#14))]
+    ├── output columns: [a (#1)]
+    ├── expressions: [numbers.number (#0) + 1]
     ├── estimated rows: 1.00
-    └── Limit
-        ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
-        ├── limit: 1
-        ├── offset: 0
-        ├── estimated rows: 1.00
-        └── AggregateFinal
-            ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
-            ├── group by: []
-            ├── aggregate functions: [count(), any(COUNT(*))]
-            ├── limit: 1
-            ├── estimated rows: 1.00
-            └── AggregatePartial
-                ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
-                ├── group by: []
-                ├── aggregate functions: [count(), any(COUNT(*))]
-                ├── estimated rows: 1.00
-                └── AggregateFinal
-                    ├── output columns: [COUNT(*) (#12)]
-                    ├── group by: []
-                    ├── aggregate functions: [count()]
-                    ├── estimated rows: 1.00
-                    └── AggregatePartial
-                        ├── output columns: [COUNT(*) (#12)]
-                        ├── group by: []
-                        ├── aggregate functions: [count()]
-                        ├── estimated rows: 1.00
-                        └── HashJoin
-                            ├── output columns: []
-                            ├── join type: INNER
-                            ├── build keys: [t2.b (#5)]
-                            ├── probe keys: [t3.b (#10)]
-                            ├── filters: []
-                            ├── estimated rows: 0.20
-                            ├── EvalScalar(Build)
-                            │   ├── output columns: [b (#5)]
-                            │   ├── expressions: [numbers.number (#3) + 1]
-                            │   ├── estimated rows: 0.20
-                            │   └── Filter
-                            │       ├── output columns: [numbers.number (#3)]
-                            │       ├── filters: [numbers.number (#3) + 1 = 1]
-                            │       ├── estimated rows: 0.20
-                            │       └── TableScan
-                            │           ├── table: default.system.numbers
-                            │           ├── output columns: [number (#3)]
-                            │           ├── read rows: 1
-                            │           ├── read bytes: 8
-                            │           ├── partitions total: 1
-                            │           ├── partitions scanned: 1
-                            │           ├── push downs: [filters: [numbers.number (#3) + 1 = 1], limit: NONE]
-                            │           └── estimated rows: 1.00
-                            └── EvalScalar(Probe)
-                                ├── output columns: [b (#10)]
-                                ├── expressions: [numbers.number (#8) + 1]
-                                ├── estimated rows: 1.00
-                                └── TableScan
-                                    ├── table: default.system.numbers
-                                    ├── output columns: [number (#8)]
-                                    ├── read rows: 1
-                                    ├── read bytes: 8
-                                    ├── partitions total: 1
-                                    ├── partitions scanned: 1
-                                    ├── push downs: [filters: [], limit: NONE]
-                                    └── estimated rows: 1.00
+    └── TableScan
+        ├── table: default.system.numbers
+        ├── output columns: [number (#0)]
+        ├── read rows: 1
+        ├── read bytes: 8
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 1.00
+
 
 query T
 explain select name from system.functions order by example
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_eval_scalar.test
similarity index 50%
rename from tests/sqllogictests/suites/mode/standalone/explain/push_down_filter.test
rename to tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_eval_scalar.test
index c339766166b1..43a459a500c7 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_eval_scalar.test
@@ -1,83 +1,3 @@
-# push down filter ProjectSet
-statement ok
-drop table if exists products;
-
-statement ok
-create table products(name varchar, details variant);
-
-statement ok
-insert into products(name, details) values ('Laptop', '{"brand": "Dell", "colors": ["Black", "Silver"], "price": 1200, "features": {"ram": "16GB", "storage": "512GB"}}'), ('Smartphone', '{"brand": "Apple", "colors": ["White", "Black"], "price": 999, "features": {"ram": "4GB", "storage": "128GB"}}'), ('Headphones', '{"brand": "Sony", "colors": ["Black", "Blue", "Red"], "price": 150, "features": {"battery": "20h", "bluetooth": "5.0"}}');
-
-query T
-explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
-----
-EvalScalar
-├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
-├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
-├── estimated rows: 0.12
-└── Filter
-    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
-    ├── filters: [is_true(TRY_CAST(get(1)(json_path_query (#2)) AS String NULL) = '512GB')]
-    ├── estimated rows: 0.12
-    └── ProjectSet
-        ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
-        ├── estimated rows: 0.60
-        ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
-        └── Filter
-            ├── output columns: [products.name (#0), products.details (#1)]
-            ├── filters: [is_true(products.name (#0) = 'Laptop'), is_true(TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')]
-            ├── estimated rows: 0.60
-            └── TableScan
-                ├── table: default.default.products
-                ├── output columns: [name (#0), details (#1)]
-                ├── read rows: 3
-                ├── read bytes: 328
-                ├── partitions total: 1
-                ├── partitions scanned: 1
-                ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
-                ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
-                └── estimated rows: 3.00
-
-query T
-select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
-----
-Laptop "512GB" "16GB"
-
-query T
-explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
-----
-EvalScalar
-├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
-├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
-├── estimated rows: 0.60
-└── ProjectSet
-    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
-    ├── estimated rows: 0.60
-    ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
-    └── Filter
-        ├── output columns: [products.name (#0), products.details (#1)]
-        ├── filters: [is_true(products.name (#0) = 'Laptop'), is_true(TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')]
-        ├── estimated rows: 0.60
-        └── TableScan
-            ├── table: default.default.products
-            ├── output columns: [name (#0), details (#1)]
-            ├── read rows: 3
-            ├── read bytes: 328
-            ├── partitions total: 1
-            ├── partitions scanned: 1
-            ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
-            ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
-            └── estimated rows: 3.00
-
-query T
-select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
-----
-Laptop "16GB" "16GB"
-Laptop "512GB" "16GB"
-
-statement ok
-drop table products;
-
 # push down filter EvalScalar
 statement ok
 drop table if exists t1;
@@ -179,15 +99,19 @@ AggregateFinal
         │               │           ├── output columns: [t2.sid (#1), sum_arg_0 (#4)]
         │               │           ├── expressions: [if(CAST(is_not_null(t3.val (#2)) AS Boolean NULL), CAST(assume_not_null(t3.val (#2)) AS Int32 NULL), true, 0, NULL)]
         │               │           ├── estimated rows: 0.00
-        │               │           └── TableScan
-        │               │               ├── table: default.default.t2
-        │               │               ├── output columns: [sid (#1), val (#2)]
-        │               │               ├── read rows: 0
-        │               │               ├── read bytes: 0
-        │               │               ├── partitions total: 0
-        │               │               ├── partitions scanned: 0
-        │               │               ├── push downs: [filters: [], limit: NONE]
-        │               │               └── estimated rows: 0.00
+        │               │           └── Filter
+        │               │               ├── output columns: [t2.sid (#1), t2.val (#2)]
+        │               │               ├── filters: [is_true(t3.sid (#1) = 1)]
+        │               │               ├── estimated rows: 0.00
+        │               │               └── TableScan
+        │               │                   ├── table: default.default.t2
+        │               │                   ├── output columns: [sid (#1), val (#2)]
+        │               │                   ├── read rows: 0
+        │               │                   ├── read bytes: 0
+        │               │                   ├── partitions total: 0
+        │               │                   ├── partitions scanned: 0
+        │               │                   ├── push downs: [filters: [is_true(t2.sid (#1) = 1)], limit: NONE]
+        │               │                   └── estimated rows: 0.00
         │               └── Filter(Probe)
         │                   ├── output columns: [t.id (#0)]
         │                   ├── filters: [is_true(t.id (#0) = 1)]
@@ -234,44 +158,13 @@ AggregateFinal
                                 └── estimated rows: 0.00
 
 statement ok
-drop table t1;
-
-statement ok
-drop table t2;
-
-statement ok
-drop view v1;
-
-statement ok
-drop view v2;
-
-# push down alias filter scan
-statement ok
-drop table if exists t;
+drop table if exists t1;
 
 statement ok
-create table t (x INT);
+drop table if exists t2;
 
 statement ok
-insert into t(x) values (1), (2);
-
-query I
-explain select * from t as a(id) where a.id > 1;
-----
-Filter
-├── output columns: [a.x (#0)]
-├── filters: [is_true(a.id (#0) > 1)]
-├── estimated rows: 0.40
-└── TableScan
-    ├── table: default.default.t
-    ├── output columns: [x (#0)]
-    ├── read rows: 2
-    ├── read bytes: 41
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [is_true(t.x (#0) > 1)], limit: NONE]
-    └── estimated rows: 2.00
+drop view if exists v1;
 
 statement ok
-drop table t;
+drop view if exists v2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_full_outer.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_full_outer.test
new file mode 100644
index 000000000000..d01e182e6db7
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_full_outer.test
@@ -0,0 +1,288 @@
+# push down filter full outer join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# convert full outer join to left outer join, can propagate t1.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t1.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# convert full outer join to right outer join, can propagate t2.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t2.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: RIGHT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# convert full outer join to left outer join, can not propagate t1.b > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t1.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 82
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.b (#1) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.b (#1) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# convert full outer join to right outer join, can not propagate t2.b > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t2.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: RIGHT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.b (#3) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.b (#3) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 90
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert full outer join to inner join, can propagate t1.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t1.a > 0 and t2.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 0), is_true(t2.b (#3) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [and_filters(t2.a (#2) > 0, t2.b (#3) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# convert full outer join to inner join, can propagate t2.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t2.a > 0 and t1.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 0), is_true(t1.b (#1) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [and_filters(t1.a (#0) > 0, t1.b (#1) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# full outer join, can not push down t2.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a and t2.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: FULL OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: [t2.a (#2) > 0]
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 82
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 90
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_inner.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_inner.test
new file mode 100644
index 000000000000..fa4ee4f10817
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_inner.test
@@ -0,0 +1,108 @@
+# push down filter inner join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# can propagate t1.a > 3
+query T
+explain select * from t1 inner join t2 on t1.a = t2.a where t1.a > 3;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 3)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 1
+│       ├── partitions scanned: 0
+│       ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 3)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 3)]
+    ├── estimated rows: 1.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 0
+        ├── read bytes: 0
+        ├── partitions total: 1
+        ├── partitions scanned: 0
+        ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 3)], limit: NONE]
+        └── estimated rows: 4.00
+
+# can propagate (t2.a > 1 or t2.a <= 2)
+query T
+explain select * from t1 inner join t2 on t1.a = t2.a where t2.a <= 2 or (t1.a > 1 and t2.a > 1);
+----
+Filter
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── filters: [is_true(t2.a (#2) <= 2 OR t1.a (#0) > 1 AND t2.a (#2) > 1)]
+├── estimated rows: 0.45
+└── HashJoin
+    ├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+    ├── join type: INNER
+    ├── build keys: [t2.a (#2)]
+    ├── probe keys: [t1.a (#0)]
+    ├── filters: []
+    ├── estimated rows: 1.26
+    ├── Filter(Build)
+    │   ├── output columns: [t2.a (#2), t2.b (#3)]
+    │   ├── filters: [is_true(t2.a (#2) <= 2 OR t2.a (#2) > 1)]
+    │   ├── estimated rows: 1.08
+    │   └── TableScan
+    │       ├── table: default.default.t2
+    │       ├── output columns: [a (#2), b (#3)]
+    │       ├── read rows: 3
+    │       ├── read bytes: 82
+    │       ├── partitions total: 1
+    │       ├── partitions scanned: 1
+    │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    │       ├── push downs: [filters: [is_true(t2.a (#2) <= 2 OR t2.a (#2) > 1)], limit: NONE]
+    │       └── estimated rows: 3.00
+    └── Filter(Probe)
+        ├── output columns: [t1.a (#0), t1.b (#1)]
+        ├── filters: [is_true(t1.a (#0) <= 2 OR t1.a (#0) > 1)]
+        ├── estimated rows: 3.50
+        └── TableScan
+            ├── table: default.default.t1
+            ├── output columns: [a (#0), b (#1)]
+            ├── read rows: 4
+            ├── read bytes: 90
+            ├── partitions total: 1
+            ├── partitions scanned: 1
+            ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+            ├── push downs: [filters: [is_true(t1.a (#0) <= 2 OR t1.a (#0) > 1)], limit: NONE]
+            └── estimated rows: 4.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_left_outer.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_left_outer.test
new file mode 100644
index 000000000000..ef83fa85610a
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_left_outer.test
@@ -0,0 +1,260 @@
+# push down filter left outer join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# left outer join, can propagate t1.a > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t1.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# left outer join, can not push down t1.a > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a and t1.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: [t1.a (#0) > 0]
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 82
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 90
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+# left outer join, can not propagate t1.b > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t1.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 82
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.b (#1) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.b (#1) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# convert left outer join to inner join, can propagate t2.a > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t2.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.a (#2) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0), t1.b (#1)]
+    ├── filters: [is_true(t1.a (#0) > 0)]
+    ├── estimated rows: 4.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 90
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+        └── estimated rows: 4.00
+
+# convert left outer join to inner join, can not propagate t2.b > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t2.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3)]
+│   ├── filters: [is_true(t2.b (#3) > 0)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 3
+│       ├── read bytes: 82
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.b (#3) > 0)], limit: NONE]
+│       └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 90
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+drop table if exists m1;
+
+statement ok
+drop table if exists j1;
+
+statement ok
+create table m1(id varchar, "context" varchar);
+
+statement ok
+create table j1(id varchar);
+
+# In `can_filter_null`, if the function is `assume_not_null` or `remove_nullable`, we cannot replace the column bindings with `Scalar::Null`.
+query T
+explain WITH base AS (SELECT id, context FROM m1), src1 AS (SELECT base.id FROM base WHERE IFNULL(base.context, '') = ''), join1 AS (SELECT id FROM j1) SELECT src1.id FROM src1 LEFT OUTER JOIN join1 ON TRUE;
+----
+HashJoin
+├── output columns: [m1.id (#0)]
+├── join type: LEFT OUTER
+├── build keys: []
+├── probe keys: []
+├── filters: []
+├── estimated rows: 0.00
+├── TableScan(Build)
+│   ├── table: default.default.j1
+│   ├── output columns: []
+│   ├── read rows: 0
+│   ├── read size: 0
+│   ├── partitions total: 0
+│   ├── partitions scanned: 0
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 0.00
+└── Filter(Probe)
+    ├── output columns: [m1.id (#0)]
+    ├── filters: [is_true(if(CAST(is_not_null(base.context (#1)) AS Boolean NULL), CAST(assume_not_null(base.context (#1)) AS String NULL), true, '', NULL) = '')]
+    ├── estimated rows: 0.00
+    └── TableScan
+        ├── table: default.default.m1
+        ├── output columns: [id (#0), context (#1)]
+        ├── read rows: 0
+        ├── read size: 0
+        ├── partitions total: 0
+        ├── partitions scanned: 0
+        ├── push downs: [filters: [is_true(if(CAST(is_not_null(m1.context (#1)) AS Boolean NULL), CAST(assume_not_null(m1.context (#1)) AS String NULL), true, '', NULL) = '')], limit: NONE]
+        └── estimated rows: 0.00
+
+statement ok
+drop table if exists m1;
+
+statement ok
+drop table if exists j1;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_semi_anti.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_semi_anti.test
new file mode 100644
index 000000000000..3b8c586910b0
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_join/push_down_filter_join_semi_anti.test
@@ -0,0 +1,104 @@
+# push down filter semi join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# left semi, can propagate t1.a > 3
+query T
+explain select t1.a from t1 where exists (select * from t2 where t1.a = t2.a) and t1.a > 3;
+----
+HashJoin
+├── output columns: [t1.a (#0)]
+├── join type: LEFT SEMI
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2)]
+│   ├── filters: [is_true(t2.a (#2) > 3)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 1
+│       ├── partitions scanned: 0
+│       ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 3)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0)]
+    ├── filters: [is_true(t1.a (#0) > 3)]
+    ├── estimated rows: 1.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0)]
+        ├── read rows: 0
+        ├── read bytes: 0
+        ├── partitions total: 1
+        ├── partitions scanned: 0
+        ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 3)], limit: NONE]
+        └── estimated rows: 4.00
+
+# left anti, can propagate t1.a > 3
+query T
+explain select t1.a from t1 where not exists (select * from t2 where t1.a = t2.a) and t1.a > 3;
+----
+HashJoin
+├── output columns: [t1.a (#0)]
+├── join type: LEFT ANTI
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 1.00
+├── Filter(Build)
+│   ├── output columns: [t2.a (#2)]
+│   ├── filters: [is_true(t2.a (#2) > 3)]
+│   ├── estimated rows: 0.60
+│   └── TableScan
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 1
+│       ├── partitions scanned: 0
+│       ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [is_true(t2.a (#2) > 3)], limit: NONE]
+│       └── estimated rows: 3.00
+└── Filter(Probe)
+    ├── output columns: [t1.a (#0)]
+    ├── filters: [is_true(t1.a (#0) > 3)]
+    ├── estimated rows: 1.00
+    └── TableScan
+        ├── table: default.default.t1
+        ├── output columns: [a (#0)]
+        ├── read rows: 0
+        ├── read bytes: 0
+        ├── partitions total: 1
+        ├── partitions scanned: 0
+        ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) > 3)], limit: NONE]
+        └── estimated rows: 4.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_project_set.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_project_set.test
new file mode 100644
index 000000000000..d068b6b420a8
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_project_set.test
@@ -0,0 +1,79 @@
+# push down filter ProjectSet
+statement ok
+drop table if exists products;
+
+statement ok
+create table products(name varchar, details variant);
+
+statement ok
+insert into products(name, details) values ('Laptop', '{"brand": "Dell", "colors": ["Black", "Silver"], "price": 1200, "features": {"ram": "16GB", "storage": "512GB"}}'), ('Smartphone', '{"brand": "Apple", "colors": ["White", "Black"], "price": 999, "features": {"ram": "4GB", "storage": "128GB"}}'), ('Headphones', '{"brand": "Sony", "colors": ["Black", "Blue", "Red"], "price": 150, "features": {"battery": "20h", "bluetooth": "5.0"}}');
+
+query T
+explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
+----
+EvalScalar
+├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
+├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
+├── estimated rows: 0.12
+└── Filter
+    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
+    ├── filters: [is_true(TRY_CAST(get(1)(json_path_query (#2)) AS String NULL) = '512GB')]
+    ├── estimated rows: 0.12
+    └── ProjectSet
+        ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
+        ├── estimated rows: 0.60
+        ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
+        └── Filter
+            ├── output columns: [products.name (#0), products.details (#1)]
+            ├── filters: [is_true(products.name (#0) = 'Laptop'), is_true(TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')]
+            ├── estimated rows: 0.60
+            └── TableScan
+                ├── table: default.default.products
+                ├── output columns: [name (#0), details (#1)]
+                ├── read rows: 3
+                ├── read bytes: 328
+                ├── partitions total: 1
+                ├── partitions scanned: 1
+                ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
+                ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
+                └── estimated rows: 3.00
+
+query T
+select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
+----
+Laptop "512GB" "16GB"
+
+query T
+explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
+----
+EvalScalar
+├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
+├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
+├── estimated rows: 0.60
+└── ProjectSet
+    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
+    ├── estimated rows: 0.60
+    ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
+    └── Filter
+        ├── output columns: [products.name (#0), products.details (#1)]
+        ├── filters: [is_true(products.name (#0) = 'Laptop'), is_true(TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')]
+        ├── estimated rows: 0.60
+        └── TableScan
+            ├── table: default.default.products
+            ├── output columns: [name (#0), details (#1)]
+            ├── read rows: 3
+            ├── read bytes: 328
+            ├── partitions total: 1
+            ├── partitions scanned: 1
+            ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
+            ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
+            └── estimated rows: 3.00
+
+query T
+select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
+----
+Laptop "16GB" "16GB"
+Laptop "512GB" "16GB"
+
+statement ok
+drop table if exists products;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_scan.test b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_scan.test
new file mode 100644
index 000000000000..00ae8d2eeee9
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain/push_down_filter/push_down_filter_scan.test
@@ -0,0 +1,30 @@
+# push down alias filter scan
+statement ok
+drop table if exists t;
+
+statement ok
+create table t (x INT);
+
+statement ok
+insert into t(x) values (1), (2);
+
+query I
+explain select * from t as a(id) where a.id > 1;
+----
+Filter
+├── output columns: [a.x (#0)]
+├── filters: [is_true(a.id (#0) > 1)]
+├── estimated rows: 0.40
+└── TableScan
+    ├── table: default.default.t
+    ├── output columns: [x (#0)]
+    ├── read rows: 2
+    ├── read bytes: 41
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t.x (#0) > 1)], limit: NONE]
+    └── estimated rows: 2.00
+
+statement ok
+drop table if exists t;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/select.test b/tests/sqllogictests/suites/mode/standalone/explain/select.test
index 9736c107cffa..5d7775e8c691 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/select.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/select.test
@@ -16,7 +16,7 @@ explain select * from (select * from numbers(1)) as t1 where number = 1
 ----
 Filter
 ├── output columns: [numbers.number (#0)]
-├── filters: [numbers.number (#0) = 1]
+├── filters: [t1.number (#0) = 1]
 ├── estimated rows: 0.00
 └── TableScan
     ├── table: default.system.numbers
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/sort.test b/tests/sqllogictests/suites/mode/standalone/explain/sort.test
index 1076a7fe9057..3c78ce7e1726 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/sort.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/sort.test
@@ -10,7 +10,7 @@ Sort
 ├── estimated rows: 0.00
 └── Filter
     ├── output columns: [t1.a (#0)]
-    ├── filters: [is_true(t1.a (#0) > 1)]
+    ├── filters: [is_true(t2.a (#0) > 1)]
     ├── estimated rows: 0.00
     └── TableScan
         ├── table: default.default.t1
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/subquery.test b/tests/sqllogictests/suites/mode/standalone/explain/subquery.test
index ac54f587ed08..b91ba286213a 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/subquery.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/subquery.test
@@ -3,13 +3,13 @@ explain select t.number from numbers(1) as t, numbers(1) as t1 where t.number =
 ----
 Filter
 ├── output columns: [t.number (#0)]
-├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(CAST(is_not_null(scalar_subquery_4 (#4)) AS Boolean NULL), scalar_subquery_4 (#4), 0))]
+├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(true, TRY_CAST(scalar_subquery_4 (#4) AS UInt64 NULL), 0))]
 ├── estimated rows: 0.20
 └── HashJoin
     ├── output columns: [t.number (#0), COUNT(*) (#4)]
-    ├── join type: LEFT SINGLE
+    ├── join type: INNER
     ├── build keys: [number (#2)]
-    ├── probe keys: [CAST(number (#0) AS UInt64 NULL)]
+    ├── probe keys: [number (#0)]
     ├── filters: []
     ├── estimated rows: 1.00
     ├── AggregateFinal(Build)
@@ -164,52 +164,52 @@ explain select t.number from numbers(1) as t where number = (select * from numbe
 HashJoin
 ├── output columns: [t.number (#0)]
 ├── join type: INNER
-├── build keys: [CAST(t.number (#0) AS UInt64 NULL)]
-├── probe keys: [_if_scalar_subquery (#4)]
+├── build keys: [_if_scalar_subquery (#4)]
+├── probe keys: [CAST(t.number (#0) AS UInt64 NULL)]
 ├── filters: []
 ├── estimated rows: 1.00
-├── TableScan(Build)
-│   ├── table: default.system.numbers
-│   ├── output columns: [number (#0)]
-│   ├── read rows: 1
-│   ├── read bytes: 8
-│   ├── partitions total: 1
-│   ├── partitions scanned: 1
-│   ├── push downs: [filters: [], limit: NONE]
-│   └── estimated rows: 1.00
-└── EvalScalar(Probe)
-    ├── output columns: [_if_scalar_subquery (#4)]
-    ├── expressions: [if(CAST(_count_scalar_subquery (#2) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#3))]
-    ├── estimated rows: 1.00
-    └── Limit
-        ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
-        ├── limit: 1
-        ├── offset: 0
-        ├── estimated rows: 1.00
-        └── AggregateFinal
-            ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
-            ├── group by: []
-            ├── aggregate functions: [count(), any(number)]
-            ├── limit: 1
-            ├── estimated rows: 1.00
-            └── AggregatePartial
-                ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
-                ├── group by: []
-                ├── aggregate functions: [count(), any(number)]
-                ├── estimated rows: 1.00
-                └── Filter
-                    ├── output columns: [numbers.number (#1)]
-                    ├── filters: [numbers.number (#1) = 0]
-                    ├── estimated rows: 0.00
-                    └── TableScan
-                        ├── table: default.system.numbers
-                        ├── output columns: [number (#1)]
-                        ├── read rows: 1
-                        ├── read bytes: 8
-                        ├── partitions total: 1
-                        ├── partitions scanned: 1
-                        ├── push downs: [filters: [numbers.number (#1) = 0], limit: NONE]
-                        └── estimated rows: 1.00
+├── EvalScalar(Build)
+│   ├── output columns: [_if_scalar_subquery (#4)]
+│   ├── expressions: [if(CAST(_count_scalar_subquery (#2) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#3))]
+│   ├── estimated rows: 1.00
+│   └── Limit
+│       ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
+│       ├── limit: 1
+│       ├── offset: 0
+│       ├── estimated rows: 1.00
+│       └── AggregateFinal
+│           ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
+│           ├── group by: []
+│           ├── aggregate functions: [count(), any(number)]
+│           ├── limit: 1
+│           ├── estimated rows: 1.00
+│           └── AggregatePartial
+│               ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
+│               ├── group by: []
+│               ├── aggregate functions: [count(), any(number)]
+│               ├── estimated rows: 1.00
+│               └── Filter
+│                   ├── output columns: [numbers.number (#1)]
+│                   ├── filters: [numbers.number (#1) = 0]
+│                   ├── estimated rows: 0.00
+│                   └── TableScan
+│                       ├── table: default.system.numbers
+│                       ├── output columns: [number (#1)]
+│                       ├── read rows: 1
+│                       ├── read bytes: 8
+│                       ├── partitions total: 1
+│                       ├── partitions scanned: 1
+│                       ├── push downs: [filters: [numbers.number (#1) = 0], limit: NONE]
+│                       └── estimated rows: 1.00
+└── TableScan(Probe)
+    ├── table: default.system.numbers
+    ├── output columns: [number (#0)]
+    ├── read rows: 1
+    ├── read bytes: 8
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 1.00
 
 query T
 explain select t.number from numbers(1) as t where exists (select * from numbers(1) where number = t.number)
@@ -310,7 +310,7 @@ HashJoin
 ├── estimated rows: 0.00
 ├── Filter(Build)
 │   ├── output columns: [numbers.number (#1)]
-│   ├── filters: [numbers.number (#1) = 0]
+│   ├── filters: [numbers.number (#1) < 10, numbers.number (#1) = 0]
 │   ├── estimated rows: 0.00
 │   └── TableScan
 │       ├── table: default.system.numbers
@@ -319,7 +319,7 @@ HashJoin
 │       ├── read bytes: 8
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
-│       ├── push downs: [filters: [numbers.number (#1) = 0], limit: NONE]
+│       ├── push downs: [filters: [and_filters(and_filters(numbers.number (#1) = 0, numbers.number (#1) < 10), numbers.number (#1) = 0)], limit: NONE]
 │       └── estimated rows: 1.00
 └── Filter(Probe)
     ├── output columns: [t.number (#0)]
@@ -396,65 +396,61 @@ HashJoin
 query T
 explain select t.number from numbers(1) as t, numbers(1) as t1 where (select count(*) = 1 from numbers(1) where t.number = number) and t.number = t1.number
 ----
-Filter
+HashJoin
 ├── output columns: [t.number (#0)]
-├── filters: [is_true(try_to_boolean(if(CAST(is_not_null(scalar_subquery_4 (#4)) AS Boolean NULL), TRY_CAST(scalar_subquery_4 (#4) AS UInt64 NULL), 0)))]
+├── join type: INNER
+├── build keys: [number (#2)]
+├── probe keys: [number (#0)]
+├── filters: []
 ├── estimated rows: 0.20
-└── HashJoin
-    ├── output columns: [t.number (#0), (count(*) = 1) (#4)]
-    ├── join type: LEFT SINGLE
-    ├── build keys: [number (#2)]
-    ├── probe keys: [CAST(number (#0) AS UInt64 NULL)]
+├── Filter(Build)
+│   ├── output columns: [numbers.number (#2)]
+│   ├── filters: [is_true(try_to_boolean(if(true, TRY_CAST(COUNT(*) (#3) = 1 AS UInt64 NULL), 0)))]
+│   ├── estimated rows: 0.20
+│   └── AggregateFinal
+│       ├── output columns: [COUNT(*) (#3), numbers.number (#2)]
+│       ├── group by: [number]
+│       ├── aggregate functions: [count()]
+│       ├── estimated rows: 1.00
+│       └── AggregatePartial
+│           ├── output columns: [COUNT(*) (#3), #_group_by_key]
+│           ├── group by: [number]
+│           ├── aggregate functions: [count()]
+│           ├── estimated rows: 1.00
+│           └── TableScan
+│               ├── table: default.system.numbers
+│               ├── output columns: [number (#2)]
+│               ├── read rows: 1
+│               ├── read bytes: 8
+│               ├── partitions total: 1
+│               ├── partitions scanned: 1
+│               ├── push downs: [filters: [], limit: NONE]
+│               └── estimated rows: 1.00
+└── HashJoin(Probe)
+    ├── output columns: [t.number (#0)]
+    ├── join type: INNER
+    ├── build keys: [t1.number (#1)]
+    ├── probe keys: [t.number (#0)]
     ├── filters: []
     ├── estimated rows: 1.00
-    ├── EvalScalar(Build)
-    │   ├── output columns: [numbers.number (#2), (count(*) = 1) (#4)]
-    │   ├── expressions: [COUNT(*) (#3) = 1]
-    │   ├── estimated rows: 1.00
-    │   └── AggregateFinal
-    │       ├── output columns: [COUNT(*) (#3), numbers.number (#2)]
-    │       ├── group by: [number]
-    │       ├── aggregate functions: [count()]
-    │       ├── estimated rows: 1.00
-    │       └── AggregatePartial
-    │           ├── output columns: [COUNT(*) (#3), #_group_by_key]
-    │           ├── group by: [number]
-    │           ├── aggregate functions: [count()]
-    │           ├── estimated rows: 1.00
-    │           └── TableScan
-    │               ├── table: default.system.numbers
-    │               ├── output columns: [number (#2)]
-    │               ├── read rows: 1
-    │               ├── read bytes: 8
-    │               ├── partitions total: 1
-    │               ├── partitions scanned: 1
-    │               ├── push downs: [filters: [], limit: NONE]
-    │               └── estimated rows: 1.00
-    └── HashJoin(Probe)
-        ├── output columns: [t.number (#0)]
-        ├── join type: INNER
-        ├── build keys: [t1.number (#1)]
-        ├── probe keys: [t.number (#0)]
-        ├── filters: []
-        ├── estimated rows: 1.00
-        ├── TableScan(Build)
-        │   ├── table: default.system.numbers
-        │   ├── output columns: [number (#1)]
-        │   ├── read rows: 1
-        │   ├── read bytes: 8
-        │   ├── partitions total: 1
-        │   ├── partitions scanned: 1
-        │   ├── push downs: [filters: [], limit: NONE]
-        │   └── estimated rows: 1.00
-        └── TableScan(Probe)
-            ├── table: default.system.numbers
-            ├── output columns: [number (#0)]
-            ├── read rows: 1
-            ├── read bytes: 8
-            ├── partitions total: 1
-            ├── partitions scanned: 1
-            ├── push downs: [filters: [], limit: NONE]
-            └── estimated rows: 1.00
+    ├── TableScan(Build)
+    │   ├── table: default.system.numbers
+    │   ├── output columns: [number (#1)]
+    │   ├── read rows: 1
+    │   ├── read bytes: 8
+    │   ├── partitions total: 1
+    │   ├── partitions scanned: 1
+    │   ├── push downs: [filters: [], limit: NONE]
+    │   └── estimated rows: 1.00
+    └── TableScan(Probe)
+        ├── table: default.system.numbers
+        ├── output columns: [number (#0)]
+        ├── read rows: 1
+        ├── read bytes: 8
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 1.00
 
 query T
 explain select t.number from numbers(1) as t where exists(select * from numbers(1) as t1 where t.number > t1.number) and not exists(select * from numbers(1) as t1 where t.number < t1.number)
@@ -506,7 +502,7 @@ explain select * from (select number as a from numbers(10)) as t(b) where t.b >
 ----
 Filter
 ├── output columns: [numbers.number (#0)]
-├── filters: [numbers.number (#0) > 5]
+├── filters: [t.b (#0) > 5]
 ├── estimated rows: 0.00
 └── TableScan
     ├── table: default.system.numbers
diff --git a/tests/sqllogictests/suites/mode/standalone/explain/window.test b/tests/sqllogictests/suites/mode/standalone/explain/window.test
index 73c33d8eccc3..c436b9018da3 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain/window.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain/window.test
@@ -64,7 +64,7 @@ explain SELECT k, v FROM (SELECT *, rank() OVER (PARTITION BY k ORDER BY v DESC)
 ----
 Filter
 ├── output columns: [test.k (#0), test.v (#1)]
-├── filters: [rank() OVER (PARTITION BY k ORDER BY v DESC) (#4) = 1]
+├── filters: [t2.rank (#4) = 1]
 ├── estimated rows: 0.00
 └── Window
     ├── output columns: [test.k (#0), test.v (#1), rank() OVER (PARTITION BY k ORDER BY v DESC) (#4)]
@@ -108,35 +108,35 @@ explain SELECT k, v FROM (SELECT *, rank() OVER (PARTITION BY v ORDER BY v DESC)
 ----
 Filter
 ├── output columns: [test.k (#0), test.v (#1)]
-├── filters: [rank() OVER (PARTITION BY v ORDER BY v DESC) (#4) = 1, is_true(t1.k (#0) = 12)]
+├── filters: [t2.rank (#4) = 1, is_true(t2.k (#0) = 12)]
 ├── estimated rows: 0.00
 └── Window
-  ├── output columns: [test.k (#0), test.v (#1), rank() OVER (PARTITION BY v ORDER BY v DESC) (#4)]
-  ├── aggregate function: [rank]
-  ├── partition by: [v]
-  ├── order by: [v]
-  ├── frame: [Range: Preceding(None) ~ CurrentRow]
-  └── UnionAll
-    ├── output columns: [test.k (#0), test.v (#1)]
-    ├── estimated rows: 0.00
-    ├── TableScan
-    │   ├── table: default.test_explain_window.test
-    │   ├── output columns: [k (#0), v (#1)]
-    │   ├── read rows: 0
-    │   ├── read bytes: 0
-    │   ├── partitions total: 0
-    │   ├── partitions scanned: 0
-    │   ├── push downs: [filters: [], limit: NONE]
-    │   └── estimated rows: 0.00
-    └── TableScan
-      ├── table: default.test_explain_window.test
-      ├── output columns: [k (#2), v (#3)]
-      ├── read rows: 0
-      ├── read bytes: 0
-      ├── partitions total: 0
-      ├── partitions scanned: 0
-      ├── push downs: [filters: [], limit: NONE]
-      └── estimated rows: 0.00
+    ├── output columns: [test.k (#0), test.v (#1), rank() OVER (PARTITION BY v ORDER BY v DESC) (#4)]
+    ├── aggregate function: [rank]
+    ├── partition by: [v]
+    ├── order by: [v]
+    ├── frame: [Range: Preceding(None) ~ CurrentRow]
+    └── UnionAll
+        ├── output columns: [test.k (#0), test.v (#1)]
+        ├── estimated rows: 0.00
+        ├── TableScan
+        │   ├── table: default.test_explain_window.test
+        │   ├── output columns: [k (#0), v (#1)]
+        │   ├── read rows: 0
+        │   ├── read bytes: 0
+        │   ├── partitions total: 0
+        │   ├── partitions scanned: 0
+        │   ├── push downs: [filters: [], limit: NONE]
+        │   └── estimated rows: 0.00
+        └── TableScan
+            ├── table: default.test_explain_window.test
+            ├── output columns: [k (#2), v (#3)]
+            ├── read rows: 0
+            ├── read bytes: 0
+            ├── partitions total: 0
+            ├── partitions scanned: 0
+            ├── push downs: [filters: [], limit: NONE]
+            └── estimated rows: 0.00
 
 # cannot push down filter in window function
 query T
@@ -144,7 +144,7 @@ explain SELECT k, v FROM (SELECT *, rank() OVER (ORDER BY v DESC) AS rank FROM (
 ----
 Filter
 ├── output columns: [test.k (#0), test.v (#1)]
-├── filters: [rank() OVER (ORDER BY v DESC) (#4) = 1, is_true(t1.k (#0) = 12)]
+├── filters: [t2.rank (#4) = 1, is_true(t2.k (#0) = 12)]
 ├── estimated rows: 0.00
 └── Window
     ├── output columns: [test.k (#0), test.v (#1), rank() OVER (ORDER BY v DESC) (#4)]
@@ -285,5 +285,91 @@ Filter
         ├── push downs: [filters: [], limit: NONE]
         └── estimated rows: 0.00
 
+# test push down limit to window function
+statement ok
+drop table if exists t
+
+statement ok
+create table t(a int)
+
+statement ok
+insert into t values(1), (2), (3), (3)
+
+# range frame (can not push down limit)
+query T
+explain pipeline select a, sum(a) over (partition by a order by a desc) from t limit 3
+----
+CompoundBlockOperator(Project) × 1 processor
+  LimitTransform × 1 processor
+    Transform Window × 1 processor
+      Merge (TransformSortMerge × 4 processors) to (Transform Window × 1)
+        TransformSortMerge × 4 processors
+          SortPartialTransform × 4 processors
+            Merge (DeserializeDataTransform × 1 processor) to (SortPartialTransform × 4)
+              DeserializeDataTransform × 1 processor
+                SyncReadParquetDataSource × 1 processor
+
+# range frame with ranking function (can push down limit)
+query T
+explain pipeline select a, dense_rank() over (partition by a order by a desc) from t limit 3
+----
+CompoundBlockOperator(Project) × 1 processor
+  LimitTransform × 1 processor
+    Transform Window × 1 processor
+      Merge (TransformSortMergeLimit × 4 processors) to (Transform Window × 1)
+        TransformSortMergeLimit × 4 processors
+          SortPartialTransform × 4 processors
+            Merge (DeserializeDataTransform × 1 processor) to (SortPartialTransform × 4)
+              DeserializeDataTransform × 1 processor
+                SyncReadParquetDataSource × 1 processor
+
+# rows frame single window (can push down limit)
+query T
+explain pipeline select a, sum(a) over (partition by a order by a desc rows between unbounded preceding and current row) from t limit 3
+----
+CompoundBlockOperator(Project) × 1 processor
+  LimitTransform × 1 processor
+    Transform Window × 1 processor
+      Merge (TransformSortMergeLimit × 4 processors) to (Transform Window × 1)
+        TransformSortMergeLimit × 4 processors
+          SortPartialTransform × 4 processors
+            Merge (DeserializeDataTransform × 1 processor) to (SortPartialTransform × 4)
+              DeserializeDataTransform × 1 processor
+                SyncReadParquetDataSource × 1 processor
+
+# rows frame single window (can not push down limit)
+query T
+explain pipeline select a, sum(a) over (partition by a order by a desc rows between unbounded preceding and unbounded following) from t limit 3
+----
+CompoundBlockOperator(Project) × 1 processor
+  LimitTransform × 1 processor
+    Transform Window × 1 processor
+      Merge (TransformSortMerge × 4 processors) to (Transform Window × 1)
+        TransformSortMerge × 4 processors
+          SortPartialTransform × 4 processors
+            Merge (DeserializeDataTransform × 1 processor) to (SortPartialTransform × 4)
+              DeserializeDataTransform × 1 processor
+                SyncReadParquetDataSource × 1 processor
+
+# rows frame multi window (can not push down limit)
+query T
+explain pipeline select a, sum(a) over (partition by a order by a desc rows between unbounded preceding and current row),
+avg(a) over (order by a rows between unbounded preceding and current row) from t limit 3
+----
+CompoundBlockOperator(Project) × 1 processor
+  LimitTransform × 1 processor
+    Transform Window × 1 processor
+      Merge (TransformSortMerge × 4 processors) to (Transform Window × 1)
+        TransformSortMerge × 4 processors
+          SortPartialTransform × 4 processors
+            Merge (Transform Window × 1 processor) to (SortPartialTransform × 4)
+              Transform Window × 1 processor
+                Merge (TransformSortMerge × 4 processors) to (Transform Window × 1)
+                  TransformSortMerge × 4 processors
+                    SortPartialTransform × 4 processors
+                      Merge (DeserializeDataTransform × 1 processor) to (SortPartialTransform × 4)
+                        DeserializeDataTransform × 1 processor
+                          SyncReadParquetDataSource × 1 processor
+
 statement ok
 DROP DATABASE test_explain_window;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/bloom_filter.test b/tests/sqllogictests/suites/mode/standalone/explain_native/bloom_filter.test
index 575bd782b6f5..52d1cbc23ef0 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/bloom_filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/bloom_filter.test
@@ -1,4 +1,7 @@
 # This case depends on explain(standalone mode), thus we put it here
+statement ok
+drop table if exists bloom_test_t;
+
 statement ok
 create table bloom_test_t(c1 int, c2 int)
 
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/explain.test b/tests/sqllogictests/suites/mode/standalone/explain_native/explain.test
index e1b7b1e6de88..1e366ae86672 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/explain.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/explain.test
@@ -679,54 +679,50 @@ create table t3 as select number as a, number as b from numbers(10)
 query T
 explain select * from t1,t2, t3 where (t1.a > 1 and t2.a > 2) or (t1.b < 3 and t2.b < 4) or t3.a = 2
 ----
-Filter
+HashJoin
 ├── output columns: [t3.a (#4), t3.b (#5), t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
+├── join type: INNER
+├── build keys: []
+├── probe keys: []
 ├── filters: [t1.a (#0) > 1 AND t2.a (#2) > 2 OR t1.b (#1) < 3 AND t2.b (#3) < 4 OR t3.a (#4) = 2]
-├── estimated rows: 21.20
-└── HashJoin
-    ├── output columns: [t3.a (#4), t3.b (#5), t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
-    ├── join type: CROSS
-    ├── build keys: []
-    ├── probe keys: []
-    ├── filters: []
-    ├── estimated rows: 50.00
-    ├── HashJoin(Build)
-    │   ├── output columns: [t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
-    │   ├── join type: CROSS
-    │   ├── build keys: []
-    │   ├── probe keys: []
-    │   ├── filters: []
-    │   ├── estimated rows: 5.00
-    │   ├── TableScan(Build)
-    │   │   ├── table: default.default.t1
-    │   │   ├── output columns: [a (#0), b (#1)]
-    │   │   ├── read rows: 1
-    │   │   ├── read bytes: 36
-    │   │   ├── partitions total: 1
-    │   │   ├── partitions scanned: 1
-    │   │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    │   │   ├── push downs: [filters: [], limit: NONE]
-    │   │   └── estimated rows: 1.00
-    │   └── TableScan(Probe)
-    │       ├── table: default.default.t2
-    │       ├── output columns: [a (#2), b (#3)]
-    │       ├── read rows: 5
-    │       ├── read bytes: 68
-    │       ├── partitions total: 1
-    │       ├── partitions scanned: 1
-    │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    │       ├── push downs: [filters: [], limit: NONE]
-    │       └── estimated rows: 5.00
-    └── TableScan(Probe)
-        ├── table: default.default.t3
-        ├── output columns: [a (#4), b (#5)]
-        ├── read rows: 10
-        ├── read bytes: 108
-        ├── partitions total: 1
-        ├── partitions scanned: 1
-        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-        ├── push downs: [filters: [], limit: NONE]
-        └── estimated rows: 10.00
+├── estimated rows: 50.00
+├── HashJoin(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3), t1.a (#0), t1.b (#1)]
+│   ├── join type: CROSS
+│   ├── build keys: []
+│   ├── probe keys: []
+│   ├── filters: []
+│   ├── estimated rows: 5.00
+│   ├── TableScan(Build)
+│   │   ├── table: default.default.t1
+│   │   ├── output columns: [a (#0), b (#1)]
+│   │   ├── read rows: 1
+│   │   ├── read bytes: 36
+│   │   ├── partitions total: 1
+│   │   ├── partitions scanned: 1
+│   │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   │   ├── push downs: [filters: [], limit: NONE]
+│   │   └── estimated rows: 1.00
+│   └── TableScan(Probe)
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 5
+│       ├── read bytes: 68
+│       ├── partitions total: 1
+│       ├── partitions scanned: 1
+│       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│       ├── push downs: [filters: [], limit: NONE]
+│       └── estimated rows: 5.00
+└── TableScan(Probe)
+    ├── table: default.default.t3
+    ├── output columns: [a (#4), b (#5)]
+    ├── read rows: 10
+    ├── read bytes: 108
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 10.00
 
 query T
 explain select * from t1,t2, t3 where ((t1.a > 1 and t2.a > 2) or (t1.b < 3 and t2.b < 4)) and t3.a > 1
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/infer_filter.test b/tests/sqllogictests/suites/mode/standalone/explain_native/infer_filter.test
index d36017e381c5..f91b67c54fa9 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/infer_filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/infer_filter.test
@@ -462,7 +462,7 @@ TableScan
 ├── read bytes: 0
 ├── partitions total: 0
 ├── partitions scanned: 0
-├── push downs: [filters: [and_filters(t1.a (#0) > 10, t1.a (#0) <= 100)], limit: NONE]
+├── push downs: [filters: [and_filters(t1.a (#0) <= 100, t1.a (#0) > 10)], limit: NONE]
 └── estimated rows: 0.00
 
 # t1.a > 10 and t2.a > 10
@@ -500,46 +500,46 @@ query T
 explain select * from t1, t2, t3 where t1.a = t2.a and t1.a = t3.a and t1.a > 5 and t3.a < 10;
 ----
 HashJoin
-├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2), t3.b (#5), t3.a (#4)]
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
 ├── join type: INNER
 ├── build keys: [t3.a (#4)]
 ├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 0.00
-├── TableScan(Build)
-│   ├── table: default.default.t3
-│   ├── output columns: [a (#4), b (#5)]
-│   ├── read rows: 0
-│   ├── read bytes: 0
-│   ├── partitions total: 0
-│   ├── partitions scanned: 0
-│   ├── push downs: [filters: [and_filters(t3.a (#4) > 5, t3.a (#4) < 10)], limit: NONE]
-│   └── estimated rows: 0.00
-└── HashJoin(Probe)
-    ├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
-    ├── join type: INNER
-    ├── build keys: [t2.a (#2)]
-    ├── probe keys: [t1.a (#0)]
-    ├── filters: []
-    ├── estimated rows: 0.00
-    ├── TableScan(Build)
-    │   ├── table: default.default.t2
-    │   ├── output columns: [a (#2), b (#3)]
-    │   ├── read rows: 0
-    │   ├── read bytes: 0
-    │   ├── partitions total: 0
-    │   ├── partitions scanned: 0
-    │   ├── push downs: [filters: [and_filters(t2.a (#2) > 5, t2.a (#2) < 10)], limit: NONE]
-    │   └── estimated rows: 0.00
-    └── TableScan(Probe)
-        ├── table: default.default.t1
-        ├── output columns: [a (#0), b (#1)]
-        ├── read rows: 0
-        ├── read bytes: 0
-        ├── partitions total: 0
-        ├── partitions scanned: 0
-        ├── push downs: [filters: [and_filters(t1.a (#0) > 5, t1.a (#0) < 10)], limit: NONE]
-        └── estimated rows: 0.00
+├── HashJoin(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
+│   ├── join type: INNER
+│   ├── build keys: [t3.a (#4)]
+│   ├── probe keys: [t2.a (#2)]
+│   ├── filters: []
+│   ├── estimated rows: 0.00
+│   ├── TableScan(Build)
+│   │   ├── table: default.default.t3
+│   │   ├── output columns: [a (#4), b (#5)]
+│   │   ├── read rows: 0
+│   │   ├── read bytes: 0
+│   │   ├── partitions total: 0
+│   │   ├── partitions scanned: 0
+│   │   ├── push downs: [filters: [and_filters(t3.a (#4) > 5, t3.a (#4) < 10)], limit: NONE]
+│   │   └── estimated rows: 0.00
+│   └── TableScan(Probe)
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [and_filters(t2.a (#2) > 5, t2.a (#2) < 10)], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [and_filters(t1.a (#0) > 5, t1.a (#0) < 10)], limit: NONE]
+    └── estimated rows: 0.00
 
 # t1.a > 5 and t2.a > 10
 query T
@@ -675,6 +675,52 @@ HashJoin
         ├── push downs: [filters: [false], limit: NONE]
         └── estimated rows: 0.00
 
+# t1.a = t2.a, t1.a = t3.a => t2.a = t3.a
+query T
+explain select * from t1, t2, t3 where t1.a = t2.a and t1.a = t3.a;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
+├── join type: INNER
+├── build keys: [t3.a (#4)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.00
+├── HashJoin(Build)
+│   ├── output columns: [t2.a (#2), t2.b (#3), t3.b (#5), t3.a (#4)]
+│   ├── join type: INNER
+│   ├── build keys: [t3.a (#4)]
+│   ├── probe keys: [t2.a (#2)]
+│   ├── filters: []
+│   ├── estimated rows: 0.00
+│   ├── TableScan(Build)
+│   │   ├── table: default.default.t3
+│   │   ├── output columns: [a (#4), b (#5)]
+│   │   ├── read rows: 0
+│   │   ├── read bytes: 0
+│   │   ├── partitions total: 0
+│   │   ├── partitions scanned: 0
+│   │   ├── push downs: [filters: [], limit: NONE]
+│   │   └── estimated rows: 0.00
+│   └── TableScan(Probe)
+│       ├── table: default.default.t2
+│       ├── output columns: [a (#2), b (#3)]
+│       ├── read rows: 0
+│       ├── read bytes: 0
+│       ├── partitions total: 0
+│       ├── partitions scanned: 0
+│       ├── push downs: [filters: [], limit: NONE]
+│       └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 0.00
+
 statement ok
 drop table if exists t1;
 
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/join.test b/tests/sqllogictests/suites/mode/standalone/explain_native/join.test
index 408086894f64..6e9cab5f2f3d 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/join.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/join.test
@@ -75,8 +75,8 @@ HashJoin
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    ├── push downs: [filters: [t1.number (#1) = t1.number (#1) + 1], limit: NONE]
+    └── estimated rows: 2.00
 
 query T
 explain select t.number from t, t1 where t.number > 1 and 1 < t1.number
@@ -87,7 +87,7 @@ HashJoin
 ├── build keys: []
 ├── probe keys: []
 ├── filters: []
-├── estimated rows: 0.18
+├── estimated rows: 1.64
 ├── TableScan(Build)
 │   ├── table: default.default.t
 │   ├── output columns: [number (#0)]
@@ -106,8 +106,8 @@ HashJoin
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [1 < t1.number (#1)], limit: NONE]
-    └── estimated rows: 0.91
+    ├── push downs: [filters: [t1.number (#1) > 1], limit: NONE]
+    └── estimated rows: 8.18
 
 query T
 explain select t.number from t, t1 where t.number + t1.number = 1
@@ -347,32 +347,32 @@ query T
 explain select * from onecolumn as a right join twocolumn as b on a.x = b.x where b.x > 42 and b.x < 45
 ----
 HashJoin
-├── output columns: [a.x (#0), b.x (#1), b.y (#2)]
-├── join type: RIGHT OUTER
-├── build keys: [b.x (#1)]
-├── probe keys: [a.x (#0)]
+├── output columns: [b.x (#1), b.y (#2), a.x (#0)]
+├── join type: LEFT OUTER
+├── build keys: [a.x (#0)]
+├── probe keys: [b.x (#1)]
 ├── filters: []
 ├── estimated rows: 3.20
 ├── TableScan(Build)
-│   ├── table: default.default.twocolumn
-│   ├── output columns: [x (#1), y (#2)]
+│   ├── table: default.default.onecolumn
+│   ├── output columns: [x (#0)]
 │   ├── read rows: 4
-│   ├── read bytes: 62
+│   ├── read bytes: 29
 │   ├── partitions total: 1
 │   ├── partitions scanned: 1
 │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-│   ├── push downs: [filters: [and_filters(twocolumn.x (#1) > 42, twocolumn.x (#1) < 45)], limit: NONE]
-│   └── estimated rows: 3.20
+│   ├── push downs: [filters: [and_filters(onecolumn.x (#0) > 42, onecolumn.x (#0) < 45)], limit: NONE]
+│   └── estimated rows: 3.00
 └── TableScan(Probe)
-    ├── table: default.default.onecolumn
-    ├── output columns: [x (#0)]
+    ├── table: default.default.twocolumn
+    ├── output columns: [x (#1), y (#2)]
     ├── read rows: 4
-    ├── read bytes: 29
+    ├── read bytes: 62
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 4.00
+    ├── push downs: [filters: [and_filters(twocolumn.x (#1) > 42, twocolumn.x (#1) < 45)], limit: NONE]
+    └── estimated rows: 3.20
 
 statement ok
 drop table t
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/chain.test b/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/chain.test
index d68450cd8be9..1db3be1d2979 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/chain.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/chain.test
@@ -29,9 +29,9 @@ query T
 explain select * from t, t1, t2 where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#2), t.a (#0), t1.a (#1)]
+├── output columns: [t2.a (#2), t1.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t1.a (#1)]
+├── build keys: [t.a (#0)]
 ├── probe keys: [t2.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -77,17 +77,17 @@ query T
 explain select * from t, t2, t1 where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#1), t.a (#0), t1.a (#2)]
+├── output columns: [t1.a (#2), t2.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t1.a (#2)]
-├── probe keys: [t2.a (#1)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t1.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t1.a (#2), t.a (#0)]
+│   ├── output columns: [t2.a (#1), t.a (#0)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#0)]
-│   ├── probe keys: [t1.a (#2)]
+│   ├── probe keys: [t2.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -101,41 +101,41 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t1
-│       ├── output columns: [a (#2)]
-│       ├── read rows: 10
-│       ├── read bytes: 54
+│       ├── table: default.join_reorder.t2
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 100
+│       ├── read bytes: 414
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 10.00
+│       └── estimated rows: 100.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t2
-    ├── output columns: [a (#1)]
-    ├── read rows: 100
-    ├── read bytes: 414
+    ├── table: default.join_reorder.t1
+    ├── output columns: [a (#2)]
+    ├── read rows: 10
+    ├── read bytes: 54
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 100.00
+    └── estimated rows: 10.00
 
 query T
 explain select * from t1, t, t2 where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#2), t.a (#1), t1.a (#0)]
+├── output columns: [t1.a (#0), t.a (#1), t2.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#0)]
-├── probe keys: [t2.a (#2)]
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t1.a (#0), t.a (#1)]
+│   ├── output columns: [t2.a (#2), t.a (#1)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#1)]
-│   ├── probe keys: [t1.a (#0)]
+│   ├── probe keys: [t2.a (#2)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -149,41 +149,41 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t1
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 10
-│       ├── read bytes: 54
+│       ├── table: default.join_reorder.t2
+│       ├── output columns: [a (#2)]
+│       ├── read rows: 100
+│       ├── read bytes: 414
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 10.00
+│       └── estimated rows: 100.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t2
-    ├── output columns: [a (#2)]
-    ├── read rows: 100
-    ├── read bytes: 414
+    ├── table: default.join_reorder.t1
+    ├── output columns: [a (#0)]
+    ├── read rows: 10
+    ├── read bytes: 54
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 100.00
+    └── estimated rows: 10.00
 
 query T
 explain select * from t1, t2, t where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#1), t.a (#2), t1.a (#0)]
+├── output columns: [t1.a (#0), t2.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#0)]
-├── probe keys: [t2.a (#1)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t1.a (#0), t.a (#2)]
+│   ├── output columns: [t2.a (#1), t.a (#2)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#2)]
-│   ├── probe keys: [t1.a (#0)]
+│   ├── probe keys: [t2.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -197,33 +197,33 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t1
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 10
-│       ├── read bytes: 54
+│       ├── table: default.join_reorder.t2
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 100
+│       ├── read bytes: 414
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 10.00
+│       └── estimated rows: 100.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t2
-    ├── output columns: [a (#1)]
-    ├── read rows: 100
-    ├── read bytes: 414
+    ├── table: default.join_reorder.t1
+    ├── output columns: [a (#0)]
+    ├── read rows: 10
+    ├── read bytes: 54
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 100.00
+    └── estimated rows: 10.00
 
 query T
 explain select * from t2, t1, t where t.a = t1.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t2.a (#0), t.a (#2), t1.a (#1)]
+├── output columns: [t2.a (#0), t1.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#1)]
+├── build keys: [t.a (#2)]
 ├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -504,9 +504,9 @@ explain join select * from t right anti join t1 on t1.a = t.a
 ----
 HashJoin: LEFT ANTI
 ├── Build
-│   └── Scan: default.join_reorder.t (read rows: 1)
+│   └── Scan: default.join_reorder.t (#0) (read rows: 1)
 └── Probe
-    └── Scan: default.join_reorder.t1 (read rows: 10)
+    └── Scan: default.join_reorder.t1 (#1) (read rows: 10)
 
 statement ok
 drop database join_reorder
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/cycles.test b/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/cycles.test
index b53e40e1f7cf..cbb25b91d504 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/cycles.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/cycles.test
@@ -22,8 +22,8 @@ explain select * from t, t1, t2 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t2.a (#2), t1.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t.a (#0), t1.a (#1)]
-├── probe keys: [t2.a (#2), t2.a (#2)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t2.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -70,8 +70,8 @@ explain select * from t, t2, t1 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t1.a (#2), t2.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t.a (#0), t2.a (#1)]
-├── probe keys: [t1.a (#2), t1.a (#2)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t1.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -116,10 +116,10 @@ query T
 explain select * from t1, t, t2 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 ----
 HashJoin
-├── output columns: [t1.a (#0), t2.a (#2), t.a (#1)]
+├── output columns: [t1.a (#0), t.a (#1), t2.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#2), t.a (#1)]
-├── probe keys: [t1.a (#0), t1.a (#0)]
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -166,8 +166,8 @@ explain select * from t1, t2, t where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t1.a (#0), t2.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t.a (#2), t2.a (#1)]
-├── probe keys: [t1.a (#0), t1.a (#0)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -214,8 +214,8 @@ explain select * from t2, t1, t where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 HashJoin
 ├── output columns: [t2.a (#0), t1.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t.a (#2), t1.a (#1)]
-├── probe keys: [t2.a (#0), t2.a (#0)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
@@ -260,10 +260,10 @@ query T
 explain select * from t2, t, t1 where t.a = t1.a and t1.a = t2.a and t2.a = t.a
 ----
 HashJoin
-├── output columns: [t2.a (#0), t1.a (#2), t.a (#1)]
+├── output columns: [t2.a (#0), t.a (#1), t1.a (#2)]
 ├── join type: INNER
-├── build keys: [t1.a (#2), t.a (#1)]
-├── probe keys: [t2.a (#0), t2.a (#0)]
+├── build keys: [t1.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/star.test b/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/star.test
index 0a1574892fee..c45c0548f8ff 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/star.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/join_reorder/star.test
@@ -20,17 +20,17 @@ query T
 explain select * from t, t1, t2 where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#1), t.a (#0), t2.a (#2)]
+├── output columns: [t2.a (#2), t1.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t2.a (#2)]
-├── probe keys: [t1.a (#1)]
+├── build keys: [t.a (#0)]
+├── probe keys: [t2.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t2.a (#2), t.a (#0)]
+│   ├── output columns: [t1.a (#1), t.a (#0)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#0)]
-│   ├── probe keys: [t2.a (#2)]
+│   ├── probe keys: [t1.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -44,33 +44,33 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t2
-│       ├── output columns: [a (#2)]
-│       ├── read rows: 100
-│       ├── read bytes: 414
+│       ├── table: default.join_reorder.t1
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 10
+│       ├── read bytes: 54
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 100.00
+│       └── estimated rows: 10.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t1
-    ├── output columns: [a (#1)]
-    ├── read rows: 10
-    ├── read bytes: 54
+    ├── table: default.join_reorder.t2
+    ├── output columns: [a (#2)]
+    ├── read rows: 100
+    ├── read bytes: 414
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    └── estimated rows: 100.00
 
 query T
 explain select * from t, t2, t1 where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#2), t.a (#0), t2.a (#1)]
+├── output columns: [t1.a (#2), t2.a (#1), t.a (#0)]
 ├── join type: INNER
-├── build keys: [t2.a (#1)]
+├── build keys: [t.a (#0)]
 ├── probe keys: [t1.a (#2)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -164,9 +164,9 @@ query T
 explain select * from t1, t2, t where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#0), t.a (#2), t2.a (#1)]
+├── output columns: [t1.a (#0), t2.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#1)]
+├── build keys: [t.a (#2)]
 ├── probe keys: [t1.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
@@ -212,17 +212,17 @@ query T
 explain select * from t2, t1, t where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#1), t.a (#2), t2.a (#0)]
+├── output columns: [t2.a (#0), t1.a (#1), t.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#0)]
-├── probe keys: [t1.a (#1)]
+├── build keys: [t.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t2.a (#0), t.a (#2)]
+│   ├── output columns: [t1.a (#1), t.a (#2)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#2)]
-│   ├── probe keys: [t2.a (#0)]
+│   ├── probe keys: [t1.a (#1)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -236,41 +236,41 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t2
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 100
-│       ├── read bytes: 414
+│       ├── table: default.join_reorder.t1
+│       ├── output columns: [a (#1)]
+│       ├── read rows: 10
+│       ├── read bytes: 54
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 100.00
+│       └── estimated rows: 10.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t1
-    ├── output columns: [a (#1)]
-    ├── read rows: 10
-    ├── read bytes: 54
+    ├── table: default.join_reorder.t2
+    ├── output columns: [a (#0)]
+    ├── read rows: 100
+    ├── read bytes: 414
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    └── estimated rows: 100.00
 
 query T
 explain select * from t2, t, t1 where t.a = t2.a and t1.a = t2.a
 ----
 HashJoin
-├── output columns: [t1.a (#2), t.a (#1), t2.a (#0)]
+├── output columns: [t2.a (#0), t.a (#1), t1.a (#2)]
 ├── join type: INNER
-├── build keys: [t2.a (#0)]
-├── probe keys: [t1.a (#2)]
+├── build keys: [t1.a (#2)]
+├── probe keys: [t2.a (#0)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── HashJoin(Build)
-│   ├── output columns: [t2.a (#0), t.a (#1)]
+│   ├── output columns: [t1.a (#2), t.a (#1)]
 │   ├── join type: INNER
 │   ├── build keys: [t.a (#1)]
-│   ├── probe keys: [t2.a (#0)]
+│   ├── probe keys: [t1.a (#2)]
 │   ├── filters: []
 │   ├── estimated rows: 1.00
 │   ├── TableScan(Build)
@@ -284,25 +284,25 @@ HashJoin
 │   │   ├── push downs: [filters: [], limit: NONE]
 │   │   └── estimated rows: 1.00
 │   └── TableScan(Probe)
-│       ├── table: default.join_reorder.t2
-│       ├── output columns: [a (#0)]
-│       ├── read rows: 100
-│       ├── read bytes: 414
+│       ├── table: default.join_reorder.t1
+│       ├── output columns: [a (#2)]
+│       ├── read rows: 10
+│       ├── read bytes: 54
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
 │       ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
 │       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 100.00
+│       └── estimated rows: 10.00
 └── TableScan(Probe)
-    ├── table: default.join_reorder.t1
-    ├── output columns: [a (#2)]
-    ├── read rows: 10
-    ├── read bytes: 54
+    ├── table: default.join_reorder.t2
+    ├── output columns: [a (#0)]
+    ├── read rows: 100
+    ├── read bytes: 414
     ├── partitions total: 1
     ├── partitions scanned: 1
     ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
     ├── push downs: [filters: [], limit: NONE]
-    └── estimated rows: 10.00
+    └── estimated rows: 100.00
 
 statement ok
 drop database join_reorder
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/limit.test b/tests/sqllogictests/suites/mode/standalone/explain_native/limit.test
index 12c1a26d4a41..c33a8410dee6 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/limit.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/limit.test
@@ -97,13 +97,13 @@ Limit
             ├── estimated rows: 0.20
             └── Filter
                 ├── output columns: [t.number (#0)]
-                ├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(CAST(is_not_null(scalar_subquery_4 (#4)) AS Boolean NULL), scalar_subquery_4 (#4), 0))]
+                ├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(true, TRY_CAST(scalar_subquery_4 (#4) AS UInt64 NULL), 0))]
                 ├── estimated rows: 0.20
                 └── HashJoin
                     ├── output columns: [t.number (#0), COUNT(*) (#4)]
-                    ├── join type: LEFT SINGLE
+                    ├── join type: INNER
                     ├── build keys: [number (#2)]
-                    ├── probe keys: [CAST(number (#0) AS UInt64 NULL)]
+                    ├── probe keys: [number (#0)]
                     ├── filters: []
                     ├── estimated rows: 1.00
                     ├── AggregateFinal(Build)
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/prune_column.test b/tests/sqllogictests/suites/mode/standalone/explain_native/prune_column.test
index a8571ac69ad6..9c6f8a75331f 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/prune_column.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/prune_column.test
@@ -117,90 +117,90 @@ explain select t1.a from (select number + 1 as a, number + 1 as b from numbers(1
 HashJoin
 ├── output columns: [a (#1)]
 ├── join type: INNER
-├── build keys: [CAST(t1.a (#1) AS UInt64 NULL)]
-├── probe keys: [_if_scalar_subquery (#15)]
+├── build keys: [_if_scalar_subquery (#15)]
+├── probe keys: [CAST(t1.a (#1) AS UInt64 NULL)]
 ├── filters: []
 ├── estimated rows: 1.00
 ├── EvalScalar(Build)
-│   ├── output columns: [a (#1)]
-│   ├── expressions: [numbers.number (#0) + 1]
+│   ├── output columns: [_if_scalar_subquery (#15)]
+│   ├── expressions: [if(CAST(_count_scalar_subquery (#13) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#14))]
 │   ├── estimated rows: 1.00
-│   └── TableScan
-│       ├── table: default.system.numbers
-│       ├── output columns: [number (#0)]
-│       ├── read rows: 1
-│       ├── read bytes: 8
-│       ├── partitions total: 1
-│       ├── partitions scanned: 1
-│       ├── push downs: [filters: [], limit: NONE]
-│       └── estimated rows: 1.00
+│   └── Limit
+│       ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
+│       ├── limit: 1
+│       ├── offset: 0
+│       ├── estimated rows: 1.00
+│       └── AggregateFinal
+│           ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
+│           ├── group by: []
+│           ├── aggregate functions: [count(), any(COUNT(*))]
+│           ├── limit: 1
+│           ├── estimated rows: 1.00
+│           └── AggregatePartial
+│               ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
+│               ├── group by: []
+│               ├── aggregate functions: [count(), any(COUNT(*))]
+│               ├── estimated rows: 1.00
+│               └── AggregateFinal
+│                   ├── output columns: [COUNT(*) (#12)]
+│                   ├── group by: []
+│                   ├── aggregate functions: [count()]
+│                   ├── estimated rows: 1.00
+│                   └── AggregatePartial
+│                       ├── output columns: [COUNT(*) (#12)]
+│                       ├── group by: []
+│                       ├── aggregate functions: [count()]
+│                       ├── estimated rows: 1.00
+│                       └── HashJoin
+│                           ├── output columns: []
+│                           ├── join type: INNER
+│                           ├── build keys: [t2.b (#5)]
+│                           ├── probe keys: [t3.b (#10)]
+│                           ├── filters: []
+│                           ├── estimated rows: 0.20
+│                           ├── EvalScalar(Build)
+│                           │   ├── output columns: [b (#5)]
+│                           │   ├── expressions: [numbers.number (#3) + 1]
+│                           │   ├── estimated rows: 0.20
+│                           │   └── Filter
+│                           │       ├── output columns: [numbers.number (#3)]
+│                           │       ├── filters: [numbers.number (#3) + 1 = 1]
+│                           │       ├── estimated rows: 0.20
+│                           │       └── TableScan
+│                           │           ├── table: default.system.numbers
+│                           │           ├── output columns: [number (#3)]
+│                           │           ├── read rows: 1
+│                           │           ├── read bytes: 8
+│                           │           ├── partitions total: 1
+│                           │           ├── partitions scanned: 1
+│                           │           ├── push downs: [filters: [numbers.number (#3) + 1 = 1], limit: NONE]
+│                           │           └── estimated rows: 1.00
+│                           └── EvalScalar(Probe)
+│                               ├── output columns: [b (#10)]
+│                               ├── expressions: [numbers.number (#8) + 1]
+│                               ├── estimated rows: 1.00
+│                               └── TableScan
+│                                   ├── table: default.system.numbers
+│                                   ├── output columns: [number (#8)]
+│                                   ├── read rows: 1
+│                                   ├── read bytes: 8
+│                                   ├── partitions total: 1
+│                                   ├── partitions scanned: 1
+│                                   ├── push downs: [filters: [], limit: NONE]
+│                                   └── estimated rows: 1.00
 └── EvalScalar(Probe)
-    ├── output columns: [_if_scalar_subquery (#15)]
-    ├── expressions: [if(CAST(_count_scalar_subquery (#13) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#14))]
+    ├── output columns: [a (#1)]
+    ├── expressions: [numbers.number (#0) + 1]
     ├── estimated rows: 1.00
-    └── Limit
-        ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
-        ├── limit: 1
-        ├── offset: 0
-        ├── estimated rows: 1.00
-        └── AggregateFinal
-            ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
-            ├── group by: []
-            ├── aggregate functions: [count(), any(COUNT(*))]
-            ├── limit: 1
-            ├── estimated rows: 1.00
-            └── AggregatePartial
-                ├── output columns: [_count_scalar_subquery (#13), _any_scalar_subquery (#14)]
-                ├── group by: []
-                ├── aggregate functions: [count(), any(COUNT(*))]
-                ├── estimated rows: 1.00
-                └── AggregateFinal
-                    ├── output columns: [COUNT(*) (#12)]
-                    ├── group by: []
-                    ├── aggregate functions: [count()]
-                    ├── estimated rows: 1.00
-                    └── AggregatePartial
-                        ├── output columns: [COUNT(*) (#12)]
-                        ├── group by: []
-                        ├── aggregate functions: [count()]
-                        ├── estimated rows: 1.00
-                        └── HashJoin
-                            ├── output columns: []
-                            ├── join type: INNER
-                            ├── build keys: [t2.b (#5)]
-                            ├── probe keys: [t3.b (#10)]
-                            ├── filters: []
-                            ├── estimated rows: 0.20
-                            ├── EvalScalar(Build)
-                            │   ├── output columns: [b (#5)]
-                            │   ├── expressions: [numbers.number (#3) + 1]
-                            │   ├── estimated rows: 0.20
-                            │   └── Filter
-                            │       ├── output columns: [numbers.number (#3)]
-                            │       ├── filters: [numbers.number (#3) + 1 = 1]
-                            │       ├── estimated rows: 0.20
-                            │       └── TableScan
-                            │           ├── table: default.system.numbers
-                            │           ├── output columns: [number (#3)]
-                            │           ├── read rows: 1
-                            │           ├── read bytes: 8
-                            │           ├── partitions total: 1
-                            │           ├── partitions scanned: 1
-                            │           ├── push downs: [filters: [numbers.number (#3) + 1 = 1], limit: NONE]
-                            │           └── estimated rows: 1.00
-                            └── EvalScalar(Probe)
-                                ├── output columns: [b (#10)]
-                                ├── expressions: [numbers.number (#8) + 1]
-                                ├── estimated rows: 1.00
-                                └── TableScan
-                                    ├── table: default.system.numbers
-                                    ├── output columns: [number (#8)]
-                                    ├── read rows: 1
-                                    ├── read bytes: 8
-                                    ├── partitions total: 1
-                                    ├── partitions scanned: 1
-                                    ├── push downs: [filters: [], limit: NONE]
-                                    └── estimated rows: 1.00
+    └── TableScan
+        ├── table: default.system.numbers
+        ├── output columns: [number (#0)]
+        ├── read rows: 1
+        ├── read bytes: 8
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 1.00
 
 query T
 explain select name from system.functions order by example
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_eval_scalar.test
similarity index 55%
rename from tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter.test
rename to tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_eval_scalar.test
index 4f3cd9bdc171..8aff143dfa86 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_eval_scalar.test
@@ -1,75 +1,3 @@
-# push down filter ProjectSet
-statement ok
-drop table if exists products;
-
-statement ok
-create table products(name varchar, details variant);
-
-statement ok
-insert into products (name, details) values ('Laptop', '{"brand": "Dell", "colors": ["Black", "Silver"], "price": 1200, "features": {"ram": "16GB", "storage": "512GB"}}'), ('Smartphone', '{"brand": "Apple", "colors": ["White", "Black"], "price": 999, "features": {"ram": "4GB", "storage": "128GB"}}'), ('Headphones', '{"brand": "Sony", "colors": ["Black", "Blue", "Red"], "price": 150, "features": {"battery": "20h", "bluetooth": "5.0"}}');
-
-query T
-explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
-----
-EvalScalar
-├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
-├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
-├── estimated rows: 0.12
-└── Filter
-    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
-    ├── filters: [is_true(TRY_CAST(get(1)(json_path_query (#2)) AS String NULL) = '512GB')]
-    ├── estimated rows: 0.12
-    └── ProjectSet
-        ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
-        ├── estimated rows: 0.60
-        ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
-        └── TableScan
-            ├── table: default.default.products
-            ├── output columns: [name (#0), details (#1)]
-            ├── read rows: 3
-            ├── read bytes: 370
-            ├── partitions total: 1
-            ├── partitions scanned: 1
-            ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
-            ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
-            └── estimated rows: 0.60
-
-query T
-select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
-----
-Laptop "512GB" "16GB"
-
-query T
-explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
-----
-EvalScalar
-├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
-├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
-├── estimated rows: 0.60
-└── ProjectSet
-    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
-    ├── estimated rows: 0.60
-    ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
-    └── TableScan
-        ├── table: default.default.products
-        ├── output columns: [name (#0), details (#1)]
-        ├── read rows: 3
-        ├── read bytes: 370
-        ├── partitions total: 1
-        ├── partitions scanned: 1
-        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
-        ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
-        └── estimated rows: 0.60
-
-query T
-select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
-----
-Laptop "16GB" "16GB"
-Laptop "512GB" "16GB"
-
-statement ok
-drop table products;
-
 # push down filter EvalScalar
 statement ok
 drop table if exists t1;
@@ -178,7 +106,7 @@ AggregateFinal
         │               │               ├── read bytes: 0
         │               │               ├── partitions total: 0
         │               │               ├── partitions scanned: 0
-        │               │               ├── push downs: [filters: [], limit: NONE]
+        │               │               ├── push downs: [filters: [is_true(t2.sid (#1) = 1)], limit: NONE]
         │               │               └── estimated rows: 0.00
         │               └── TableScan(Probe)
         │                   ├── table: default.default.t1
@@ -218,44 +146,13 @@ AggregateFinal
                             └── estimated rows: 0.00
 
 statement ok
-drop table t1;
-
-statement ok
-drop table t2;
-
-statement ok
-drop view v1;
-
-statement ok
-drop view v2;
-
-# push down alias filter scan
-statement ok
-drop table if exists t;
+drop table if exists t1;
 
 statement ok
-create table t (x INT);
+drop table if exists t2;
 
 statement ok
-insert into t(x) values (1), (2);
-
-query I
-explain select * from t as a(id) where a.id > 1;
-----
-Filter
-├── output columns: [a.x (#0)]
-├── filters: [is_true(a.id (#0) > 1)]
-├── estimated rows: 0.40
-└── TableScan
-    ├── table: default.default.t
-    ├── output columns: [x (#0)]
-    ├── read rows: 2
-    ├── read bytes: 24
-    ├── partitions total: 1
-    ├── partitions scanned: 1
-    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
-    ├── push downs: [filters: [is_true(t.x (#0) > 1)], limit: NONE]
-    └── estimated rows: 2.00
+drop view if exists v1;
 
 statement ok
-drop table t;
+drop view if exists v2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_full_outer.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_full_outer.test
new file mode 100644
index 000000000000..a669551bcbe4
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_full_outer.test
@@ -0,0 +1,248 @@
+# push down filter full outer join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# convert full outer join to inner join, can propagate t1.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t1.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert full outer join to inner join, can propagate t2.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t2.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: RIGHT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert full outer join to left outer join, can not propagate t1.b > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t1.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.b (#1) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert full outer join to right outer join, can not propagate t2.b > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t2.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: RIGHT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.b (#3) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert full outer join to inner join, can propagate t1.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t1.a > 0 and t2.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [and_filters(t2.a (#2) > 0, t2.b (#3) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert full outer join to inner join, can propagate t2.a > 0
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a where t2.a > 0 and t1.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [and_filters(t1.a (#0) > 0, t1.b (#1) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# full outer join, can not push down
+query T
+explain select * from t1 full outer join t2 on t1.a = t2.a and t2.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: FULL OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: [t2.a (#2) > 0]
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_inner.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_inner.test
new file mode 100644
index 000000000000..7008a66cbe7f
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_inner.test
@@ -0,0 +1,92 @@
+# push down filter inner join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# can propagate t1.a > 3
+query T
+explain select * from t1 inner join t2 on t1.a = t2.a where t1.a > 3;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 0
+│   ├── read bytes: 0
+│   ├── partitions total: 1
+│   ├── partitions scanned: 0
+│   ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 3)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 1
+    ├── partitions scanned: 0
+    ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 3)], limit: NONE]
+    └── estimated rows: 1.00
+
+# can propagate (t2.a > 1 or t2.a <= 2)
+query T
+explain select * from t1 inner join t2 on t1.a = t2.a where t2.a <= 2 or (t1.a > 1 and t2.a > 1);
+----
+Filter
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── filters: [is_true(t2.a (#2) <= 2 OR t1.a (#0) > 1 AND t2.a (#2) > 1)]
+├── estimated rows: 0.45
+└── HashJoin
+    ├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+    ├── join type: INNER
+    ├── build keys: [t2.a (#2)]
+    ├── probe keys: [t1.a (#0)]
+    ├── filters: []
+    ├── estimated rows: 1.26
+    ├── TableScan(Build)
+    │   ├── table: default.default.t2
+    │   ├── output columns: [a (#2), b (#3)]
+    │   ├── read rows: 3
+    │   ├── read bytes: 56
+    │   ├── partitions total: 1
+    │   ├── partitions scanned: 1
+    │   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    │   ├── push downs: [filters: [is_true(t2.a (#2) <= 2 OR t2.a (#2) > 1)], limit: NONE]
+    │   └── estimated rows: 1.08
+    └── TableScan(Probe)
+        ├── table: default.default.t1
+        ├── output columns: [a (#0), b (#1)]
+        ├── read rows: 4
+        ├── read bytes: 66
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+        ├── push downs: [filters: [is_true(t1.a (#0) <= 2 OR t1.a (#0) > 1)], limit: NONE]
+        └── estimated rows: 3.50
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_left_outer.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_left_outer.test
new file mode 100644
index 000000000000..436b194698f7
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_left_outer.test
@@ -0,0 +1,232 @@
+# push down filter left outer join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# left outer join, can propagate t1.a > 0 and convert it to inner join
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t1.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# left outer join, can not propagate t1.a > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a and t1.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: [t1.a (#0) > 0]
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+# left outer join, can not propagate t1.b > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t1.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.a (#2), t2.b (#3)]
+├── join type: LEFT OUTER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 4.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 3.00
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.b (#1) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert left outer join to inner join, can propagate t2.a > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t2.a > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 0)], limit: NONE]
+    └── estimated rows: 4.00
+
+# convert left outer join to inner join, can not propagate t2.b > 0
+query T
+explain select * from t1 left join t2 on t1.a = t2.a where t2.b > 0;
+----
+HashJoin
+├── output columns: [t1.a (#0), t1.b (#1), t2.b (#3), t2.a (#2)]
+├── join type: INNER
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.80
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2), b (#3)]
+│   ├── read rows: 3
+│   ├── read bytes: 56
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.b (#3) > 0)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0), b (#1)]
+    ├── read rows: 4
+    ├── read bytes: 66
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 4.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+drop table if exists m1;
+
+statement ok
+drop table if exists j1;
+
+statement ok
+create table m1(id varchar, "context" varchar);
+
+statement ok
+create table j1(id varchar);
+
+# In `can_filter_null`, if the function is `assume_not_null` or `remove_nullable`, we cannot replace the column bindings with `Scalar::Null`.
+query T
+explain WITH base AS (SELECT id, context FROM m1), src1 AS (SELECT base.id FROM base WHERE IFNULL(base.context, '') = ''), join1 AS (SELECT id FROM j1) SELECT src1.id FROM src1 LEFT OUTER JOIN join1 ON TRUE;
+----
+HashJoin
+├── output columns: [m1.id (#0)]
+├── join type: LEFT OUTER
+├── build keys: []
+├── probe keys: []
+├── filters: []
+├── estimated rows: 0.00
+├── TableScan(Build)
+│   ├── table: default.default.j1
+│   ├── output columns: []
+│   ├── read rows: 0
+│   ├── read size: 0
+│   ├── partitions total: 0
+│   ├── partitions scanned: 0
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 0.00
+└── TableScan(Probe)
+    ├── table: default.default.m1
+    ├── output columns: [id (#0)]
+    ├── read rows: 0
+    ├── read size: 0
+    ├── partitions total: 0
+    ├── partitions scanned: 0
+    ├── push downs: [filters: [is_true(if(CAST(is_not_null(m1.context (#1)) AS Boolean NULL), CAST(assume_not_null(m1.context (#1)) AS String NULL), true, '', NULL) = '')], limit: NONE]
+    └── estimated rows: 0.00
+
+statement ok
+drop table if exists m1;
+
+statement ok
+drop table if exists j1;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_semi_anti.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_semi_anti.test
new file mode 100644
index 000000000000..9e30d7cc5a57
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_join/push_down_filter_join_semi_anti.test
@@ -0,0 +1,88 @@
+# push down filter semi join
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
+
+statement ok
+create table t1(a int, b int);
+
+statement ok
+create table t2(a int, b int);
+
+statement ok
+insert into t1 values(null, null), (1, 1), (2, 2), (3, 3);
+
+statement ok
+insert into t2 values(null, null), (1, 1), (2, 2);
+
+# left semi, can propagate t1.a > 3
+query T
+explain select t1.a from t1 where exists (select * from t2 where t1.a = t2.a) and t1.a > 3;
+----
+HashJoin
+├── output columns: [t1.a (#0)]
+├── join type: LEFT SEMI
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 0.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2)]
+│   ├── read rows: 0
+│   ├── read bytes: 0
+│   ├── partitions total: 1
+│   ├── partitions scanned: 0
+│   ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 3)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 1
+    ├── partitions scanned: 0
+    ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 3)], limit: NONE]
+    └── estimated rows: 1.00
+
+# left anti, can propagate t1.a > 3
+query T
+explain select t1.a from t1 where not exists (select * from t2 where t1.a = t2.a) and t1.a > 3;
+----
+HashJoin
+├── output columns: [t1.a (#0)]
+├── join type: LEFT ANTI
+├── build keys: [t2.a (#2)]
+├── probe keys: [t1.a (#0)]
+├── filters: []
+├── estimated rows: 1.00
+├── TableScan(Build)
+│   ├── table: default.default.t2
+│   ├── output columns: [a (#2)]
+│   ├── read rows: 0
+│   ├── read bytes: 0
+│   ├── partitions total: 1
+│   ├── partitions scanned: 0
+│   ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+│   ├── push downs: [filters: [is_true(t2.a (#2) > 3)], limit: NONE]
+│   └── estimated rows: 0.60
+└── TableScan(Probe)
+    ├── table: default.default.t1
+    ├── output columns: [a (#0)]
+    ├── read rows: 0
+    ├── read bytes: 0
+    ├── partitions total: 1
+    ├── partitions scanned: 0
+    ├── pruning stats: [segments: <range pruning: 1 to 0>, blocks: <range pruning: 0 to 0, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t1.a (#0) > 3)], limit: NONE]
+    └── estimated rows: 1.00
+
+statement ok
+drop table if exists t1;
+
+statement ok
+drop table if exists t2;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_project_set.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_project_set.test
new file mode 100644
index 000000000000..b9e52300e2a4
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_project_set.test
@@ -0,0 +1,71 @@
+# push down filter ProjectSet
+statement ok
+drop table if exists products;
+
+statement ok
+create table products(name varchar, details variant);
+
+statement ok
+insert into products(name, details) values ('Laptop', '{"brand": "Dell", "colors": ["Black", "Silver"], "price": 1200, "features": {"ram": "16GB", "storage": "512GB"}}'), ('Smartphone', '{"brand": "Apple", "colors": ["White", "Black"], "price": 999, "features": {"ram": "4GB", "storage": "128GB"}}'), ('Headphones', '{"brand": "Sony", "colors": ["Black", "Blue", "Red"], "price": 150, "features": {"battery": "20h", "bluetooth": "5.0"}}');
+
+query T
+explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
+----
+EvalScalar
+├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
+├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
+├── estimated rows: 0.12
+└── Filter
+    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
+    ├── filters: [is_true(TRY_CAST(get(1)(json_path_query (#2)) AS String NULL) = '512GB')]
+    ├── estimated rows: 0.12
+    └── ProjectSet
+        ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
+        ├── estimated rows: 0.60
+        ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
+        └── TableScan
+            ├── table: default.default.products
+            ├── output columns: [name (#0), details (#1)]
+            ├── read rows: 3
+            ├── read bytes: 370
+            ├── partitions total: 1
+            ├── partitions scanned: 1
+            ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
+            ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
+            └── estimated rows: 0.60
+
+query T
+select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB' and all_features = '512GB';
+----
+Laptop "512GB" "16GB"
+
+query T
+explain select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
+----
+EvalScalar
+├── output columns: [products.name (#0), all_features (#3), first_feature (#4)]
+├── expressions: [get(1)(json_path_query (#2)), json_path_query_first(products.details (#1), '$.features.*')]
+├── estimated rows: 0.60
+└── ProjectSet
+    ├── output columns: [products.name (#0), products.details (#1), json_path_query (#2)]
+    ├── estimated rows: 0.60
+    ├── set returning functions: json_path_query(products.details (#1), '$.features.*')
+    └── TableScan
+        ├── table: default.default.products
+        ├── output columns: [name (#0), details (#1)]
+        ├── read rows: 3
+        ├── read bytes: 370
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 1 to 1>]
+        ├── push downs: [filters: [and_filters(products.name (#0) = 'Laptop', TRY_CAST(json_path_query_first(products.details (#1), '$.features.*') AS String NULL) = '16GB')], limit: NONE]
+        └── estimated rows: 0.60
+
+query T
+select name, json_path_query(details, '$.features.*') as all_features, json_path_query_first(details, '$.features.*') as first_feature from products where name = 'Laptop' and first_feature = '16GB';
+----
+Laptop "16GB" "16GB"
+Laptop "512GB" "16GB"
+
+statement ok
+drop table if exists products;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_scan.test b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_scan.test
new file mode 100644
index 000000000000..b7cfcc2c15b1
--- /dev/null
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/push_down_filter/push_down_filter_scan.test
@@ -0,0 +1,30 @@
+# push down alias filter scan
+statement ok
+drop table if exists t;
+
+statement ok
+create table t (x INT);
+
+statement ok
+insert into t(x) values (1), (2);
+
+query I
+explain select * from t as a(id) where a.id > 1;
+----
+Filter
+├── output columns: [a.x (#0)]
+├── filters: [is_true(a.id (#0) > 1)]
+├── estimated rows: 0.40
+└── TableScan
+    ├── table: default.default.t
+    ├── output columns: [x (#0)]
+    ├── read rows: 2
+    ├── read bytes: 24
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1, bloom pruning: 0 to 0>]
+    ├── push downs: [filters: [is_true(t.x (#0) > 1)], limit: NONE]
+    └── estimated rows: 2.00
+
+statement ok
+drop table if exists t;
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/select.test b/tests/sqllogictests/suites/mode/standalone/explain_native/select.test
index f95415b3ba9b..f74b8aeaea9e 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/select.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/select.test
@@ -16,7 +16,7 @@ explain select * from (select * from numbers(1)) as t1 where number = 1
 ----
 Filter
 ├── output columns: [numbers.number (#0)]
-├── filters: [numbers.number (#0) = 1]
+├── filters: [t1.number (#0) = 1]
 ├── estimated rows: 0.00
 └── TableScan
     ├── table: default.system.numbers
diff --git a/tests/sqllogictests/suites/mode/standalone/explain_native/subquery.test b/tests/sqllogictests/suites/mode/standalone/explain_native/subquery.test
index fdabb1d44936..f0181b37bd1e 100644
--- a/tests/sqllogictests/suites/mode/standalone/explain_native/subquery.test
+++ b/tests/sqllogictests/suites/mode/standalone/explain_native/subquery.test
@@ -3,13 +3,13 @@ explain select t.number from numbers(1) as t, numbers(1) as t1 where t.number =
 ----
 Filter
 ├── output columns: [t.number (#0)]
-├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(CAST(is_not_null(scalar_subquery_4 (#4)) AS Boolean NULL), scalar_subquery_4 (#4), 0))]
+├── filters: [is_true(CAST(t.number (#0) AS UInt64 NULL) = if(true, TRY_CAST(scalar_subquery_4 (#4) AS UInt64 NULL), 0))]
 ├── estimated rows: 0.20
 └── HashJoin
     ├── output columns: [t.number (#0), COUNT(*) (#4)]
-    ├── join type: LEFT SINGLE
+    ├── join type: INNER
     ├── build keys: [number (#2)]
-    ├── probe keys: [CAST(number (#0) AS UInt64 NULL)]
+    ├── probe keys: [number (#0)]
     ├── filters: []
     ├── estimated rows: 1.00
     ├── AggregateFinal(Build)
@@ -164,52 +164,52 @@ explain select t.number from numbers(1) as t where number = (select * from numbe
 HashJoin
 ├── output columns: [t.number (#0)]
 ├── join type: INNER
-├── build keys: [CAST(t.number (#0) AS UInt64 NULL)]
-├── probe keys: [_if_scalar_subquery (#4)]
+├── build keys: [_if_scalar_subquery (#4)]
+├── probe keys: [CAST(t.number (#0) AS UInt64 NULL)]
 ├── filters: []
 ├── estimated rows: 1.00
-├── TableScan(Build)
-│   ├── table: default.system.numbers
-│   ├── output columns: [number (#0)]
-│   ├── read rows: 1
-│   ├── read bytes: 8
-│   ├── partitions total: 1
-│   ├── partitions scanned: 1
-│   ├── push downs: [filters: [], limit: NONE]
-│   └── estimated rows: 1.00
-└── EvalScalar(Probe)
-    ├── output columns: [_if_scalar_subquery (#4)]
-    ├── expressions: [if(CAST(_count_scalar_subquery (#2) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#3))]
-    ├── estimated rows: 1.00
-    └── Limit
-        ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
-        ├── limit: 1
-        ├── offset: 0
-        ├── estimated rows: 1.00
-        └── AggregateFinal
-            ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
-            ├── group by: []
-            ├── aggregate functions: [count(), any(number)]
-            ├── limit: 1
-            ├── estimated rows: 1.00
-            └── AggregatePartial
-                ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
-                ├── group by: []
-                ├── aggregate functions: [count(), any(number)]
-                ├── estimated rows: 1.00
-                └── Filter
-                    ├── output columns: [numbers.number (#1)]
-                    ├── filters: [numbers.number (#1) = 0]
-                    ├── estimated rows: 0.00
-                    └── TableScan
-                        ├── table: default.system.numbers
-                        ├── output columns: [number (#1)]
-                        ├── read rows: 1
-                        ├── read bytes: 8
-                        ├── partitions total: 1
-                        ├── partitions scanned: 1
-                        ├── push downs: [filters: [numbers.number (#1) = 0], limit: NONE]
-                        └── estimated rows: 1.00
+├── EvalScalar(Build)
+│   ├── output columns: [_if_scalar_subquery (#4)]
+│   ├── expressions: [if(CAST(_count_scalar_subquery (#2) = 0 AS Boolean NULL), NULL, _any_scalar_subquery (#3))]
+│   ├── estimated rows: 1.00
+│   └── Limit
+│       ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
+│       ├── limit: 1
+│       ├── offset: 0
+│       ├── estimated rows: 1.00
+│       └── AggregateFinal
+│           ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
+│           ├── group by: []
+│           ├── aggregate functions: [count(), any(number)]
+│           ├── limit: 1
+│           ├── estimated rows: 1.00
+│           └── AggregatePartial
+│               ├── output columns: [_count_scalar_subquery (#2), _any_scalar_subquery (#3)]
+│               ├── group by: []
+│               ├── aggregate functions: [count(), any(number)]
+│               ├── estimated rows: 1.00
+│               └── Filter
+│                   ├── output columns: [numbers.number (#1)]
+│                   ├── filters: [numbers.number (#1) = 0]
+│                   ├── estimated rows: 0.00
+│                   └── TableScan
+│                       ├── table: default.system.numbers
+│                       ├── output columns: [number (#1)]
+│                       ├── read rows: 1
+│                       ├── read bytes: 8
+│                       ├── partitions total: 1
+│                       ├── partitions scanned: 1
+│                       ├── push downs: [filters: [numbers.number (#1) = 0], limit: NONE]
+│                       └── estimated rows: 1.00
+└── TableScan(Probe)
+    ├── table: default.system.numbers
+    ├── output columns: [number (#0)]
+    ├── read rows: 1
+    ├── read bytes: 8
+    ├── partitions total: 1
+    ├── partitions scanned: 1
+    ├── push downs: [filters: [], limit: NONE]
+    └── estimated rows: 1.00
 
 query T
 explain select t.number from numbers(1) as t where exists (select * from numbers(1) where number = t.number)
@@ -310,7 +310,7 @@ HashJoin
 ├── estimated rows: 0.00
 ├── Filter(Build)
 │   ├── output columns: [numbers.number (#1)]
-│   ├── filters: [numbers.number (#1) = 0]
+│   ├── filters: [numbers.number (#1) < 10, numbers.number (#1) = 0]
 │   ├── estimated rows: 0.00
 │   └── TableScan
 │       ├── table: default.system.numbers
@@ -319,7 +319,7 @@ HashJoin
 │       ├── read bytes: 8
 │       ├── partitions total: 1
 │       ├── partitions scanned: 1
-│       ├── push downs: [filters: [numbers.number (#1) = 0], limit: NONE]
+│       ├── push downs: [filters: [and_filters(and_filters(numbers.number (#1) = 0, numbers.number (#1) < 10), numbers.number (#1) = 0)], limit: NONE]
 │       └── estimated rows: 1.00
 └── Filter(Probe)
     ├── output columns: [t.number (#0)]
@@ -396,65 +396,61 @@ HashJoin
 query T
 explain select t.number from numbers(1) as t, numbers(1) as t1 where (select count(*) = 1 from numbers(1) where t.number = number) and t.number = t1.number
 ----
-Filter
+HashJoin
 ├── output columns: [t.number (#0)]
-├── filters: [is_true(try_to_boolean(if(CAST(is_not_null(scalar_subquery_4 (#4)) AS Boolean NULL), TRY_CAST(scalar_subquery_4 (#4) AS UInt64 NULL), 0)))]
+├── join type: INNER
+├── build keys: [number (#2)]
+├── probe keys: [number (#0)]
+├── filters: []
 ├── estimated rows: 0.20
-└── HashJoin
-    ├── output columns: [t.number (#0), (count(*) = 1) (#4)]
-    ├── join type: LEFT SINGLE
-    ├── build keys: [number (#2)]
-    ├── probe keys: [CAST(number (#0) AS UInt64 NULL)]
+├── Filter(Build)
+│   ├── output columns: [numbers.number (#2)]
+│   ├── filters: [is_true(try_to_boolean(if(true, TRY_CAST(COUNT(*) (#3) = 1 AS UInt64 NULL), 0)))]
+│   ├── estimated rows: 0.20
+│   └── AggregateFinal
+│       ├── output columns: [COUNT(*) (#3), numbers.number (#2)]
+│       ├── group by: [number]
+│       ├── aggregate functions: [count()]
+│       ├── estimated rows: 1.00
+│       └── AggregatePartial
+│           ├── output columns: [COUNT(*) (#3), #_group_by_key]
+│           ├── group by: [number]
+│           ├── aggregate functions: [count()]
+│           ├── estimated rows: 1.00
+│           └── TableScan
+│               ├── table: default.system.numbers
+│               ├── output columns: [number (#2)]
+│               ├── read rows: 1
+│               ├── read bytes: 8
+│               ├── partitions total: 1
+│               ├── partitions scanned: 1
+│               ├── push downs: [filters: [], limit: NONE]
+│               └── estimated rows: 1.00
+└── HashJoin(Probe)
+    ├── output columns: [t.number (#0)]
+    ├── join type: INNER
+    ├── build keys: [t1.number (#1)]
+    ├── probe keys: [t.number (#0)]
     ├── filters: []
     ├── estimated rows: 1.00
-    ├── EvalScalar(Build)
-    │   ├── output columns: [numbers.number (#2), (count(*) = 1) (#4)]
-    │   ├── expressions: [COUNT(*) (#3) = 1]
-    │   ├── estimated rows: 1.00
-    │   └── AggregateFinal
-    │       ├── output columns: [COUNT(*) (#3), numbers.number (#2)]
-    │       ├── group by: [number]
-    │       ├── aggregate functions: [count()]
-    │       ├── estimated rows: 1.00
-    │       └── AggregatePartial
-    │           ├── output columns: [COUNT(*) (#3), #_group_by_key]
-    │           ├── group by: [number]
-    │           ├── aggregate functions: [count()]
-    │           ├── estimated rows: 1.00
-    │           └── TableScan
-    │               ├── table: default.system.numbers
-    │               ├── output columns: [number (#2)]
-    │               ├── read rows: 1
-    │               ├── read bytes: 8
-    │               ├── partitions total: 1
-    │               ├── partitions scanned: 1
-    │               ├── push downs: [filters: [], limit: NONE]
-    │               └── estimated rows: 1.00
-    └── HashJoin(Probe)
-        ├── output columns: [t.number (#0)]
-        ├── join type: INNER
-        ├── build keys: [t1.number (#1)]
-        ├── probe keys: [t.number (#0)]
-        ├── filters: []
-        ├── estimated rows: 1.00
-        ├── TableScan(Build)
-        │   ├── table: default.system.numbers
-        │   ├── output columns: [number (#1)]
-        │   ├── read rows: 1
-        │   ├── read bytes: 8
-        │   ├── partitions total: 1
-        │   ├── partitions scanned: 1
-        │   ├── push downs: [filters: [], limit: NONE]
-        │   └── estimated rows: 1.00
-        └── TableScan(Probe)
-            ├── table: default.system.numbers
-            ├── output columns: [number (#0)]
-            ├── read rows: 1
-            ├── read bytes: 8
-            ├── partitions total: 1
-            ├── partitions scanned: 1
-            ├── push downs: [filters: [], limit: NONE]
-            └── estimated rows: 1.00
+    ├── TableScan(Build)
+    │   ├── table: default.system.numbers
+    │   ├── output columns: [number (#1)]
+    │   ├── read rows: 1
+    │   ├── read bytes: 8
+    │   ├── partitions total: 1
+    │   ├── partitions scanned: 1
+    │   ├── push downs: [filters: [], limit: NONE]
+    │   └── estimated rows: 1.00
+    └── TableScan(Probe)
+        ├── table: default.system.numbers
+        ├── output columns: [number (#0)]
+        ├── read rows: 1
+        ├── read bytes: 8
+        ├── partitions total: 1
+        ├── partitions scanned: 1
+        ├── push downs: [filters: [], limit: NONE]
+        └── estimated rows: 1.00
 
 query T
 explain select t.number from numbers(1) as t where exists(select * from numbers(1) as t1 where t.number > t1.number) and not exists(select * from numbers(1) as t1 where t.number < t1.number)
diff --git a/tests/sqllogictests/suites/query/02_function/02_0000_function_arithmetic.test b/tests/sqllogictests/suites/query/02_function/02_0000_function_arithmetic.test
index cb02c41b11c3..782e3272847a 100644
--- a/tests/sqllogictests/suites/query/02_function/02_0000_function_arithmetic.test
+++ b/tests/sqllogictests/suites/query/02_function/02_0000_function_arithmetic.test
@@ -146,3 +146,25 @@ query I
 select * from numbers(4) where -number > -1;
 ----
 0
+
+
+## decimal
+
+statement ok
+create table test3 (
+	amount1 DECIMAL(38, 18) NULL,
+	amount2 DECIMAL(28, 8) NULL
+);
+
+statement ok
+insert into test3 values('30.606168460000000000','30.60616846');
+
+
+query TTTTTT
+select sum(amount1)a ,  sum(amount2) b , a  +  b, a - b, a * b, a / b from test3;
+----
+30.606168460000000000 30.60616846 61.212336920000000000 0.000000000000000000 936.73754780189877160000000000 1.000000000000000000
+
+statement ok
+drop table test3
+
diff --git a/tests/sqllogictests/suites/query/02_function/02_0005_function_substring.test b/tests/sqllogictests/suites/query/02_function/02_0005_function_substring.test
index 48be2d7c2222..f212bdeaa8be 100644
--- a/tests/sqllogictests/suites/query/02_function/02_0005_function_substring.test
+++ b/tests/sqllogictests/suites/query/02_function/02_0005_function_substring.test
@@ -195,7 +195,7 @@ SELECT SUBSTRING('12345', 0, 1)
 query TI
 select substr('城区主城区其他', 1, 6), length('我爱中国')
 ----
-城区	12
+城区主城区其 4
 
 
 statement ok
@@ -208,4 +208,4 @@ select substr('城区主城区其他', 1, 6), length('我爱中国')
 城区主城区其	4
 
 statement ok
-unset sql_dialect
\ No newline at end of file
+unset sql_dialect
diff --git a/tests/sqllogictests/suites/query/02_function/02_0040_function_strings_length.test b/tests/sqllogictests/suites/query/02_function/02_0040_function_strings_length.test
index 0ef21f1f73de..2595460d89ca 100644
--- a/tests/sqllogictests/suites/query/02_function/02_0040_function_strings_length.test
+++ b/tests/sqllogictests/suites/query/02_function/02_0040_function_strings_length.test
@@ -7,17 +7,17 @@ SELECT LENGTH('word')
 query I
 SELECT LENGTH('кириллица')
 ----
-18
+9
 
 query I
 SELECT LENGTH('кириллица and latin')
 ----
-28
+19
 
 query I
 SELECT LENGTH('你好')
 ----
-6
+2
 
 query T
 SELECT LENGTH(NULL)
diff --git a/tests/sqllogictests/suites/query/02_function/02_0043_function_char.test b/tests/sqllogictests/suites/query/02_function/02_0043_function_char.test
index 78bcd636aa10..599bfa1d8422 100644
--- a/tests/sqllogictests/suites/query/02_function/02_0043_function_char.test
+++ b/tests/sqllogictests/suites/query/02_function/02_0043_function_char.test
@@ -9,30 +9,30 @@ SELECT CHAR(97, NULL)
 NULL
 
 query T
-SELECT CHAR(97)
+SELECT CHAR(97)::STRING
 ----
 a
 
 query TT
-select char(97,98) from numbers(2);
+select char(97,98)::STRING from numbers(2);
 ----
 ab
 ab
 
 query T
-SELECT CHAR(77,121,83,81,76)
+SELECT CHAR(77,121,83,81,76)::STRING
 ----
 MySQL
 
 
 query T
-SELECT char(0xD0, 0xBF, 0xD1, 0x80, 0xD0, 0xB8, 0xD0, 0xB2, 0xD0, 0xB5, 0xD1, 0x82) AS hello
+SELECT char(0xD0, 0xBF, 0xD1, 0x80, 0xD0, 0xB8, 0xD0, 0xB2, 0xD0, 0xB5, 0xD1, 0x82)::STRING AS hello
 ----
 привет
 
 
 query T
-SELECT char(0xe4, 0xbd, 0xa0, 0xe5, 0xa5, 0xbd) AS hello
+SELECT char(0xe4, 0xbd, 0xa0, 0xe5, 0xa5, 0xbd)::STRING AS hello
 ----
 你好
 
@@ -43,7 +43,7 @@ statement ok
 create database if not exists db1
 
 statement ok
-create table db1.t as SELECT CHAR(77,121,83,81,76) AS hello
+create table db1.t as SELECT CHAR(77,121,83,81,76)::STRING AS hello
 
 query T
 select * from db1.t
diff --git a/tests/sqllogictests/suites/query/join/runtime_filter.test b/tests/sqllogictests/suites/query/join/runtime_filter.test
index b875356187dd..bdb84d3eb936 100644
--- a/tests/sqllogictests/suites/query/join/runtime_filter.test
+++ b/tests/sqllogictests/suites/query/join/runtime_filter.test
@@ -1,6 +1,3 @@
-statement ok
-set enable_runtime_filter = 1;
-
 statement ok
 CREATE TABLE table1 (
     key1 String,
@@ -89,9 +86,6 @@ NULL NULL
 NULL NULL
 NULL NULL
 
-statement ok
-set enable_runtime_filter = 0;
-
 statement ok
 drop table table1;
 
diff --git a/tests/sqllogictests/suites/query/subquery.test b/tests/sqllogictests/suites/query/subquery.test
index c88ffc8bd1d9..eaced5013cd6 100644
--- a/tests/sqllogictests/suites/query/subquery.test
+++ b/tests/sqllogictests/suites/query/subquery.test
@@ -733,3 +733,18 @@ drop table mortgage_records;
 
 statement ok
 drop table transaction_history;
+
+statement ok
+drop table if exists t1;
+
+statement ok
+create table t1(a int);
+
+statement ok
+insert into t1 values(1), (2), (3);
+
+query error 1001.*Scalar subquery can't return more than one row
+select (select sum(a) from t1 where t1.a >= t2.a group by t1.a) from t1 as t2;
+
+statement ok
+drop table t1;
diff --git a/tests/sqllogictests/suites/stage/formats/parquet/options/parquet_missing_field.test b/tests/sqllogictests/suites/stage/formats/parquet/options/parquet_missing_field.test
index bf8078f50ec3..522f9167dc50 100644
--- a/tests/sqllogictests/suites/stage/formats/parquet/options/parquet_missing_field.test
+++ b/tests/sqllogictests/suites/stage/formats/parquet/options/parquet_missing_field.test
@@ -19,7 +19,7 @@ c5 BIGINT 1 1
 c2 BIGINT 1 2
 c4 VARCHAR 1 3
 
-query error get diff schema
+query error
 copy into t1 from @data/parquet/diff_schema/ file_format=(type=parquet) pattern='.*[.]parquet'
 
 query 
diff --git a/tests/sqllogictests/suites/task/task_dag_test.test b/tests/sqllogictests/suites/task/task_dag_test.test
index 2d6dd58d71f3..0629fdf965cf 100644
--- a/tests/sqllogictests/suites/task/task_dag_test.test
+++ b/tests/sqllogictests/suites/task/task_dag_test.test
@@ -1,5 +1,5 @@
 # Please start the UDF Server first before running this test:
-#   python3 tests/cloud-control-server/simple_server.py
+#   python3 tests/cloud_control_server/simple_server.py
 #
 statement ok
 DROP TASK IF EXISTS mytaskroot
diff --git a/tests/sqllogictests/suites/task/task_ddl_test.test b/tests/sqllogictests/suites/task/task_ddl_test.test
index 8b83ab8d405e..678b5c651ab3 100644
--- a/tests/sqllogictests/suites/task/task_ddl_test.test
+++ b/tests/sqllogictests/suites/task/task_ddl_test.test
@@ -4,6 +4,9 @@
 statement ok
 DROP TASK IF EXISTS mytask
 
+statement ok
+DROP TASK IF EXISTS sessionTask
+
 statement ok
 CREATE TASK mytask
   WAREHOUSE = 'mywh'
@@ -94,5 +97,38 @@ select state from system.task_history where name = 'mytask'
 ----
 SUCCEEDED
 
+statement ok
+CREATE TASK sessionTask
+  WAREHOUSE = 'mywh'
+  SCHEDULE = USING CRON '0 0 0 1 1 ? 2100'
+  DATABASE = 'mydb', TIMEZONE = 'America/Los_Angeles'
+  AS SELECT 1;
+
+query SSS
+select name, state, session_parameters from system.tasks where name = 'sessionTask'
+----
+sessionTask Suspended {"database":"mydb","timezone":"America/Los_Angeles"}
+
+statement ok
+EXECUTE TASK sessionTask
+
+query SSSS
+select name, session_parameters from system.task_history where name = 'sessionTask'
+----
+sessionTask {"database":"mydb","timezone":"America/Los_Angeles"}
+
+statement ok
+ALTER TASK sessionTask
+  SET
+  DATABASE = 'db2', TIMEZONE = 'Pacific/Honolulu'
+
+query SSS
+select name, state, session_parameters from system.tasks where name = 'sessionTask'
+----
+sessionTask Suspended {"database":"db2","timezone":"Pacific/Honolulu"}
+
 statement ok
 DROP TASK mytask
+
+statement ok
+DROP TASK sessionTask
\ No newline at end of file
diff --git a/tests/sqllogictests/suites/tpcds/queries.test b/tests/sqllogictests/suites/tpcds/queries.test
index 72739a04b429..f8cef15ea28a 100644
--- a/tests/sqllogictests/suites/tpcds/queries.test
+++ b/tests/sqllogictests/suites/tpcds/queries.test
@@ -1,9 +1,6 @@
 statement ok
 set sandbox_tenant = 'test_tenant';
 
-statement ok
-set enable_runtime_filter = 1;
-
 statement ok
 use tpcds;
 
@@ -7979,6 +7976,3 @@ Conventional childr NEXT DAY ny metro 256 251 235 0 0
 Conventional childr OVERNIGHT ny metro 188 181 188 0 0
 Conventional childr REGULAR ny metro 179 150 215 0 0
 Conventional childr TWO DAY ny metro 185 183 158 0 0
-
-statement ok
-set enable_runtime_filter = 0
diff --git a/tests/sqllogictests/suites/tpcds/tpcds_join_order.test b/tests/sqllogictests/suites/tpcds/tpcds_join_order.test
index 1d46b29367df..f37aedc31742 100644
--- a/tests/sqllogictests/suites/tpcds/tpcds_join_order.test
+++ b/tests/sqllogictests/suites/tpcds/tpcds_join_order.test
@@ -1,9 +1,6 @@
 statement ok
 set sandbox_tenant = 'test_tenant';
 
-statement ok
-set enable_runtime_filter = 1;
-
 statement ok
 use tpcds;
 
@@ -34,27 +31,27 @@ WHERE ctr1.ctr_total_return >
 ORDER BY c_customer_id
 LIMIT 100;
 ----
-HashJoin: LEFT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── HashJoin: INNER
+│       │       ├── Build
+│       │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
+│       │       └── Probe
+│       │           └── Scan: default.tpcds.store_returns (#4) (read rows: 2810)
 │       └── Probe
-│           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│           └── HashJoin: INNER
+│               ├── Build
+│               │   └── Scan: default.tpcds.store (#2) (read rows: 1)
+│               └── Probe
+│                   └── HashJoin: INNER
+│                       ├── Build
+│                       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
+│                       └── Probe
+│                           └── Scan: default.tpcds.store_returns (#0) (read rows: 2810)
 └── Probe
-    └── HashJoin: INNER
-        ├── Build
-        │   └── HashJoin: INNER
-        │       ├── Build
-        │       │   └── Scan: default.tpcds.store (read rows: 1)
-        │       └── Probe
-        │           └── HashJoin: INNER
-        │               ├── Build
-        │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │               └── Probe
-        │                   └── Scan: default.tpcds.store_returns (read rows: 2810)
-        └── Probe
-            └── Scan: default.tpcds.customer (read rows: 1000)
+    └── Scan: default.tpcds.customer (#3) (read rows: 1000)
 
 # Q2
 query I
@@ -143,31 +140,31 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
 │               │   └── UnionAll
 │               │       ├── Left
-│               │       │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│               │       │   └── Scan: default.tpcds.web_sales (#4) (read rows: 7212)
 │               │       └── Right
-│               │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │           └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
 │               └── Probe
-│                   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
                 │   └── UnionAll
                 │       ├── Left
-                │       │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+                │       │   └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
                 │       └── Right
-                │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+                │           └── Scan: default.tpcds.catalog_sales (#1) (read rows: 14313)
                 └── Probe
-                    └── Scan: default.tpcds.date_dim (read rows: 73049)
+                    └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 
 # Q3
 query I
@@ -195,11 +192,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.item (read rows: 0)
+│       │   └── Scan: default.tpcds.item (#2) (read rows: 0)
 │       └── Probe
-│           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│           └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 └── Probe
-    └── Scan: default.tpcds.date_dim (read rows: 73049)
+    └── Scan: default.tpcds.date_dim (#0) (read rows: 73049)
 
 # Q4
 query I
@@ -342,31 +339,31 @@ HashJoin: INNER
 │       │       │       │       │       │       │       ├── Build
 │       │       │       │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       │       │       │       ├── Build
-│       │       │       │       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │       │       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#20) (read rows: 0)
 │       │       │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │       │       │       │       │       │           └── Scan: default.tpcds.store_sales (#19) (read rows: 0)
 │       │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │       │           └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │       │       │       │       │           └── Scan: default.tpcds.customer (#18) (read rows: 0)
 │       │       │       │       │       │       └── Right
 │       │       │       │       │       │           └── HashJoin: INNER
 │       │       │       │       │       │               ├── Build
 │       │       │       │       │       │               │   └── HashJoin: INNER
 │       │       │       │       │       │               │       ├── Build
-│       │       │       │       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │       │               │       │   └── Scan: default.tpcds.date_dim (#23) (read rows: 73049)
 │       │       │       │       │       │               │       └── Probe
-│       │       │       │       │       │               │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │       │       │       │       │               │           └── Scan: default.tpcds.catalog_sales (#22) (read rows: 14313)
 │       │       │       │       │       │               └── Probe
-│       │       │       │       │       │                   └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │       │       │       │                   └── Scan: default.tpcds.customer (#21) (read rows: 1000)
 │       │       │       │       │       └── Right
 │       │       │       │       │           └── HashJoin: INNER
 │       │       │       │       │               ├── Build
 │       │       │       │       │               │   └── HashJoin: INNER
 │       │       │       │       │               │       ├── Build
-│       │       │       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │       │       │               │       │   └── Scan: default.tpcds.date_dim (#26) (read rows: 0)
 │       │       │       │       │               │       └── Probe
-│       │       │       │       │               │           └── Scan: default.tpcds.web_sales (read rows: 0)
+│       │       │       │       │               │           └── Scan: default.tpcds.web_sales (#25) (read rows: 0)
 │       │       │       │       │               └── Probe
-│       │       │       │       │                   └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │       │       │                   └── Scan: default.tpcds.customer (#24) (read rows: 0)
 │       │       │       │       └── Probe
 │       │       │       │           └── UnionAll
 │       │       │       │               ├── Left
@@ -376,31 +373,31 @@ HashJoin: INNER
 │       │       │       │               │       │       ├── Build
 │       │       │       │               │       │       │   └── HashJoin: INNER
 │       │       │       │               │       │       │       ├── Build
-│       │       │       │               │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │               │       │       │       │   └── Scan: default.tpcds.date_dim (#38) (read rows: 0)
 │       │       │       │               │       │       │       └── Probe
-│       │       │       │               │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │               │       │       │           └── Scan: default.tpcds.store_sales (#37) (read rows: 0)
 │       │       │       │               │       │       └── Probe
-│       │       │       │               │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │       │               │       │           └── Scan: default.tpcds.customer (#36) (read rows: 0)
 │       │       │       │               │       └── Right
 │       │       │       │               │           └── HashJoin: INNER
 │       │       │       │               │               ├── Build
 │       │       │       │               │               │   └── HashJoin: INNER
 │       │       │       │               │               │       ├── Build
-│       │       │       │               │               │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │       │               │               │       │   └── Scan: default.tpcds.date_dim (#41) (read rows: 0)
 │       │       │       │               │               │       └── Probe
-│       │       │       │               │               │           └── Scan: default.tpcds.catalog_sales (read rows: 0)
+│       │       │       │               │               │           └── Scan: default.tpcds.catalog_sales (#40) (read rows: 0)
 │       │       │       │               │               └── Probe
-│       │       │       │               │                   └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │       │               │                   └── Scan: default.tpcds.customer (#39) (read rows: 0)
 │       │       │       │               └── Right
 │       │       │       │                   └── HashJoin: INNER
 │       │       │       │                       ├── Build
 │       │       │       │                       │   └── HashJoin: INNER
 │       │       │       │                       │       ├── Build
-│       │       │       │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │       │                       │       │   └── Scan: default.tpcds.date_dim (#44) (read rows: 73049)
 │       │       │       │                       │       └── Probe
-│       │       │       │                       │           └── Scan: default.tpcds.web_sales (read rows: 0)
+│       │       │       │                       │           └── Scan: default.tpcds.web_sales (#43) (read rows: 7212)
 │       │       │       │                       └── Probe
-│       │       │       │                           └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │       │                           └── Scan: default.tpcds.customer (#42) (read rows: 1000)
 │       │       │       └── Probe
 │       │       │           └── UnionAll
 │       │       │               ├── Left
@@ -410,31 +407,31 @@ HashJoin: INNER
 │       │       │               │       │       ├── Build
 │       │       │               │       │       │   └── HashJoin: INNER
 │       │       │               │       │       │       ├── Build
-│       │       │               │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │               │       │       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │       │               │       │       │       └── Probe
-│       │       │               │       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │               │       │       │           └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 │       │       │               │       │       └── Probe
-│       │       │               │       │           └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │               │       │           └── Scan: default.tpcds.customer (#0) (read rows: 1000)
 │       │       │               │       └── Right
 │       │       │               │           └── HashJoin: INNER
 │       │       │               │               ├── Build
 │       │       │               │               │   └── HashJoin: INNER
 │       │       │               │               │       ├── Build
-│       │       │               │               │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │               │               │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 0)
 │       │       │               │               │       └── Probe
-│       │       │               │               │           └── Scan: default.tpcds.catalog_sales (read rows: 0)
+│       │       │               │               │           └── Scan: default.tpcds.catalog_sales (#4) (read rows: 0)
 │       │       │               │               └── Probe
-│       │       │               │                   └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │               │                   └── Scan: default.tpcds.customer (#3) (read rows: 0)
 │       │       │               └── Right
 │       │       │                   └── HashJoin: INNER
 │       │       │                       ├── Build
 │       │       │                       │   └── HashJoin: INNER
 │       │       │                       │       ├── Build
-│       │       │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │                       │       │   └── Scan: default.tpcds.date_dim (#8) (read rows: 0)
 │       │       │                       │       └── Probe
-│       │       │                       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │                       │           └── Scan: default.tpcds.web_sales (#7) (read rows: 0)
 │       │       │                       └── Probe
-│       │       │                           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │                           └── Scan: default.tpcds.customer (#6) (read rows: 0)
 │       │       └── Probe
 │       │           └── UnionAll
 │       │               ├── Left
@@ -444,31 +441,31 @@ HashJoin: INNER
 │       │               │       │       ├── Build
 │       │               │       │       │   └── HashJoin: INNER
 │       │               │       │       │       ├── Build
-│       │               │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │               │       │       │       │   └── Scan: default.tpcds.date_dim (#29) (read rows: 0)
 │       │               │       │       │       └── Probe
-│       │               │       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │               │       │       │           └── Scan: default.tpcds.store_sales (#28) (read rows: 0)
 │       │               │       │       └── Probe
-│       │               │       │           └── Scan: default.tpcds.customer (read rows: 0)
+│       │               │       │           └── Scan: default.tpcds.customer (#27) (read rows: 0)
 │       │               │       └── Right
 │       │               │           └── HashJoin: INNER
 │       │               │               ├── Build
 │       │               │               │   └── HashJoin: INNER
 │       │               │               │       ├── Build
-│       │               │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │               │       │   └── Scan: default.tpcds.date_dim (#32) (read rows: 73049)
 │       │               │               │       └── Probe
-│       │               │               │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │               │               │           └── Scan: default.tpcds.catalog_sales (#31) (read rows: 14313)
 │       │               │               └── Probe
-│       │               │                   └── Scan: default.tpcds.customer (read rows: 1000)
+│       │               │                   └── Scan: default.tpcds.customer (#30) (read rows: 1000)
 │       │               └── Right
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
 │       │                       │   └── HashJoin: INNER
 │       │                       │       ├── Build
-│       │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │                       │       │   └── Scan: default.tpcds.date_dim (#35) (read rows: 0)
 │       │                       │       └── Probe
-│       │                       │           └── Scan: default.tpcds.web_sales (read rows: 0)
+│       │                       │           └── Scan: default.tpcds.web_sales (#34) (read rows: 0)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.customer (read rows: 0)
+│       │                           └── Scan: default.tpcds.customer (#33) (read rows: 0)
 │       └── Probe
 │           └── UnionAll
 │               ├── Left
@@ -478,31 +475,31 @@ HashJoin: INNER
 │               │       │       ├── Build
 │               │       │       │   └── HashJoin: INNER
 │               │       │       │       ├── Build
-│               │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│               │       │       │       │   └── Scan: default.tpcds.date_dim (#47) (read rows: 0)
 │               │       │       │       └── Probe
-│               │       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│               │       │       │           └── Scan: default.tpcds.store_sales (#46) (read rows: 0)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.customer (read rows: 0)
+│               │       │           └── Scan: default.tpcds.customer (#45) (read rows: 0)
 │               │       └── Right
 │               │           └── HashJoin: INNER
 │               │               ├── Build
 │               │               │   └── HashJoin: INNER
 │               │               │       ├── Build
-│               │               │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│               │               │       │   └── Scan: default.tpcds.date_dim (#50) (read rows: 0)
 │               │               │       └── Probe
-│               │               │           └── Scan: default.tpcds.catalog_sales (read rows: 0)
+│               │               │           └── Scan: default.tpcds.catalog_sales (#49) (read rows: 0)
 │               │               └── Probe
-│               │                   └── Scan: default.tpcds.customer (read rows: 0)
+│               │                   └── Scan: default.tpcds.customer (#48) (read rows: 0)
 │               └── Right
 │                   └── HashJoin: INNER
 │                       ├── Build
 │                       │   └── HashJoin: INNER
 │                       │       ├── Build
-│                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │       │   └── Scan: default.tpcds.date_dim (#53) (read rows: 73049)
 │                       │       └── Probe
-│                       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│                       │           └── Scan: default.tpcds.web_sales (#52) (read rows: 7212)
 │                       └── Probe
-│                           └── Scan: default.tpcds.customer (read rows: 1000)
+│                           └── Scan: default.tpcds.customer (#51) (read rows: 1000)
 └── Probe
     └── UnionAll
         ├── Left
@@ -512,31 +509,31 @@ HashJoin: INNER
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │       │   └── Scan: default.tpcds.date_dim (#11) (read rows: 73049)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │       │           └── Scan: default.tpcds.store_sales (#10) (read rows: 28810)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │       │           └── Scan: default.tpcds.customer (#9) (read rows: 1000)
         │       └── Right
         │           └── HashJoin: INNER
         │               ├── Build
         │               │   └── HashJoin: INNER
         │               │       ├── Build
-        │               │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+        │               │       │   └── Scan: default.tpcds.date_dim (#14) (read rows: 0)
         │               │       └── Probe
-        │               │           └── Scan: default.tpcds.catalog_sales (read rows: 0)
+        │               │           └── Scan: default.tpcds.catalog_sales (#13) (read rows: 0)
         │               └── Probe
-        │                   └── Scan: default.tpcds.customer (read rows: 0)
+        │                   └── Scan: default.tpcds.customer (#12) (read rows: 0)
         └── Right
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+                │       │   └── Scan: default.tpcds.date_dim (#17) (read rows: 0)
                 │       └── Probe
-                │           └── Scan: default.tpcds.web_sales (read rows: 0)
+                │           └── Scan: default.tpcds.web_sales (#16) (read rows: 0)
                 └── Probe
-                    └── Scan: default.tpcds.customer (read rows: 0)
+                    └── Scan: default.tpcds.customer (#15) (read rows: 0)
 
 
 # Q5
@@ -661,49 +658,49 @@ UnionAll
 │       ├── Left
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.store (read rows: 1)
+│       │               │   └── Scan: default.tpcds.store (#3) (read rows: 1)
 │       │               └── Probe
 │       │                   └── UnionAll
 │       │                       ├── Left
-│       │                       │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                       │   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │                       └── Right
-│       │                           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │                           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
 │       └── Right
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.catalog_page (read rows: 11718)
+│               │   └── Scan: default.tpcds.catalog_page (#7) (read rows: 11718)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │                       └── Probe
 │                           └── UnionAll
 │                               ├── Left
-│                               │   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│                               │   └── Scan: default.tpcds.catalog_sales (#4) (read rows: 14313)
 │                               └── Right
-│                                   └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│                                   └── Scan: default.tpcds.catalog_returns (#5) (read rows: 1358)
 └── Right
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#11) (read rows: 73049)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.web_site (read rows: 1)
+                │   └── Scan: default.tpcds.web_site (#12) (read rows: 1)
                 └── Probe
                     └── UnionAll
                         ├── Left
-                        │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+                        │   └── Scan: default.tpcds.web_sales (#8) (read rows: 7212)
                         └── Right
                             └── HashJoin: RIGHT OUTER
                                 ├── Build
-                                │   └── Scan: default.tpcds.web_returns (read rows: 679)
+                                │   └── Scan: default.tpcds.web_returns (#9) (read rows: 679)
                                 └── Probe
-                                    └── Scan: default.tpcds.web_sales (read rows: 7212)
+                                    └── Scan: default.tpcds.web_sales (#10) (read rows: 7212)
 
 # Q6
 query I
@@ -734,31 +731,31 @@ ORDER BY cnt NULLS FIRST,
          a.ca_state NULLS FIRST
 LIMIT 100;
 ----
-HashJoin: LEFT SINGLE
+HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.item (read rows: 180)
+│   └── Scan: default.tpcds.item (#6) (read rows: 180)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.item (read rows: 180)
+                │   └── HashJoin: INNER
+                │       ├── Build
+                │       │   └── Scan: default.tpcds.customer_address (#0) (read rows: 500)
+                │       └── Probe
+                │           └── Scan: default.tpcds.customer (#1) (read rows: 1000)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
-                        │   └── HashJoin: INNER
-                        │       ├── Build
-                        │       │   └── HashJoin: INNER
-                        │       │       ├── Build
-                        │       │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
-                        │       │       └── Probe
-                        │       │           └── Scan: default.tpcds.customer (read rows: 1000)
-                        │       └── Probe
-                        │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                        │   └── Scan: default.tpcds.item (#4) (read rows: 180)
                         └── Probe
-                            └── Scan: default.tpcds.date_dim (read rows: 73049)
+                            └── HashJoin: INNER
+                                ├── Build
+                                │   └── Scan: default.tpcds.store_sales (#2) (read rows: 28810)
+                                └── Probe
+                                    └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 
 # Q7
 query I
@@ -789,21 +786,21 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.promotion (read rows: 3)
+│   └── Scan: default.tpcds.promotion (#4) (read rows: 3)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#3) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
                 │       └── Probe
-                │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
                 └── Probe
-                    └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+                    └── Scan: default.tpcds.customer_demographics (#1) (read rows: 19208)
 
 
 # Q8
@@ -1243,21 +1240,21 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │   └── Scan: default.tpcds.customer (#5) (read rows: 1000)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │           └── Scan: default.tpcds.customer_address (#4) (read rows: 500)
 │       └── Probe
-│           └── Scan: default.tpcds.customer_address (read rows: 500)
+│           └── Scan: default.tpcds.customer_address (#3) (read rows: 500)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.store (read rows: 1)
+        │   └── Scan: default.tpcds.store (#2) (read rows: 1)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
                 └── Probe
-                    └── Scan: default.tpcds.store_sales (read rows: 28810)
+                    └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 
 # Q9
 query I
@@ -1332,65 +1329,65 @@ WHERE r_reason_sk = 1 ;
 ----
 HashJoin: CROSS
 ├── Build
-│   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│   └── Scan: default.tpcds.store_sales (#15) (read rows: 28810)
 └── Probe
     └── HashJoin: CROSS
         ├── Build
-        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │   └── Scan: default.tpcds.store_sales (#14) (read rows: 28810)
         └── Probe
             └── HashJoin: CROSS
                 ├── Build
-                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                │   └── Scan: default.tpcds.store_sales (#13) (read rows: 28810)
                 └── Probe
                     └── HashJoin: CROSS
                         ├── Build
-                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                        │   └── Scan: default.tpcds.store_sales (#12) (read rows: 28810)
                         └── Probe
                             └── HashJoin: CROSS
                                 ├── Build
-                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                │   └── Scan: default.tpcds.store_sales (#11) (read rows: 28810)
                                 └── Probe
                                     └── HashJoin: CROSS
                                         ├── Build
-                                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                        │   └── Scan: default.tpcds.store_sales (#10) (read rows: 28810)
                                         └── Probe
                                             └── HashJoin: CROSS
                                                 ├── Build
-                                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                │   └── Scan: default.tpcds.store_sales (#9) (read rows: 28810)
                                                 └── Probe
                                                     └── HashJoin: CROSS
                                                         ├── Build
-                                                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                        │   └── Scan: default.tpcds.store_sales (#8) (read rows: 28810)
                                                         └── Probe
                                                             └── HashJoin: CROSS
                                                                 ├── Build
-                                                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                │   └── Scan: default.tpcds.store_sales (#7) (read rows: 28810)
                                                                 └── Probe
                                                                     └── HashJoin: CROSS
                                                                         ├── Build
-                                                                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                        │   └── Scan: default.tpcds.store_sales (#6) (read rows: 28810)
                                                                         └── Probe
                                                                             └── HashJoin: CROSS
                                                                                 ├── Build
-                                                                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                                │   └── Scan: default.tpcds.store_sales (#5) (read rows: 28810)
                                                                                 └── Probe
                                                                                     └── HashJoin: CROSS
                                                                                         ├── Build
-                                                                                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                                        │   └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
                                                                                         └── Probe
                                                                                             └── HashJoin: CROSS
                                                                                                 ├── Build
-                                                                                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                                                │   └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
                                                                                                 └── Probe
                                                                                                     └── HashJoin: CROSS
                                                                                                         ├── Build
-                                                                                                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                                                        │   └── Scan: default.tpcds.store_sales (#2) (read rows: 28810)
                                                                                                         └── Probe
                                                                                                             └── HashJoin: CROSS
                                                                                                                 ├── Build
-                                                                                                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                                                                                │   └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
                                                                                                                 └── Probe
-                                                                                                                    └── Scan: default.tpcds.reason (read rows: 1)
+                                                                                                                    └── Scan: default.tpcds.reason (#0) (read rows: 1)
 
 # Q10
 query I
@@ -1461,39 +1458,39 @@ ORDER BY cd_gender,
          cd_dep_college_count
 LIMIT 100;
 ----
-HashJoin: RIGHT MARK
+HashJoin: LEFT MARK
 ├── Build
-│   └── HashJoin: INNER
+│   └── HashJoin: LEFT MARK
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── HashJoin: LEFT SEMI
+│       │       ├── Build
+│       │       │   └── HashJoin: INNER
+│       │       │       ├── Build
+│       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
+│       │       │       └── Probe
+│       │       │           └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
+│       │       └── Probe
+│       │           └── HashJoin: INNER
+│       │               ├── Build
+│       │               │   └── HashJoin: INNER
+│       │               │       ├── Build
+│       │               │       │   └── Scan: default.tpcds.customer_address (#1) (read rows: 500)
+│       │               │       └── Probe
+│       │               │           └── Scan: default.tpcds.customer (#0) (read rows: 1000)
+│       │               └── Probe
+│       │                   └── Scan: default.tpcds.customer_demographics (#2) (read rows: 19208)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── HashJoin: INNER
+│               ├── Build
+│               │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
+│               └── Probe
+│                   └── Scan: default.tpcds.web_sales (#5) (read rows: 7212)
 └── Probe
-    └── HashJoin: RIGHT MARK
+    └── HashJoin: INNER
         ├── Build
-        │   └── HashJoin: INNER
-        │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       └── Probe
-        │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │   └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
         └── Probe
-            └── HashJoin: LEFT SEMI
-                ├── Build
-                │   └── HashJoin: INNER
-                │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-                │       └── Probe
-                │           └── Scan: default.tpcds.store_sales (read rows: 28810)
-                └── Probe
-                    └── HashJoin: INNER
-                        ├── Build
-                        │   └── HashJoin: INNER
-                        │       ├── Build
-                        │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
-                        │       └── Probe
-                        │           └── Scan: default.tpcds.customer (read rows: 1000)
-                        └── Probe
-                            └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+            └── Scan: default.tpcds.catalog_sales (#7) (read rows: 14313)
 
 # Q11
 query I
@@ -1591,21 +1588,21 @@ HashJoin: INNER
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#14) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │       │       │           └── Scan: default.tpcds.store_sales (#13) (read rows: 0)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │       │           └── Scan: default.tpcds.customer (#12) (read rows: 0)
 │       │       │       └── Right
 │       │       │           └── HashJoin: INNER
 │       │       │               ├── Build
 │       │       │               │   └── HashJoin: INNER
 │       │       │               │       ├── Build
-│       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │               │       │   └── Scan: default.tpcds.date_dim (#17) (read rows: 73049)
 │       │       │               │       └── Probe
-│       │       │               │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │               │           └── Scan: default.tpcds.web_sales (#16) (read rows: 7212)
 │       │       │               └── Probe
-│       │       │                   └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │                   └── Scan: default.tpcds.customer (#15) (read rows: 1000)
 │       │       └── Probe
 │       │           └── UnionAll
 │       │               ├── Left
@@ -1613,21 +1610,21 @@ HashJoin: INNER
 │       │               │       ├── Build
 │       │               │       │   └── HashJoin: INNER
 │       │               │       │       ├── Build
-│       │               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │               │       │       └── Probe
-│       │               │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │               │       │           └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 │       │               │       └── Probe
-│       │               │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │               │           └── Scan: default.tpcds.customer (#0) (read rows: 1000)
 │       │               └── Right
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
 │       │                       │   └── HashJoin: INNER
 │       │                       │       ├── Build
-│       │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │                       │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 0)
 │       │                       │       └── Probe
-│       │                       │           └── Scan: default.tpcds.web_sales (read rows: 0)
+│       │                       │           └── Scan: default.tpcds.web_sales (#4) (read rows: 0)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.customer (read rows: 0)
+│       │                           └── Scan: default.tpcds.customer (#3) (read rows: 0)
 │       └── Probe
 │           └── UnionAll
 │               ├── Left
@@ -1635,21 +1632,21 @@ HashJoin: INNER
 │               │       ├── Build
 │               │       │   └── HashJoin: INNER
 │               │       │       ├── Build
-│               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│               │       │       │   └── Scan: default.tpcds.date_dim (#20) (read rows: 0)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│               │       │           └── Scan: default.tpcds.store_sales (#19) (read rows: 0)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.customer (read rows: 0)
+│               │           └── Scan: default.tpcds.customer (#18) (read rows: 0)
 │               └── Right
 │                   └── HashJoin: INNER
 │                       ├── Build
 │                       │   └── HashJoin: INNER
 │                       │       ├── Build
-│                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │       │   └── Scan: default.tpcds.date_dim (#23) (read rows: 73049)
 │                       │       └── Probe
-│                       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│                       │           └── Scan: default.tpcds.web_sales (#22) (read rows: 7212)
 │                       └── Probe
-│                           └── Scan: default.tpcds.customer (read rows: 1000)
+│                           └── Scan: default.tpcds.customer (#21) (read rows: 1000)
 └── Probe
     └── UnionAll
         ├── Left
@@ -1657,21 +1654,21 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │   └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │           └── Scan: default.tpcds.store_sales (#7) (read rows: 28810)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │           └── Scan: default.tpcds.customer (#6) (read rows: 1000)
         └── Right
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+                │       │   └── Scan: default.tpcds.date_dim (#11) (read rows: 0)
                 │       └── Probe
-                │           └── Scan: default.tpcds.web_sales (read rows: 0)
+                │           └── Scan: default.tpcds.web_sales (#10) (read rows: 0)
                 └── Probe
-                    └── Scan: default.tpcds.customer (read rows: 0)
+                    └── Scan: default.tpcds.customer (#9) (read rows: 0)
 
 # Q12
 query I
@@ -1708,11 +1705,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#1) (read rows: 180)
 
 # Q13
 query I
@@ -1769,17 +1766,17 @@ HashJoin: INNER
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │       │   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │           └── Scan: default.tpcds.customer_address (#4) (read rows: 500)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │           └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.store (read rows: 1)
+│       │           └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
 │       └── Probe
-│           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│           └── Scan: default.tpcds.store (#1) (read rows: 1)
 └── Probe
-    └── Scan: default.tpcds.household_demographics (read rows: 7200)
+    └── Scan: default.tpcds.customer_demographics (#2) (read rows: 19208)
 
 
 # Q14
@@ -2059,13 +2056,13 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 │       └── Probe
-│           └── Scan: default.tpcds.customer (read rows: 1000)
+│           └── Scan: default.tpcds.customer (#1) (read rows: 1000)
 └── Probe
-    └── Scan: default.tpcds.customer_address (read rows: 500)
+    └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 
 # Q16
 query I
@@ -2105,17 +2102,17 @@ HashJoin: RIGHT ANTI
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │       │       │           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │           └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.call_center (read rows: 1)
+│       │           └── Scan: default.tpcds.call_center (#3) (read rows: 1)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── Scan: default.tpcds.catalog_sales (#4) (read rows: 14313)
 └── Probe
-    └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+    └── Scan: default.tpcds.catalog_returns (#5) (read rows: 1358)
 
 # Q17
 query I
@@ -2180,23 +2177,23 @@ HashJoin: INNER
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │       │       │       │           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
 │       │       │       │       └── Probe
 │       │       │       │           └── HashJoin: INNER
 │       │       │       │               ├── Build
-│       │       │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │               │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 │       │       │       │               └── Probe
-│       │       │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │                   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.store (read rows: 1)
+│       │       │           └── Scan: default.tpcds.store (#6) (read rows: 1)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │           └── Scan: default.tpcds.item (#7) (read rows: 180)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── Scan: default.tpcds.catalog_sales (#2) (read rows: 14313)
 └── Probe
-    └── Scan: default.tpcds.date_dim (read rows: 73049)
+    └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 
 # Q18
 query I
@@ -2263,19 +2260,19 @@ HashJoin: INNER
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │       │       │       │           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │       │           └── Scan: default.tpcds.customer (#3) (read rows: 1000)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │       │           └── Scan: default.tpcds.customer_demographics (#1) (read rows: 19208)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │           └── Scan: default.tpcds.customer_demographics (#2) (read rows: 19208)
 │       └── Probe
-│           └── Scan: default.tpcds.customer_address (read rows: 500)
+│           └── Scan: default.tpcds.customer_address (#4) (read rows: 500)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#6) (read rows: 180)
 
 # Q19
 query I
@@ -2313,7 +2310,7 @@ LIMIT 100 ;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.store (read rows: 1)
+│   └── Scan: default.tpcds.store (#5) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -2321,17 +2318,17 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+        │       │       │   └── Scan: default.tpcds.item (#2) (read rows: 180)
         │       │       └── Probe
         │       │           └── HashJoin: INNER
         │       │               ├── Build
-        │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │               │   └── Scan: default.tpcds.date_dim (#0) (read rows: 73049)
         │       │               └── Probe
-        │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │                   └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │           └── Scan: default.tpcds.customer (#3) (read rows: 1000)
         └── Probe
-            └── Scan: default.tpcds.customer_address (read rows: 500)
+            └── Scan: default.tpcds.customer_address (#4) (read rows: 500)
 
 # Q20
 query I
@@ -2368,11 +2365,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#1) (read rows: 180)
 
 # Q21
 query I
@@ -2412,15 +2409,15 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │   └── Scan: default.tpcds.item (#2) (read rows: 180)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 │               └── Probe
-│                   └── Scan: default.tpcds.inventory (read rows: 23490)
+│                   └── Scan: default.tpcds.inventory (#0) (read rows: 23490)
 └── Probe
-    └── Scan: default.tpcds.warehouse (read rows: 1)
+    └── Scan: default.tpcds.warehouse (#1) (read rows: 1)
 
 # Q22
 query I
@@ -2448,11 +2445,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.inventory (read rows: 23490)
+│           └── Scan: default.tpcds.inventory (#0) (read rows: 23490)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#2) (read rows: 180)
 
 # Q23
 query I
@@ -2554,15 +2551,15 @@ UnionAll
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │       │       │       │   └── Scan: default.tpcds.item (#5) (read rows: 180)
 │       │       │       │       │       └── Probe
 │       │       │       │       │           └── HashJoin: INNER
 │       │       │       │       │               ├── Build
-│       │       │       │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │               │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       │       │       │       │               └── Probe
-│       │       │       │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │       │                   └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │       │       │           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 │       │       │       └── Probe
 │       │       │           └── HashJoin: CROSS
 │       │       │               ├── Build
@@ -2570,21 +2567,21 @@ UnionAll
 │       │       │               │       ├── Build
 │       │       │               │       │   └── HashJoin: INNER
 │       │       │               │       │       ├── Build
-│       │       │               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │               │       │       │   └── Scan: default.tpcds.date_dim (#10) (read rows: 73049)
 │       │       │               │       │       └── Probe
-│       │       │               │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │               │       │           └── Scan: default.tpcds.store_sales (#8) (read rows: 28810)
 │       │       │               │       └── Probe
-│       │       │               │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │               │           └── Scan: default.tpcds.customer (#9) (read rows: 1000)
 │       │       │               └── Probe
 │       │       │                   └── HashJoin: INNER
 │       │       │                       ├── Build
-│       │       │                       │   └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │                       │   └── Scan: default.tpcds.customer (#7) (read rows: 1000)
 │       │       │                       └── Probe
-│       │       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │                           └── Scan: default.tpcds.store_sales (#6) (read rows: 28810)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │           └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.customer (read rows: 1000)
+│           └── Scan: default.tpcds.customer (#1) (read rows: 1000)
 └── Right
     └── HashJoin: INNER
         ├── Build
@@ -2596,15 +2593,15 @@ UnionAll
         │       │       │       ├── Build
         │       │       │       │   └── HashJoin: INNER
         │       │       │       │       ├── Build
-        │       │       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+        │       │       │       │       │   └── Scan: default.tpcds.item (#16) (read rows: 180)
         │       │       │       │       └── Probe
         │       │       │       │           └── HashJoin: INNER
         │       │       │       │               ├── Build
-        │       │       │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │       │               │   └── Scan: default.tpcds.date_dim (#15) (read rows: 73049)
         │       │       │       │               └── Probe
-        │       │       │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │       │       │                   └── Scan: default.tpcds.store_sales (#14) (read rows: 28810)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │       │           └── Scan: default.tpcds.web_sales (#11) (read rows: 7212)
         │       │       └── Probe
         │       │           └── HashJoin: CROSS
         │       │               ├── Build
@@ -2612,21 +2609,21 @@ UnionAll
         │       │               │       ├── Build
         │       │               │       │   └── HashJoin: INNER
         │       │               │       │       ├── Build
-        │       │               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │               │       │       │   └── Scan: default.tpcds.date_dim (#21) (read rows: 73049)
         │       │               │       │       └── Probe
-        │       │               │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │               │       │           └── Scan: default.tpcds.store_sales (#19) (read rows: 28810)
         │       │               │       └── Probe
-        │       │               │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │       │               │           └── Scan: default.tpcds.customer (#20) (read rows: 1000)
         │       │               └── Probe
         │       │                   └── HashJoin: INNER
         │       │                       ├── Build
-        │       │                       │   └── Scan: default.tpcds.customer (read rows: 1000)
+        │       │                       │   └── Scan: default.tpcds.customer (#18) (read rows: 1000)
         │       │                       └── Probe
-        │       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │                           └── Scan: default.tpcds.store_sales (#17) (read rows: 28810)
         │       └── Probe
-        │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │           └── Scan: default.tpcds.date_dim (#13) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.customer (read rows: 1000)
+            └── Scan: default.tpcds.customer (#12) (read rows: 1000)
 
 # Q24
 query I
@@ -2696,17 +2693,17 @@ RangeJoin: INNER
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.store (#2) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │       │       │           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.item (read rows: 0)
+│       │       │           └── Scan: default.tpcds.item (#3) (read rows: 0)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │           └── Scan: default.tpcds.customer (#4) (read rows: 1000)
 │       └── Probe
-│           └── Scan: default.tpcds.customer_address (read rows: 500)
+│           └── Scan: default.tpcds.customer_address (#5) (read rows: 500)
 └── Right
     └── HashJoin: INNER
         ├── Build
@@ -2718,17 +2715,17 @@ RangeJoin: INNER
         │       │       │       ├── Build
         │       │       │       │   └── HashJoin: INNER
         │       │       │       │       ├── Build
-        │       │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+        │       │       │       │       │   └── Scan: default.tpcds.store (#8) (read rows: 0)
         │       │       │       │       └── Probe
-        │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │       │       │           └── Scan: default.tpcds.store_sales (#6) (read rows: 28810)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+        │       │       │           └── Scan: default.tpcds.store_returns (#7) (read rows: 2810)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.item (read rows: 180)
+        │       │           └── Scan: default.tpcds.item (#9) (read rows: 180)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │           └── Scan: default.tpcds.customer (#10) (read rows: 1000)
         └── Probe
-            └── Scan: default.tpcds.customer_address (read rows: 500)
+            └── Scan: default.tpcds.customer_address (#11) (read rows: 500)
 
 # Q25
 query I
@@ -2776,33 +2773,33 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── HashJoin: INNER
-│       ├── Build
-│       │   └── Scan: default.tpcds.store (read rows: 1)
-│       └── Probe
-│           └── HashJoin: INNER
-│               ├── Build
-│               │   └── HashJoin: INNER
-│               │       ├── Build
-│               │       │   └── HashJoin: INNER
-│               │       │       ├── Build
-│               │       │       │   └── HashJoin: INNER
-│               │       │       │       ├── Build
-│               │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-│               │       │       │       └── Probe
-│               │       │       │           └── Scan: default.tpcds.store_returns (read rows: 2810)
-│               │       │       └── Probe
-│               │       │           └── HashJoin: INNER
-│               │       │               ├── Build
-│               │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-│               │       │               └── Probe
-│               │       │                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
-│               │       └── Probe
-│               │           └── Scan: default.tpcds.store_sales (read rows: 28810)
-│               └── Probe
-│                   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│   └── Scan: default.tpcds.store (#6) (read rows: 1)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── HashJoin: INNER
+        ├── Build
+        │   └── HashJoin: INNER
+        │       ├── Build
+        │       │   └── HashJoin: INNER
+        │       │       ├── Build
+        │       │       │   └── HashJoin: INNER
+        │       │       │       ├── Build
+        │       │       │       │   └── HashJoin: INNER
+        │       │       │       │       ├── Build
+        │       │       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
+        │       │       │       │       └── Probe
+        │       │       │       │           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
+        │       │       │       └── Probe
+        │       │       │           └── HashJoin: INNER
+        │       │       │               ├── Build
+        │       │       │               │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
+        │       │       │               └── Probe
+        │       │       │                   └── Scan: default.tpcds.catalog_sales (#2) (read rows: 14313)
+        │       │       └── Probe
+        │       │           └── Scan: default.tpcds.item (#7) (read rows: 180)
+        │       └── Probe
+        │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
+        └── Probe
+            └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 
 # Q26
 query I
@@ -2833,7 +2830,7 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.promotion (read rows: 3)
+│   └── Scan: default.tpcds.promotion (#4) (read rows: 3)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -2841,13 +2838,13 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+        │       │           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+        │           └── Scan: default.tpcds.customer_demographics (#1) (read rows: 19208)
         └── Probe
-            └── Scan: default.tpcds.item (read rows: 180)
+            └── Scan: default.tpcds.item (#3) (read rows: 180)
 
 # Q27
 query I
@@ -2919,57 +2916,57 @@ UnionAll
 │       ├── Left
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │   └── Scan: default.tpcds.item (#4) (read rows: 180)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.store (read rows: 1)
+│       │               │   └── Scan: default.tpcds.store (#3) (read rows: 1)
 │       │               └── Probe
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
 │       │                       │   └── HashJoin: INNER
 │       │                       │       ├── Build
-│       │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │                       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │                       │       └── Probe
-│       │                       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │                           └── Scan: default.tpcds.customer_demographics (#1) (read rows: 19208)
 │       └── Right
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │   └── Scan: default.tpcds.item (#9) (read rows: 180)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.store (read rows: 1)
+│                       │   └── Scan: default.tpcds.store (#8) (read rows: 1)
 │                       └── Probe
 │                           └── HashJoin: INNER
 │                               ├── Build
 │                               │   └── HashJoin: INNER
 │                               │       ├── Build
-│                               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                               │       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │                               │       └── Probe
-│                               │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                               │           └── Scan: default.tpcds.store_sales (#5) (read rows: 28810)
 │                               └── Probe
-│                                   └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│                                   └── Scan: default.tpcds.customer_demographics (#6) (read rows: 19208)
 └── Right
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#14) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.store (read rows: 1)
+                │   └── Scan: default.tpcds.store (#13) (read rows: 1)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
                         │   └── HashJoin: INNER
                         │       ├── Build
-                        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                        │       │   └── Scan: default.tpcds.date_dim (#12) (read rows: 73049)
                         │       └── Probe
-                        │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                        │           └── Scan: default.tpcds.store_sales (#10) (read rows: 28810)
                         └── Probe
-                            └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+                            └── Scan: default.tpcds.customer_demographics (#11) (read rows: 19208)
 
 # Q28
 query I
@@ -3028,25 +3025,25 @@ LIMIT 100;
 ----
 HashJoin: CROSS
 ├── Build
-│   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│   └── Scan: default.tpcds.store_sales (#5) (read rows: 28810)
 └── Probe
     └── HashJoin: CROSS
         ├── Build
-        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │   └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
         └── Probe
             └── HashJoin: CROSS
                 ├── Build
-                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                │   └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
                 └── Probe
                     └── HashJoin: CROSS
                         ├── Build
-                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                        │   └── Scan: default.tpcds.store_sales (#2) (read rows: 28810)
                         └── Probe
                             └── HashJoin: CROSS
                                 ├── Build
-                                │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                │   └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
                                 └── Probe
-                                    └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                    └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 
 
 # Q29
@@ -3102,27 +3099,27 @@ HashJoin: INNER
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │       │       │   └── Scan: default.tpcds.store (#6) (read rows: 1)
 │       │       │       └── Probe
 │       │       │           └── HashJoin: INNER
 │       │       │               ├── Build
 │       │       │               │   └── HashJoin: INNER
 │       │       │               │       ├── Build
-│       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │               │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       │       │               │       └── Probe
-│       │       │               │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │       │               │           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
 │       │       │               └── Probe
 │       │       │                   └── HashJoin: INNER
 │       │       │                       ├── Build
-│       │       │                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │                       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 │       │       │                       └── Probe
-│       │       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │                           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │           └── Scan: default.tpcds.item (#7) (read rows: 180)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── Scan: default.tpcds.catalog_sales (#2) (read rows: 14313)
 └── Probe
-    └── Scan: default.tpcds.date_dim (read rows: 73049)
+    └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 
 
 # Q30
@@ -3178,35 +3175,35 @@ ORDER BY c_customer_id NULLS FIRST,
          ctr_total_return NULLS FIRST
 LIMIT 100;
 ----
-HashJoin: RIGHT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │   └── HashJoin: INNER
+│       │       ├── Build
+│       │       │   └── HashJoin: INNER
+│       │       │       ├── Build
+│       │       │       │   └── HashJoin: INNER
+│       │       │       │       ├── Build
+│       │       │       │       │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
+│       │       │       │       └── Probe
+│       │       │       │           └── Scan: default.tpcds.web_returns (#5) (read rows: 679)
+│       │       │       └── Probe
+│       │       │           └── Scan: default.tpcds.customer_address (#7) (read rows: 500)
+│       │       └── Probe
+│       │           └── HashJoin: INNER
+│       │               ├── Build
+│       │               │   └── HashJoin: INNER
+│       │               │       ├── Build
+│       │               │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
+│       │               │       └── Probe
+│       │               │           └── Scan: default.tpcds.web_returns (#0) (read rows: 679)
+│       │               └── Probe
+│       │                   └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 │       └── Probe
-│           └── HashJoin: INNER
-│               ├── Build
-│               │   └── HashJoin: INNER
-│               │       ├── Build
-│               │       │   └── HashJoin: INNER
-│               │       │       ├── Build
-│               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-│               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.web_returns (read rows: 679)
-│               │       └── Probe
-│               │           └── Scan: default.tpcds.customer_address (read rows: 500)
-│               └── Probe
-│                   └── Scan: default.tpcds.customer (read rows: 1000)
+│           └── Scan: default.tpcds.customer (#4) (read rows: 1000)
 └── Probe
-    └── HashJoin: INNER
-        ├── Build
-        │   └── HashJoin: INNER
-        │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       └── Probe
-        │           └── Scan: default.tpcds.web_returns (read rows: 679)
-        └── Probe
-            └── Scan: default.tpcds.customer_address (read rows: 500)
+    └── Scan: default.tpcds.customer_address (#3) (read rows: 500)
 
 # Q31
 query I
@@ -3296,61 +3293,61 @@ HashJoin: INNER
 │       │       │       │       │       ├── Build
 │       │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       │       ├── Build
-│       │       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#16) (read rows: 73049)
 │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │       │       │       │           └── Scan: default.tpcds.web_sales (#15) (read rows: 7212)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │       │           └── Scan: default.tpcds.customer_address (#17) (read rows: 500)
 │       │       │       │       └── Probe
 │       │       │       │           └── HashJoin: INNER
 │       │       │       │               ├── Build
 │       │       │       │               │   └── HashJoin: INNER
 │       │       │       │               │       ├── Build
-│       │       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │               │       │   └── Scan: default.tpcds.date_dim (#13) (read rows: 73049)
 │       │       │       │               │       └── Probe
-│       │       │       │               │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │       │               │           └── Scan: default.tpcds.web_sales (#12) (read rows: 7212)
 │       │       │       │               └── Probe
-│       │       │       │                   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │                   └── Scan: default.tpcds.customer_address (#14) (read rows: 500)
 │       │       │       └── Probe
 │       │       │           └── HashJoin: INNER
 │       │       │               ├── Build
 │       │       │               │   └── HashJoin: INNER
 │       │       │               │       ├── Build
-│       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │               │       │   └── Scan: default.tpcds.date_dim (#10) (read rows: 73049)
 │       │       │               │       └── Probe
-│       │       │               │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │               │           └── Scan: default.tpcds.web_sales (#9) (read rows: 7212)
 │       │       │               └── Probe
-│       │       │                   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │                   └── Scan: default.tpcds.customer_address (#11) (read rows: 500)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
 │       │               │   └── HashJoin: INNER
 │       │               │       ├── Build
-│       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │       │               │       └── Probe
-│       │               │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │               │           └── Scan: default.tpcds.store_sales (#6) (read rows: 28810)
 │       │               └── Probe
-│       │                   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │                   └── Scan: default.tpcds.customer_address (#8) (read rows: 500)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│               │           └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
 │               └── Probe
-│                   └── Scan: default.tpcds.customer_address (read rows: 500)
+│                   └── Scan: default.tpcds.customer_address (#5) (read rows: 500)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
         └── Probe
-            └── Scan: default.tpcds.customer_address (read rows: 500)
+            └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 
 # Q32
 query I
@@ -3372,23 +3369,23 @@ WHERE i_manufact_id = 977
        AND d_date_sk = cs_sold_date_sk )
 LIMIT 100;
 ----
-HashJoin: RIGHT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │   └── HashJoin: INNER
+│       │       │       ├── Build
+│       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
+│       │       │       └── Probe
+│       │       │           └── Scan: default.tpcds.catalog_sales (#3) (read rows: 14313)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │           └── Scan: default.tpcds.item (#1) (read rows: 0)
 │       └── Probe
-│           └── Scan: default.tpcds.item (read rows: 0)
+│           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 └── Probe
-    └── HashJoin: INNER
-        ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        └── Probe
-            └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+    └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 
 # Q33
 query I
@@ -3472,15 +3469,15 @@ UnionAll
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (#2) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │           └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │       │           └── Scan: default.tpcds.item (#3) (read rows: 180)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │           └── Scan: default.tpcds.item (#4) (read rows: 180)
 │       └── Right
 │           └── HashJoin: RIGHT SEMI
 │               ├── Build
@@ -3490,15 +3487,15 @@ UnionAll
 │               │       │       ├── Build
 │               │       │       │   └── HashJoin: INNER
 │               │       │       │       ├── Build
-│               │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│               │       │       │       │   └── Scan: default.tpcds.customer_address (#7) (read rows: 0)
 │               │       │       │       └── Probe
-│               │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │       │       │           └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │           └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.item (read rows: 180)
+│               │           └── Scan: default.tpcds.item (#8) (read rows: 180)
 │               └── Probe
-│                   └── Scan: default.tpcds.item (read rows: 180)
+│                   └── Scan: default.tpcds.item (#9) (read rows: 180)
 └── Right
     └── HashJoin: RIGHT SEMI
         ├── Build
@@ -3508,15 +3505,15 @@ UnionAll
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+        │       │       │       │   └── Scan: default.tpcds.customer_address (#12) (read rows: 0)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │       │           └── Scan: default.tpcds.web_sales (#10) (read rows: 7212)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │           └── Scan: default.tpcds.date_dim (#11) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.item (read rows: 180)
+        │           └── Scan: default.tpcds.item (#13) (read rows: 180)
         └── Probe
-            └── Scan: default.tpcds.item (read rows: 180)
+            └── Scan: default.tpcds.item (#14) (read rows: 180)
 
 # Q34
 query I
@@ -3566,19 +3563,19 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │   └── Scan: default.tpcds.store (#2) (read rows: 1)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│               │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │               └── Probe
-│                   └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│                   └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
 └── Probe
-    └── Scan: default.tpcds.customer (read rows: 1000)
+    └── Scan: default.tpcds.customer (#4) (read rows: 1000)
 
 # Q35
 onlyif todo
@@ -3726,45 +3723,45 @@ UnionAll
 │       ├── Left
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │       │   └── Scan: default.tpcds.store (#3) (read rows: 1)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.item (read rows: 180)
+│       │               │   └── Scan: default.tpcds.item (#2) (read rows: 180)
 │       │               └── Probe
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
-│       │                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │                       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       └── Right
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.store (read rows: 1)
+│               │   └── Scan: default.tpcds.store (#7) (read rows: 1)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.item (read rows: 180)
+│                       │   └── Scan: default.tpcds.item (#6) (read rows: 180)
 │                       └── Probe
 │                           └── HashJoin: INNER
 │                               ├── Build
-│                               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                               │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 │                               └── Probe
-│                                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                                   └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
 └── Right
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.store (read rows: 1)
+        │   └── Scan: default.tpcds.store (#11) (read rows: 1)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.item (read rows: 180)
+                │   └── Scan: default.tpcds.item (#10) (read rows: 180)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
-                        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                        │   └── Scan: default.tpcds.date_dim (#9) (read rows: 73049)
                         └── Probe
-                            └── Scan: default.tpcds.store_sales (read rows: 28810)
+                            └── Scan: default.tpcds.store_sales (#8) (read rows: 28810)
 
 # Q37
 query I
@@ -3796,15 +3793,15 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │   └── Scan: default.tpcds.item (#0) (read rows: 180)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │               └── Probe
-│                   └── Scan: default.tpcds.inventory (read rows: 23490)
+│                   └── Scan: default.tpcds.inventory (#1) (read rows: 23490)
 └── Probe
-    └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+    └── Scan: default.tpcds.catalog_sales (#3) (read rows: 14313)
 
 # Q38
 query I
@@ -3846,31 +3843,31 @@ HashJoin: INNER
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │       │           └── Scan: default.tpcds.catalog_sales (#3) (read rows: 14313)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │           └── Scan: default.tpcds.customer (#5) (read rows: 1000)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│               │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │               └── Probe
-│                   └── Scan: default.tpcds.customer (read rows: 1000)
+│                   └── Scan: default.tpcds.customer (#2) (read rows: 1000)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │           └── Scan: default.tpcds.web_sales (#6) (read rows: 7212)
         └── Probe
-            └── Scan: default.tpcds.customer (read rows: 1000)
+            └── Scan: default.tpcds.customer (#8) (read rows: 1000)
 
 # Q39
 query I
@@ -3938,31 +3935,31 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.warehouse (read rows: 1)
+│       │   └── Scan: default.tpcds.warehouse (#6) (read rows: 1)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │   └── Scan: default.tpcds.item (#5) (read rows: 180)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │                       └── Probe
-│                           └── Scan: default.tpcds.inventory (read rows: 23490)
+│                           └── Scan: default.tpcds.inventory (#4) (read rows: 23490)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.warehouse (read rows: 1)
+        │   └── Scan: default.tpcds.warehouse (#2) (read rows: 1)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.item (read rows: 180)
+                │   └── Scan: default.tpcds.item (#1) (read rows: 180)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
-                        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                        │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
                         └── Probe
-                            └── Scan: default.tpcds.inventory (read rows: 23490)
+                            └── Scan: default.tpcds.inventory (#0) (read rows: 23490)
 
 # Q40
 query I
@@ -3997,19 +3994,19 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │   └── Scan: default.tpcds.item (#3) (read rows: 180)
 │               └── Probe
 │                   └── HashJoin: LEFT OUTER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│                       │   └── Scan: default.tpcds.catalog_returns (#1) (read rows: 1358)
 │                       └── Probe
-│                           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│                           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 └── Probe
-    └── Scan: default.tpcds.warehouse (read rows: 1)
+    └── Scan: default.tpcds.warehouse (#2) (read rows: 1)
 
 # Q41
 query I
@@ -4081,15 +4078,15 @@ WHERE i_manufact_id BETWEEN 738 AND 738+40
 ORDER BY i_product_name
 LIMIT 100;
 ----
-HashJoin: LEFT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │   └── Scan: default.tpcds.item (#0) (read rows: 180)
 │       └── Probe
-│           └── Scan: default.tpcds.item (read rows: 180)
+│           └── Scan: default.tpcds.item (#1) (read rows: 180)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#0) (read rows: 180)
 
 # Q42
 query I
@@ -4116,13 +4113,13 @@ LIMIT 100 ;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.item (read rows: 180)
+│   └── Scan: default.tpcds.item (#2) (read rows: 180)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#0) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q43
 query I
@@ -4179,13 +4176,13 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.store (read rows: 0)
+│   └── Scan: default.tpcds.store (#2) (read rows: 0)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#0) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q44
 query I
@@ -4246,19 +4243,19 @@ HashJoin: INNER
 │       │       ├── Build
 │       │       │   └── RangeJoin: INNER
 │       │       │       ├── Left
-│       │       │       │   └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │       │   └── Scan: default.tpcds.store_sales (#2) (read rows: 0)
 │       │       │       └── Right
-│       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │           └── Scan: default.tpcds.store_sales (#3) (read rows: 0)
 │       │       └── Probe
 │       │           └── RangeJoin: INNER
 │       │               ├── Left
-│       │               │   └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │               │   └── Scan: default.tpcds.store_sales (#0) (read rows: 0)
 │       │               └── Right
-│       │                   └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │                   └── Scan: default.tpcds.store_sales (#1) (read rows: 0)
 │       └── Probe
-│           └── Scan: default.tpcds.item (read rows: 180)
+│           └── Scan: default.tpcds.item (#4) (read rows: 180)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#5) (read rows: 180)
 
 # Q45
 query I
@@ -4307,7 +4304,7 @@ LIMIT 100;
 ----
 HashJoin: RIGHT MARK
 ├── Build
-│   └── Scan: default.tpcds.item (read rows: 180)
+│   └── Scan: default.tpcds.item (#5) (read rows: 180)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -4317,15 +4314,15 @@ HashJoin: RIGHT MARK
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │       │           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │       │           └── Scan: default.tpcds.customer (#1) (read rows: 1000)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer_address (read rows: 500)
+        │           └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
         └── Probe
-            └── Scan: default.tpcds.item (read rows: 180)
+            └── Scan: default.tpcds.item (#4) (read rows: 180)
 
 # Q46
 query I
@@ -4379,29 +4376,29 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.customer_address (read rows: 500)
+│   └── Scan: default.tpcds.customer_address (#6) (read rows: 500)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
+        │       │   └── Scan: default.tpcds.customer_address (#4) (read rows: 500)
         │       └── Probe
         │           └── HashJoin: INNER
         │               ├── Build
-        │               │   └── Scan: default.tpcds.store (read rows: 1)
+        │               │   └── Scan: default.tpcds.store (#2) (read rows: 1)
         │               └── Probe
         │                   └── HashJoin: INNER
         │                       ├── Build
         │                       │   └── HashJoin: INNER
         │                       │       ├── Build
-        │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │                       │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
         │                       │       └── Probe
-        │                       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │                       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
         │                       └── Probe
-        │                           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+        │                           └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
         └── Probe
-            └── Scan: default.tpcds.customer (read rows: 1000)
+            └── Scan: default.tpcds.customer (#5) (read rows: 1000)
 
 # Q47
 query I
@@ -4484,45 +4481,45 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │   └── Scan: default.tpcds.item (#0) (read rows: 180)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.store (read rows: 1)
+│       │               │   └── Scan: default.tpcds.store (#3) (read rows: 1)
 │       │               └── Probe
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
-│       │                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │                       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                           └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │   └── Scan: default.tpcds.item (#4) (read rows: 180)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.store (read rows: 1)
+│                       │   └── Scan: default.tpcds.store (#7) (read rows: 1)
 │                       └── Probe
 │                           └── HashJoin: INNER
 │                               ├── Build
-│                               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                               │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │                               └── Probe
-│                                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                                   └── Scan: default.tpcds.store_sales (#5) (read rows: 28810)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#8) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.store (read rows: 1)
+                │   └── Scan: default.tpcds.store (#11) (read rows: 1)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
-                        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                        │   └── Scan: default.tpcds.date_dim (#10) (read rows: 73049)
                         └── Probe
-                            └── Scan: default.tpcds.store_sales (read rows: 28810)
+                            └── Scan: default.tpcds.store_sales (#9) (read rows: 28810)
 
 # Q48
 query I
@@ -4575,15 +4572,15 @@ HashJoin: INNER
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │           └── Scan: default.tpcds.customer_address (#3) (read rows: 500)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │           └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│           └── Scan: default.tpcds.customer_demographics (#2) (read rows: 19208)
 └── Probe
-    └── Scan: default.tpcds.store (read rows: 1)
+    └── Scan: default.tpcds.store (#1) (read rows: 1)
 
 # Q49
 query I
@@ -4696,33 +4693,33 @@ UnionAll
 │       ├── Left
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── HashJoin: RIGHT OUTER
+│       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │       │   └── Scan: default.tpcds.web_returns (#1) (read rows: 679)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.web_returns (read rows: 679)
+│       │       │           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │           └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       └── Right
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── HashJoin: RIGHT OUTER
+│               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │       │   └── Scan: default.tpcds.catalog_returns (#4) (read rows: 1358)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│               │           └── Scan: default.tpcds.catalog_sales (#3) (read rows: 14313)
 │               └── Probe
-│                   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 └── Right
     └── HashJoin: INNER
         ├── Build
-        │   └── HashJoin: RIGHT OUTER
-        │       ├── Build
-        │       │   └── Scan: default.tpcds.store_sales (read rows: 28810)
-        │       └── Probe
-        │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+        │   └── Scan: default.tpcds.store_returns (#7) (read rows: 2810)
         └── Probe
-            └── Scan: default.tpcds.date_dim (read rows: 73049)
+            └── HashJoin: INNER
+                ├── Build
+                │   └── Scan: default.tpcds.store_sales (#6) (read rows: 28810)
+                └── Probe
+                    └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 
 # Q50
 query I
@@ -4799,19 +4796,19 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │   └── Scan: default.tpcds.store (#2) (read rows: 1)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│               │           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
 │               └── Probe
-│                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 └── Probe
-    └── Scan: default.tpcds.date_dim (read rows: 73049)
+    └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 
 # Q51
 query I
@@ -4873,15 +4870,15 @@ HashJoin: FULL OUTER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│           └── Scan: default.tpcds.store_sales (#2) (read rows: 28810)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.web_sales (read rows: 7212)
+            └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 
 # Q52
 query I
@@ -4908,13 +4905,13 @@ LIMIT 100 ;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.item (read rows: 180)
+│   └── Scan: default.tpcds.item (#2) (read rows: 180)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#0) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q53
 query I
@@ -4969,17 +4966,17 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.store (read rows: 1)
+│   └── Scan: default.tpcds.store (#3) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#0) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
                 └── Probe
-                    └── Scan: default.tpcds.store_sales (read rows: 28810)
+                    └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q54
 query I
@@ -5044,11 +5041,11 @@ LIMIT 100;
 ----
 RangeJoin: INNER
 ├── Left
-│   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│   └── Scan: default.tpcds.date_dim (#10) (read rows: 73049)
 └── Right
     └── RangeJoin: INNER
         ├── Left
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#9) (read rows: 73049)
         └── Right
             └── HashJoin: INNER
                 ├── Build
@@ -5058,31 +5055,31 @@ RangeJoin: INNER
                 │       │       ├── Build
                 │       │       │   └── HashJoin: INNER
                 │       │       │       ├── Build
-                │       │       │       │   └── Scan: default.tpcds.store (read rows: 1)
+                │       │       │       │   └── Scan: default.tpcds.store (#7) (read rows: 1)
                 │       │       │       └── Probe
-                │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+                │       │       │           └── Scan: default.tpcds.customer_address (#6) (read rows: 500)
                 │       │       └── Probe
                 │       │           └── HashJoin: INNER
                 │       │               ├── Build
-                │       │               │   └── Scan: default.tpcds.customer (read rows: 1000)
+                │       │               │   └── Scan: default.tpcds.customer (#4) (read rows: 1000)
                 │       │               └── Probe
                 │       │                   └── HashJoin: INNER
                 │       │                       ├── Build
-                │       │                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │       │                       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
                 │       │                       └── Probe
                 │       │                           └── HashJoin: INNER
                 │       │                               ├── Build
-                │       │                               │   └── Scan: default.tpcds.item (read rows: 180)
+                │       │                               │   └── Scan: default.tpcds.item (#2) (read rows: 180)
                 │       │                               └── Probe
                 │       │                                   └── UnionAll
                 │       │                                       ├── Left
-                │       │                                       │   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+                │       │                                       │   └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
                 │       │                                       └── Right
-                │       │                                           └── Scan: default.tpcds.web_sales (read rows: 7212)
+                │       │                                           └── Scan: default.tpcds.web_sales (#1) (read rows: 7212)
                 │       └── Probe
-                │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                │           └── Scan: default.tpcds.store_sales (#5) (read rows: 28810)
                 └── Probe
-                    └── Scan: default.tpcds.date_dim (read rows: 73049)
+                    └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 
 # Q55
 query I
@@ -5106,13 +5103,13 @@ LIMIT 100 ;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.item (read rows: 180)
+│   └── Scan: default.tpcds.item (#2) (read rows: 180)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#0) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q56
 query I
@@ -5203,15 +5200,15 @@ UnionAll
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (#2) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │           └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │       │           └── Scan: default.tpcds.item (#3) (read rows: 180)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │           └── Scan: default.tpcds.item (#4) (read rows: 180)
 │       └── Right
 │           └── HashJoin: RIGHT SEMI
 │               ├── Build
@@ -5221,15 +5218,15 @@ UnionAll
 │               │       │       ├── Build
 │               │       │       │   └── HashJoin: INNER
 │               │       │       │       ├── Build
-│               │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│               │       │       │       │   └── Scan: default.tpcds.customer_address (#7) (read rows: 0)
 │               │       │       │       └── Probe
-│               │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │       │       │           └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │           └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.item (read rows: 180)
+│               │           └── Scan: default.tpcds.item (#8) (read rows: 180)
 │               └── Probe
-│                   └── Scan: default.tpcds.item (read rows: 180)
+│                   └── Scan: default.tpcds.item (#9) (read rows: 180)
 └── Right
     └── HashJoin: RIGHT SEMI
         ├── Build
@@ -5239,15 +5236,15 @@ UnionAll
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+        │       │       │       │   └── Scan: default.tpcds.customer_address (#12) (read rows: 0)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │       │           └── Scan: default.tpcds.web_sales (#10) (read rows: 7212)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │           └── Scan: default.tpcds.date_dim (#11) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.item (read rows: 180)
+        │           └── Scan: default.tpcds.item (#13) (read rows: 180)
         └── Probe
-            └── Scan: default.tpcds.item (read rows: 180)
+            └── Scan: default.tpcds.item (#14) (read rows: 180)
 
 # Q57
 query I
@@ -5323,45 +5320,45 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │   └── Scan: default.tpcds.item (#0) (read rows: 180)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.call_center (read rows: 1)
+│       │               │   └── Scan: default.tpcds.call_center (#3) (read rows: 1)
 │       │               └── Probe
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
-│       │                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │                       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │                           └── Scan: default.tpcds.catalog_sales (#1) (read rows: 14313)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │   └── Scan: default.tpcds.item (#4) (read rows: 180)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.call_center (read rows: 1)
+│                       │   └── Scan: default.tpcds.call_center (#7) (read rows: 1)
 │                       └── Probe
 │                           └── HashJoin: INNER
 │                               ├── Build
-│                               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                               │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │                               └── Probe
-│                                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│                                   └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#8) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.call_center (read rows: 1)
+                │   └── Scan: default.tpcds.call_center (#11) (read rows: 1)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
-                        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                        │   └── Scan: default.tpcds.date_dim (#10) (read rows: 73049)
                         └── Probe
-                            └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+                            └── Scan: default.tpcds.catalog_sales (#9) (read rows: 14313)
 
 # Q58
 query I
@@ -5439,61 +5436,61 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── HashJoin: RIGHT SEMI
+│   └── HashJoin: INNER
 │       ├── Build
-│       │   └── HashJoin: INNER
+│       │   └── HashJoin: RIGHT SEMI
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │       │   └── HashJoin: INNER
+│       │       │       │       ├── Build
+│       │       │       │       │   └── Scan: default.tpcds.item (#11) (read rows: 180)
+│       │       │       │       └── Probe
+│       │       │       │           └── Scan: default.tpcds.web_sales (#10) (read rows: 7212)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │           └── Scan: default.tpcds.date_dim (#12) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │           └── HashJoin: INNER
+│       │               ├── Build
+│       │               │   └── Scan: default.tpcds.date_dim (#14) (read rows: 73049)
+│       │               └── Probe
+│       │                   └── Scan: default.tpcds.date_dim (#13) (read rows: 73049)
 │       └── Probe
-│           └── HashJoin: INNER
+│           └── HashJoin: RIGHT SEMI
 │               ├── Build
-│               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │   └── HashJoin: INNER
+│               │       ├── Build
+│               │       │   └── HashJoin: INNER
+│               │       │       ├── Build
+│               │       │       │   └── Scan: default.tpcds.item (#6) (read rows: 180)
+│               │       │       └── Probe
+│               │       │           └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
+│               │       └── Probe
+│               │           └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │               └── Probe
-│                   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                   └── HashJoin: INNER
+│                       ├── Build
+│                       │   └── Scan: default.tpcds.date_dim (#9) (read rows: 73049)
+│                       └── Probe
+│                           └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 └── Probe
-    └── HashJoin: INNER
+    └── HashJoin: RIGHT SEMI
         ├── Build
-        │   └── HashJoin: RIGHT SEMI
+        │   └── HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── HashJoin: INNER
-        │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
-        │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+        │       │       │   └── Scan: default.tpcds.item (#1) (read rows: 180)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
         │       └── Probe
-        │           └── HashJoin: INNER
-        │               ├── Build
-        │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │               └── Probe
-        │                   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │           └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
         └── Probe
-            └── HashJoin: RIGHT SEMI
+            └── HashJoin: INNER
                 ├── Build
-                │   └── HashJoin: INNER
-                │       ├── Build
-                │       │   └── HashJoin: INNER
-                │       │       ├── Build
-                │       │       │   └── Scan: default.tpcds.item (read rows: 180)
-                │       │       └── Probe
-                │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
-                │       └── Probe
-                │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
                 └── Probe
-                    └── HashJoin: INNER
-                        ├── Build
-                        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-                        └── Probe
-                            └── Scan: default.tpcds.date_dim (read rows: 73049)
+                    └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 
 # Q59
 query I
@@ -5588,31 +5585,31 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │   └── Scan: default.tpcds.store (#6) (read rows: 1)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                       │   └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
 │                       └── Probe
-│                           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                           └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.store (read rows: 1)
+        │   └── Scan: default.tpcds.store (#2) (read rows: 1)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
                 └── Probe
                     └── HashJoin: INNER
                         ├── Build
-                        │   └── Scan: default.tpcds.store_sales (read rows: 28810)
+                        │   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
                         └── Probe
-                            └── Scan: default.tpcds.date_dim (read rows: 73049)
+                            └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 
 # Q60
 query I
@@ -5697,15 +5694,15 @@ UnionAll
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (#2) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │           └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │       │           └── Scan: default.tpcds.item (#3) (read rows: 180)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │           └── Scan: default.tpcds.item (#4) (read rows: 180)
 │       └── Right
 │           └── HashJoin: RIGHT SEMI
 │               ├── Build
@@ -5715,15 +5712,15 @@ UnionAll
 │               │       │       ├── Build
 │               │       │       │   └── HashJoin: INNER
 │               │       │       │       ├── Build
-│               │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│               │       │       │       │   └── Scan: default.tpcds.customer_address (#7) (read rows: 0)
 │               │       │       │       └── Probe
-│               │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │       │       │           └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │           └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.item (read rows: 180)
+│               │           └── Scan: default.tpcds.item (#8) (read rows: 180)
 │               └── Probe
-│                   └── Scan: default.tpcds.item (read rows: 180)
+│                   └── Scan: default.tpcds.item (#9) (read rows: 180)
 └── Right
     └── HashJoin: RIGHT SEMI
         ├── Build
@@ -5733,15 +5730,15 @@ UnionAll
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+        │       │       │       │   └── Scan: default.tpcds.customer_address (#12) (read rows: 0)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │       │           └── Scan: default.tpcds.web_sales (#10) (read rows: 7212)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │           └── Scan: default.tpcds.date_dim (#11) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.item (read rows: 180)
+        │           └── Scan: default.tpcds.item (#13) (read rows: 180)
         └── Probe
-            └── Scan: default.tpcds.item (read rows: 180)
+            └── Scan: default.tpcds.item (#14) (read rows: 180)
 
 # Q61
 query I
@@ -5803,19 +5800,19 @@ HashJoin: CROSS
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│       │       │       │       │   └── Scan: default.tpcds.customer_address (#11) (read rows: 0)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │       │           └── Scan: default.tpcds.customer (#10) (read rows: 1000)
 │       │       │       └── Probe
 │       │       │           └── HashJoin: INNER
 │       │       │               ├── Build
-│       │       │               │   └── Scan: default.tpcds.store (read rows: 0)
+│       │       │               │   └── Scan: default.tpcds.store (#8) (read rows: 0)
 │       │       │               └── Probe
-│       │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │                   └── Scan: default.tpcds.store_sales (#7) (read rows: 28810)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.item (read rows: 180)
+│       │           └── Scan: default.tpcds.item (#12) (read rows: 180)
 │       └── Probe
-│           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│           └── Scan: default.tpcds.date_dim (#9) (read rows: 73049)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -5827,21 +5824,21 @@ HashJoin: CROSS
         │       │       │       ├── Build
         │       │       │       │   └── HashJoin: INNER
         │       │       │       │       ├── Build
-        │       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+        │       │       │       │       │   └── Scan: default.tpcds.customer_address (#5) (read rows: 0)
         │       │       │       │       └── Probe
-        │       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │       │       │       │           └── Scan: default.tpcds.customer (#4) (read rows: 1000)
         │       │       │       └── Probe
         │       │       │           └── HashJoin: INNER
         │       │       │               ├── Build
-        │       │       │               │   └── Scan: default.tpcds.store (read rows: 0)
+        │       │       │               │   └── Scan: default.tpcds.store (#1) (read rows: 0)
         │       │       │               └── Probe
-        │       │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │       │                   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.item (read rows: 180)
+        │       │           └── Scan: default.tpcds.item (#6) (read rows: 180)
         │       └── Probe
-        │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │           └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.promotion (read rows: 3)
+            └── Scan: default.tpcds.promotion (#2) (read rows: 3)
 
 # Q62
 query I
@@ -5894,21 +5891,21 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.web_site (read rows: 1)
+│   └── Scan: default.tpcds.web_site (#3) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.warehouse (read rows: 1)
+        │       │   └── Scan: default.tpcds.warehouse (#1) (read rows: 1)
         │       └── Probe
         │           └── HashJoin: INNER
         │               ├── Build
-        │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │               │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
         │               └── Probe
-        │                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │                   └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
         └── Probe
-            └── Scan: default.tpcds.ship_mode (read rows: 20)
+            └── Scan: default.tpcds.ship_mode (#2) (read rows: 20)
 
 # Q63
 query I
@@ -5963,17 +5960,17 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.store (read rows: 1)
+│   └── Scan: default.tpcds.store (#3) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#0) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
                 └── Probe
-                    └── Scan: default.tpcds.store_sales (read rows: 28810)
+                    └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q64
 query I
@@ -6119,65 +6116,65 @@ HashJoin: INNER
 │       │       │       │       │       │       ├── Build
 │       │       │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       │       │       ├── Build
-│       │       │       │       │       │       │       │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │       ├── Build
-│       │       │       │       │       │       │       │       │   └── Scan: default.tpcds.promotion (read rows: 3)
-│       │       │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │       │       │           └── HashJoin: INNER
-│       │       │       │       │       │       │       │               ├── Build
-│       │       │       │       │       │       │       │               │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       ├── Build
-│       │       │       │       │       │       │       │               │       │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       ├── Build
-│       │       │       │       │       │       │       │               │       │       │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       │       ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │       │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       │       │       ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │       │       │   └── Scan: default.tpcds.store (read rows: 1)
-│       │       │       │       │       │       │       │               │       │       │       │       │       └── Probe
-│       │       │       │       │       │       │       │               │       │       │       │       │           └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       │       │               ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │       │               │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       │   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       │       ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       │       │   └── Scan: default.tpcds.item (read rows: 180)
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       │       └── Probe
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
-│       │       │       │       │       │       │       │               │       │       │       │       │               │       └── Probe
-│       │       │       │       │       │       │       │               │       │       │       │       │               │           └── Scan: default.tpcds.date_dim (read rows: 73049)
-│       │       │       │       │       │       │       │               │       │       │       │       │               └── Probe
-│       │       │       │       │       │       │       │               │       │       │       │       │                   └── HashJoin: INNER
-│       │       │       │       │       │       │       │               │       │       │       │       │                       ├── Build
-│       │       │       │       │       │       │       │               │       │       │       │       │                       │   └── Scan: default.tpcds.catalog_returns (read rows: 1358)
-│       │       │       │       │       │       │       │               │       │       │       │       │                       └── Probe
-│       │       │       │       │       │       │       │               │       │       │       │       │                           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
-│       │       │       │       │       │       │       │               │       │       │       │       └── Probe
-│       │       │       │       │       │       │       │               │       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
-│       │       │       │       │       │       │       │               │       │       │       └── Probe
-│       │       │       │       │       │       │       │               │       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
-│       │       │       │       │       │       │       │               │       │       └── Probe
-│       │       │       │       │       │       │       │               │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
-│       │       │       │       │       │       │       │               │       └── Probe
-│       │       │       │       │       │       │       │               │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
-│       │       │       │       │       │       │       │               └── Probe
-│       │       │       │       │       │       │       │                   └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │       │       │       │       │       │       │   └── Scan: default.tpcds.promotion (#30) (read rows: 3)
 │       │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│       │       │       │       │       │       │           └── HashJoin: INNER
+│       │       │       │       │       │       │               ├── Build
+│       │       │       │       │       │       │               │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       ├── Build
+│       │       │       │       │       │       │               │       │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       ├── Build
+│       │       │       │       │       │       │               │       │       │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       ├── Build
+│       │       │       │       │       │       │               │       │       │       │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       │       ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │       │   └── Scan: default.tpcds.store (#26) (read rows: 1)
+│       │       │       │       │       │       │               │       │       │       │       │       └── Probe
+│       │       │       │       │       │       │               │       │       │       │       │           └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       │               ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │               │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       │               │       ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │               │       │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │   └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │       ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │       │   └── Scan: default.tpcds.item (#37) (read rows: 180)
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │       └── Probe
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │           └── HashJoin: INNER
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │               ├── Build
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │               │   └── Scan: default.tpcds.catalog_returns (#22) (read rows: 1358)
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │               └── Probe
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       │                   └── Scan: default.tpcds.catalog_sales (#21) (read rows: 14313)
+│       │       │       │       │       │       │               │       │       │       │       │               │       │       └── Probe
+│       │       │       │       │       │       │               │       │       │       │       │               │       │           └── Scan: default.tpcds.store_returns (#20) (read rows: 2810)
+│       │       │       │       │       │       │               │       │       │       │       │               │       └── Probe
+│       │       │       │       │       │       │               │       │       │       │       │               │           └── Scan: default.tpcds.store_sales (#19) (read rows: 28810)
+│       │       │       │       │       │       │               │       │       │       │       │               └── Probe
+│       │       │       │       │       │       │               │       │       │       │       │                   └── Scan: default.tpcds.date_dim (#23) (read rows: 73049)
+│       │       │       │       │       │       │               │       │       │       │       └── Probe
+│       │       │       │       │       │       │               │       │       │       │           └── Scan: default.tpcds.customer (#27) (read rows: 1000)
+│       │       │       │       │       │       │               │       │       │       └── Probe
+│       │       │       │       │       │       │               │       │       │           └── Scan: default.tpcds.date_dim (#24) (read rows: 73049)
+│       │       │       │       │       │       │               │       │       └── Probe
+│       │       │       │       │       │       │               │       │           └── Scan: default.tpcds.date_dim (#25) (read rows: 73049)
+│       │       │       │       │       │       │               │       └── Probe
+│       │       │       │       │       │       │               │           └── Scan: default.tpcds.customer_demographics (#28) (read rows: 19208)
+│       │       │       │       │       │       │               └── Probe
+│       │       │       │       │       │       │                   └── Scan: default.tpcds.customer_demographics (#29) (read rows: 19208)
 │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│       │       │       │       │       │           └── Scan: default.tpcds.household_demographics (#31) (read rows: 7200)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │       │           └── Scan: default.tpcds.household_demographics (#32) (read rows: 7200)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │           └── Scan: default.tpcds.customer_address (#33) (read rows: 500)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.income_band (read rows: 20)
+│       │       │           └── Scan: default.tpcds.customer_address (#34) (read rows: 500)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.income_band (read rows: 20)
+│       │           └── Scan: default.tpcds.income_band (#35) (read rows: 20)
 │       └── Probe
-│           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│           └── Scan: default.tpcds.income_band (#36) (read rows: 20)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -6193,65 +6190,65 @@ HashJoin: INNER
         │       │       │       │       │       ├── Build
         │       │       │       │       │       │   └── HashJoin: INNER
         │       │       │       │       │       │       ├── Build
-        │       │       │       │       │       │       │   └── HashJoin: INNER
-        │       │       │       │       │       │       │       ├── Build
-        │       │       │       │       │       │       │       │   └── Scan: default.tpcds.promotion (read rows: 3)
-        │       │       │       │       │       │       │       └── Probe
-        │       │       │       │       │       │       │           └── HashJoin: INNER
-        │       │       │       │       │       │       │               ├── Build
-        │       │       │       │       │       │       │               │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       ├── Build
-        │       │       │       │       │       │       │               │       │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       ├── Build
-        │       │       │       │       │       │       │               │       │       │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       ├── Build
-        │       │       │       │       │       │       │               │       │       │       │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       │       ├── Build
-        │       │       │       │       │       │       │               │       │       │       │       │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       │       │       ├── Build
-        │       │       │       │       │       │       │               │       │       │       │       │       │   └── Scan: default.tpcds.store (read rows: 1)
-        │       │       │       │       │       │       │               │       │       │       │       │       └── Probe
-        │       │       │       │       │       │       │               │       │       │       │       │           └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       │       │               ├── Build
-        │       │       │       │       │       │       │               │       │       │       │       │               │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       │       │               │       ├── Build
-        │       │       │       │       │       │       │               │       │       │       │       │               │       │   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       │       │               │       │       ├── Build
-        │       │       │       │       │       │       │               │       │       │       │       │               │       │       │   └── Scan: default.tpcds.item (read rows: 180)
-        │       │       │       │       │       │       │               │       │       │       │       │               │       │       └── Probe
-        │       │       │       │       │       │       │               │       │       │       │       │               │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
-        │       │       │       │       │       │       │               │       │       │       │       │               │       └── Probe
-        │       │       │       │       │       │       │               │       │       │       │       │               │           └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       │       │       │       │       │       │               │       │       │       │       │               └── Probe
-        │       │       │       │       │       │       │               │       │       │       │       │                   └── HashJoin: INNER
-        │       │       │       │       │       │       │               │       │       │       │       │                       ├── Build
-        │       │       │       │       │       │       │               │       │       │       │       │                       │   └── Scan: default.tpcds.catalog_returns (read rows: 1358)
-        │       │       │       │       │       │       │               │       │       │       │       │                       └── Probe
-        │       │       │       │       │       │       │               │       │       │       │       │                           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
-        │       │       │       │       │       │       │               │       │       │       │       └── Probe
-        │       │       │       │       │       │       │               │       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
-        │       │       │       │       │       │       │               │       │       │       └── Probe
-        │       │       │       │       │       │       │               │       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       │       │       │       │       │       │               │       │       └── Probe
-        │       │       │       │       │       │       │               │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       │       │       │       │       │       │               │       └── Probe
-        │       │       │       │       │       │       │               │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
-        │       │       │       │       │       │       │               └── Probe
-        │       │       │       │       │       │       │                   └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+        │       │       │       │       │       │       │   └── Scan: default.tpcds.promotion (#11) (read rows: 3)
         │       │       │       │       │       │       └── Probe
-        │       │       │       │       │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+        │       │       │       │       │       │           └── HashJoin: INNER
+        │       │       │       │       │       │               ├── Build
+        │       │       │       │       │       │               │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       ├── Build
+        │       │       │       │       │       │               │       │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       ├── Build
+        │       │       │       │       │       │               │       │       │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       ├── Build
+        │       │       │       │       │       │               │       │       │       │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       ├── Build
+        │       │       │       │       │       │               │       │       │       │       │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       │       ├── Build
+        │       │       │       │       │       │               │       │       │       │       │       │   └── Scan: default.tpcds.store (#7) (read rows: 1)
+        │       │       │       │       │       │               │       │       │       │       │       └── Probe
+        │       │       │       │       │       │               │       │       │       │       │           └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       │               ├── Build
+        │       │       │       │       │       │               │       │       │       │       │               │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       │               │       ├── Build
+        │       │       │       │       │       │               │       │       │       │       │               │       │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       │               │       │       ├── Build
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │   └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │       ├── Build
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │       │   └── Scan: default.tpcds.item (#18) (read rows: 180)
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │       └── Probe
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │           └── HashJoin: INNER
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │               ├── Build
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │               │   └── Scan: default.tpcds.catalog_returns (#3) (read rows: 1358)
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │               └── Probe
+        │       │       │       │       │       │               │       │       │       │       │               │       │       │                   └── Scan: default.tpcds.catalog_sales (#2) (read rows: 14313)
+        │       │       │       │       │       │               │       │       │       │       │               │       │       └── Probe
+        │       │       │       │       │       │               │       │       │       │       │               │       │           └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
+        │       │       │       │       │       │               │       │       │       │       │               │       └── Probe
+        │       │       │       │       │       │               │       │       │       │       │               │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
+        │       │       │       │       │       │               │       │       │       │       │               └── Probe
+        │       │       │       │       │       │               │       │       │       │       │                   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
+        │       │       │       │       │       │               │       │       │       │       └── Probe
+        │       │       │       │       │       │               │       │       │       │           └── Scan: default.tpcds.customer (#8) (read rows: 1000)
+        │       │       │       │       │       │               │       │       │       └── Probe
+        │       │       │       │       │       │               │       │       │           └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
+        │       │       │       │       │       │               │       │       └── Probe
+        │       │       │       │       │       │               │       │           └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
+        │       │       │       │       │       │               │       └── Probe
+        │       │       │       │       │       │               │           └── Scan: default.tpcds.customer_demographics (#9) (read rows: 19208)
+        │       │       │       │       │       │               └── Probe
+        │       │       │       │       │       │                   └── Scan: default.tpcds.customer_demographics (#10) (read rows: 19208)
         │       │       │       │       │       └── Probe
-        │       │       │       │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+        │       │       │       │       │           └── Scan: default.tpcds.household_demographics (#12) (read rows: 7200)
         │       │       │       │       └── Probe
-        │       │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+        │       │       │       │           └── Scan: default.tpcds.household_demographics (#13) (read rows: 7200)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+        │       │       │           └── Scan: default.tpcds.customer_address (#14) (read rows: 500)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.income_band (read rows: 20)
+        │       │           └── Scan: default.tpcds.customer_address (#15) (read rows: 500)
         │       └── Probe
-        │           └── Scan: default.tpcds.income_band (read rows: 20)
+        │           └── Scan: default.tpcds.income_band (#16) (read rows: 20)
         └── Probe
-            └── Scan: default.tpcds.store_returns (read rows: 2810)
+            └── Scan: default.tpcds.income_band (#17) (read rows: 20)
 
 # Q65
 query I
@@ -6296,25 +6293,25 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.store (read rows: 1)
+│   └── HashJoin: INNER
+│       ├── Build
+│       │   └── HashJoin: INNER
+│       │       ├── Build
+│       │       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
+│       │       └── Probe
+│       │           └── Scan: default.tpcds.store_sales (#2) (read rows: 28810)
+│       └── Probe
+│           └── Scan: default.tpcds.store (#0) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── HashJoin: INNER
-        │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       │       └── Probe
-        │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
         │       └── Probe
-        │           └── HashJoin: INNER
-        │               ├── Build
-        │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │               └── Probe
-        │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │           └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
         └── Probe
-            └── Scan: default.tpcds.item (read rows: 180)
+            └── Scan: default.tpcds.item (#1) (read rows: 180)
 
 
 # Q66
@@ -6542,39 +6539,39 @@ UnionAll
 ├── Left
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.warehouse (read rows: 1)
+│       │   └── Scan: default.tpcds.warehouse (#1) (read rows: 1)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.ship_mode (read rows: 20)
+│               │       │   └── Scan: default.tpcds.ship_mode (#4) (read rows: 20)
 │               │       └── Probe
 │               │           └── HashJoin: INNER
 │               │               ├── Build
-│               │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │               │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │               │               └── Probe
-│               │                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│               │                   └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 │               └── Probe
-│                   └── Scan: default.tpcds.time_dim (read rows: 86400)
+│                   └── Scan: default.tpcds.time_dim (#3) (read rows: 86400)
 └── Right
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.warehouse (read rows: 1)
+        │   └── Scan: default.tpcds.warehouse (#6) (read rows: 1)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.ship_mode (read rows: 20)
+                │       │   └── Scan: default.tpcds.ship_mode (#9) (read rows: 20)
                 │       └── Probe
                 │           └── HashJoin: INNER
                 │               ├── Build
-                │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │               │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
                 │               └── Probe
-                │                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+                │                   └── Scan: default.tpcds.catalog_sales (#5) (read rows: 14313)
                 └── Probe
-                    └── Scan: default.tpcds.time_dim (read rows: 86400)
+                    └── Scan: default.tpcds.time_dim (#8) (read rows: 86400)
 
 
 # Q67
@@ -6629,15 +6626,15 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │   └── Scan: default.tpcds.store (#2) (read rows: 1)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │               └── Probe
-│                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#3) (read rows: 180)
 
 # Q68
 query I
@@ -6689,33 +6686,33 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.customer_address (read rows: 500)
+│   └── Scan: default.tpcds.customer_address (#6) (read rows: 500)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
+        │       │   └── Scan: default.tpcds.customer_address (#4) (read rows: 500)
         │       └── Probe
         │           └── HashJoin: INNER
         │               ├── Build
-        │               │   └── Scan: default.tpcds.store (read rows: 1)
+        │               │   └── Scan: default.tpcds.store (#2) (read rows: 1)
         │               └── Probe
         │                   └── HashJoin: INNER
         │                       ├── Build
         │                       │   └── HashJoin: INNER
         │                       │       ├── Build
-        │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │                       │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
         │                       │       └── Probe
-        │                       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │                       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
         │                       └── Probe
-        │                           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+        │                           └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
         └── Probe
-            └── Scan: default.tpcds.customer (read rows: 1000)
+            └── Scan: default.tpcds.customer (#5) (read rows: 1000)
 
 
 # Q69
-query I
+query
 explain join
 SELECT cd_gender,
        cd_marital_status,
@@ -6769,39 +6766,39 @@ ORDER BY cd_gender,
          cd_credit_rating
 LIMIT 100;
 ----
-HashJoin: RIGHT MARK
+HashJoin: LEFT MARK
 ├── Build
-│   └── HashJoin: INNER
+│   └── HashJoin: LEFT MARK
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── HashJoin: LEFT MARK
+│       │       ├── Build
+│       │       │   └── HashJoin: INNER
+│       │       │       ├── Build
+│       │       │       │   └── HashJoin: INNER
+│       │       │       │       ├── Build
+│       │       │       │       │   └── Scan: default.tpcds.customer_address (#1) (read rows: 500)
+│       │       │       │       └── Probe
+│       │       │       │           └── Scan: default.tpcds.customer (#0) (read rows: 1000)
+│       │       │       └── Probe
+│       │       │           └── Scan: default.tpcds.customer_demographics (#2) (read rows: 19208)
+│       │       └── Probe
+│       │           └── HashJoin: INNER
+│       │               ├── Build
+│       │               │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
+│       │               └── Probe
+│       │                   └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── HashJoin: INNER
+│               ├── Build
+│               │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
+│               └── Probe
+│                   └── Scan: default.tpcds.web_sales (#5) (read rows: 7212)
 └── Probe
-    └── HashJoin: RIGHT MARK
+    └── HashJoin: INNER
         ├── Build
-        │   └── HashJoin: INNER
-        │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       └── Probe
-        │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │   └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
         └── Probe
-            └── HashJoin: RIGHT MARK
-                ├── Build
-                │   └── HashJoin: INNER
-                │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-                │       └── Probe
-                │           └── Scan: default.tpcds.store_sales (read rows: 28810)
-                └── Probe
-                    └── HashJoin: INNER
-                        ├── Build
-                        │   └── HashJoin: INNER
-                        │       ├── Build
-                        │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
-                        │       └── Probe
-                        │           └── Scan: default.tpcds.customer (read rows: 1000)
-                        └── Probe
-                            └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+            └── Scan: default.tpcds.catalog_sales (#7) (read rows: 14313)
 
 
 # Q70
@@ -6846,7 +6843,7 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -6854,17 +6851,17 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.store (read rows: 1)
+        │       │       │   └── Scan: default.tpcds.store (#4) (read rows: 1)
         │       │       └── Probe
         │       │           └── HashJoin: INNER
         │       │               ├── Build
-        │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │               │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
         │       │               └── Probe
-        │       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │                   └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
         │       └── Probe
-        │           └── Scan: default.tpcds.store (read rows: 1)
+        │           └── Scan: default.tpcds.store (#2) (read rows: 1)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 
 # Q71
 query I
@@ -6920,7 +6917,7 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │   └── Scan: default.tpcds.item (#0) (read rows: 180)
 │       └── Probe
 │           └── UnionAll
 │               ├── Left
@@ -6928,23 +6925,23 @@ HashJoin: INNER
 │               │       ├── Left
 │               │       │   └── HashJoin: INNER
 │               │       │       ├── Build
-│               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│               │       │           └── Scan: default.tpcds.web_sales (#1) (read rows: 7212)
 │               │       └── Right
 │               │           └── HashJoin: INNER
 │               │               ├── Build
-│               │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │               │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │               │               └── Probe
-│               │                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │                   └── Scan: default.tpcds.catalog_sales (#3) (read rows: 14313)
 │               └── Right
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │                       └── Probe
-│                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│                           └── Scan: default.tpcds.store_sales (#5) (read rows: 28810)
 └── Probe
-    └── Scan: default.tpcds.time_dim (read rows: 86400)
+    └── Scan: default.tpcds.time_dim (#7) (read rows: 86400)
 
 
 # Q72
@@ -6993,11 +6990,11 @@ HashJoin: RIGHT OUTER
 ├── Build
 │   └── HashJoin: LEFT OUTER
 │       ├── Build
-│       │   └── Scan: default.tpcds.promotion (read rows: 3)
+│       │   └── Scan: default.tpcds.promotion (#9) (read rows: 3)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.warehouse (read rows: 1)
+│               │   └── Scan: default.tpcds.warehouse (#2) (read rows: 1)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
@@ -7013,23 +7010,23 @@ HashJoin: RIGHT OUTER
 │                       │       │       │       │       │       ├── Build
 │                       │       │       │       │       │       │   └── HashJoin: INNER
 │                       │       │       │       │       │       │       ├── Build
-│                       │       │       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │       │       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │                       │       │       │       │       │       │       └── Probe
-│                       │       │       │       │       │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│                       │       │       │       │       │       │           └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
 │                       │       │       │       │       │       └── Probe
-│                       │       │       │       │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│                       │       │       │       │       │           └── Scan: default.tpcds.household_demographics (#5) (read rows: 7200)
 │                       │       │       │       │       └── Probe
-│                       │       │       │       │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│                       │       │       │       │           └── Scan: default.tpcds.customer_demographics (#4) (read rows: 19208)
 │                       │       │       │       └── Probe
-│                       │       │       │           └── Scan: default.tpcds.item (read rows: 180)
+│                       │       │       │           └── Scan: default.tpcds.item (#3) (read rows: 180)
 │                       │       │       └── Probe
-│                       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │       │           └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 │                       │       └── Probe
-│                       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │           └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │                       └── Probe
-│                           └── Scan: default.tpcds.inventory (read rows: 23490)
+│                           └── Scan: default.tpcds.inventory (#1) (read rows: 23490)
 └── Probe
-    └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+    └── Scan: default.tpcds.catalog_returns (#10) (read rows: 1358)
 
 
 # Q73
@@ -7081,17 +7078,17 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │       │   └── Scan: default.tpcds.store (#2) (read rows: 1)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │               └── Probe
-│       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       └── Probe
-│           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│           └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
 └── Probe
-    └── Scan: default.tpcds.customer (read rows: 1000)
+    └── Scan: default.tpcds.customer (#4) (read rows: 1000)
 
 
 # Q74
@@ -7174,21 +7171,21 @@ HashJoin: INNER
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.date_dim (#14) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │       │       │           └── Scan: default.tpcds.store_sales (#13) (read rows: 0)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer (read rows: 0)
+│       │       │       │           └── Scan: default.tpcds.customer (#12) (read rows: 0)
 │       │       │       └── Right
 │       │       │           └── HashJoin: INNER
 │       │       │               ├── Build
 │       │       │               │   └── HashJoin: INNER
 │       │       │               │       ├── Build
-│       │       │               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │               │       │   └── Scan: default.tpcds.date_dim (#17) (read rows: 73049)
 │       │       │               │       └── Probe
-│       │       │               │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │               │           └── Scan: default.tpcds.web_sales (#16) (read rows: 7212)
 │       │       │               └── Probe
-│       │       │                   └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │                   └── Scan: default.tpcds.customer (#15) (read rows: 1000)
 │       │       └── Probe
 │       │           └── UnionAll
 │       │               ├── Left
@@ -7196,21 +7193,21 @@ HashJoin: INNER
 │       │               │       ├── Build
 │       │               │       │   └── HashJoin: INNER
 │       │               │       │       ├── Build
-│       │               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │               │       │       └── Probe
-│       │               │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │               │       │           └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 │       │               │       └── Probe
-│       │               │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │               │           └── Scan: default.tpcds.customer (#0) (read rows: 1000)
 │       │               └── Right
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
 │       │                       │   └── HashJoin: INNER
 │       │                       │       ├── Build
-│       │                       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│       │                       │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 0)
 │       │                       │       └── Probe
-│       │                       │           └── Scan: default.tpcds.web_sales (read rows: 0)
+│       │                       │           └── Scan: default.tpcds.web_sales (#4) (read rows: 0)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.customer (read rows: 0)
+│       │                           └── Scan: default.tpcds.customer (#3) (read rows: 0)
 │       └── Probe
 │           └── UnionAll
 │               ├── Left
@@ -7218,21 +7215,21 @@ HashJoin: INNER
 │               │       ├── Build
 │               │       │   └── HashJoin: INNER
 │               │       │       ├── Build
-│               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+│               │       │       │   └── Scan: default.tpcds.date_dim (#20) (read rows: 0)
 │               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│               │       │           └── Scan: default.tpcds.store_sales (#19) (read rows: 0)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.customer (read rows: 0)
+│               │           └── Scan: default.tpcds.customer (#18) (read rows: 0)
 │               └── Right
 │                   └── HashJoin: INNER
 │                       ├── Build
 │                       │   └── HashJoin: INNER
 │                       │       ├── Build
-│                       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │       │   └── Scan: default.tpcds.date_dim (#23) (read rows: 73049)
 │                       │       └── Probe
-│                       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│                       │           └── Scan: default.tpcds.web_sales (#22) (read rows: 7212)
 │                       └── Probe
-│                           └── Scan: default.tpcds.customer (read rows: 1000)
+│                           └── Scan: default.tpcds.customer (#21) (read rows: 1000)
 └── Probe
     └── UnionAll
         ├── Left
@@ -7240,21 +7237,21 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │   └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │           └── Scan: default.tpcds.store_sales (#7) (read rows: 28810)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │           └── Scan: default.tpcds.customer (#6) (read rows: 1000)
         └── Right
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 0)
+                │       │   └── Scan: default.tpcds.date_dim (#11) (read rows: 0)
                 │       └── Probe
-                │           └── Scan: default.tpcds.web_sales (read rows: 0)
+                │           └── Scan: default.tpcds.web_sales (#10) (read rows: 0)
                 └── Probe
-                    └── Scan: default.tpcds.customer (read rows: 0)
+                    └── Scan: default.tpcds.customer (#9) (read rows: 0)
 
 
 # Q75
@@ -7346,43 +7343,43 @@ HashJoin: INNER
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │       │       │   └── Scan: default.tpcds.item (#13) (read rows: 180)
 │       │       │       │       └── Probe
 │       │       │       │           └── HashJoin: INNER
 │       │       │       │               ├── Build
-│       │       │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │               │   └── Scan: default.tpcds.date_dim (#14) (read rows: 73049)
 │       │       │       │               └── Probe
-│       │       │       │                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │       │       │                   └── Scan: default.tpcds.catalog_sales (#12) (read rows: 14313)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│       │       │           └── Scan: default.tpcds.catalog_returns (#15) (read rows: 1358)
 │       │       └── Right
 │       │           └── HashJoin: RIGHT OUTER
 │       │               ├── Build
 │       │               │   └── HashJoin: INNER
 │       │               │       ├── Build
-│       │               │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │               │       │   └── Scan: default.tpcds.item (#17) (read rows: 180)
 │       │               │       └── Probe
 │       │               │           └── HashJoin: INNER
 │       │               │               ├── Build
-│       │               │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │               │   └── Scan: default.tpcds.date_dim (#18) (read rows: 73049)
 │       │               │               └── Probe
-│       │               │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │               │                   └── Scan: default.tpcds.store_sales (#16) (read rows: 28810)
 │       │               └── Probe
-│       │                   └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │                   └── Scan: default.tpcds.store_returns (#19) (read rows: 2810)
 │       └── Right
 │           └── HashJoin: RIGHT OUTER
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.item (read rows: 180)
+│               │       │   └── Scan: default.tpcds.item (#21) (read rows: 180)
 │               │       └── Probe
 │               │           └── HashJoin: INNER
 │               │               ├── Build
-│               │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │               │   └── Scan: default.tpcds.date_dim (#22) (read rows: 73049)
 │               │               └── Probe
-│               │                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│               │                   └── Scan: default.tpcds.web_sales (#20) (read rows: 7212)
 │               └── Probe
-│                   └── Scan: default.tpcds.web_returns (read rows: 679)
+│                   └── Scan: default.tpcds.web_returns (#23) (read rows: 679)
 └── Probe
     └── UnionAll
         ├── Left
@@ -7392,43 +7389,43 @@ HashJoin: INNER
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+        │       │       │       │   └── Scan: default.tpcds.item (#1) (read rows: 180)
         │       │       │       └── Probe
         │       │       │           └── HashJoin: INNER
         │       │       │               ├── Build
-        │       │       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │               │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
         │       │       │               └── Probe
-        │       │       │                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+        │       │       │                   └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+        │       │           └── Scan: default.tpcds.catalog_returns (#3) (read rows: 1358)
         │       └── Right
         │           └── HashJoin: RIGHT OUTER
         │               ├── Build
         │               │   └── HashJoin: INNER
         │               │       ├── Build
-        │               │       │   └── Scan: default.tpcds.item (read rows: 180)
+        │               │       │   └── Scan: default.tpcds.item (#5) (read rows: 180)
         │               │       └── Probe
         │               │           └── HashJoin: INNER
         │               │               ├── Build
-        │               │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │               │               │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
         │               │               └── Probe
-        │               │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │               │                   └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
         │               └── Probe
-        │                   └── Scan: default.tpcds.store_returns (read rows: 2810)
+        │                   └── Scan: default.tpcds.store_returns (#7) (read rows: 2810)
         └── Right
             └── HashJoin: RIGHT OUTER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.item (read rows: 180)
+                │       │   └── Scan: default.tpcds.item (#9) (read rows: 180)
                 │       └── Probe
                 │           └── HashJoin: INNER
                 │               ├── Build
-                │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │               │   └── Scan: default.tpcds.date_dim (#10) (read rows: 73049)
                 │               └── Probe
-                │                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+                │                   └── Scan: default.tpcds.web_sales (#8) (read rows: 7212)
                 └── Probe
-                    └── Scan: default.tpcds.web_returns (read rows: 679)
+                    └── Scan: default.tpcds.web_returns (#11) (read rows: 679)
 
 # Q76
 query I
@@ -7497,31 +7494,31 @@ UnionAll
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │       │   └── Scan: default.tpcds.item (#1) (read rows: 180)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.store_sales (read rows: 0)
+│       │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 0)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │           └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       └── Right
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │   └── Scan: default.tpcds.item (#4) (read rows: 180)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.web_sales (read rows: 0)
+│                       │   └── Scan: default.tpcds.web_sales (#3) (read rows: 0)
 │                       └── Probe
-│                           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                           └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
 └── Right
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.item (read rows: 180)
+        │       │   └── Scan: default.tpcds.item (#7) (read rows: 180)
         │       └── Probe
-        │           └── Scan: default.tpcds.catalog_sales (read rows: 0)
+        │           └── Scan: default.tpcds.catalog_sales (#6) (read rows: 0)
         └── Probe
-            └── Scan: default.tpcds.date_dim (read rows: 73049)
+            └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 
 # Q77
 query I
@@ -7631,35 +7628,35 @@ UnionAll
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │       │       │           └── Scan: default.tpcds.store_returns (#3) (read rows: 2810)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.store (read rows: 1)
+│       │       │           └── Scan: default.tpcds.store (#5) (read rows: 1)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.store (read rows: 1)
+│       │               │   └── Scan: default.tpcds.store (#2) (read rows: 1)
 │       │               └── Probe
 │       │                   └── HashJoin: INNER
 │       │                       ├── Build
-│       │                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │                       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │                       └── Probe
-│       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       └── Right
 │           └── HashJoin: CROSS
 │               ├── Build
 │               │   └── HashJoin: INNER
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │   └── Scan: default.tpcds.date_dim (#9) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│               │           └── Scan: default.tpcds.catalog_returns (#8) (read rows: 1358)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
-│                       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │                       └── Probe
-│                           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│                           └── Scan: default.tpcds.catalog_sales (#6) (read rows: 14313)
 └── Right
     └── HashJoin: LEFT OUTER
         ├── Build
@@ -7667,21 +7664,21 @@ UnionAll
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │   └── Scan: default.tpcds.date_dim (#14) (read rows: 73049)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.web_returns (read rows: 679)
+        │       │           └── Scan: default.tpcds.web_returns (#13) (read rows: 679)
         │       └── Probe
-        │           └── Scan: default.tpcds.web_page (read rows: 1)
+        │           └── Scan: default.tpcds.web_page (#15) (read rows: 1)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │       │   └── Scan: default.tpcds.date_dim (#11) (read rows: 73049)
                 │       └── Probe
-                │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+                │           └── Scan: default.tpcds.web_sales (#10) (read rows: 7212)
                 └── Probe
-                    └── Scan: default.tpcds.web_page (read rows: 1)
+                    └── Scan: default.tpcds.web_page (#12) (read rows: 1)
 
 # Q78
 query I
@@ -7767,35 +7764,35 @@ HashJoin: LEFT OUTER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── HashJoin: LEFT OUTER
-│       │       ├── Build
-│       │       │   └── Scan: default.tpcds.catalog_returns (read rows: 1358)
-│       │       └── Probe
-│       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│       │   └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│           └── HashJoin: LEFT OUTER
+│               ├── Build
+│               │   └── Scan: default.tpcds.catalog_returns (#7) (read rows: 1358)
+│               └── Probe
+│                   └── Scan: default.tpcds.catalog_sales (#6) (read rows: 14313)
 └── Probe
     └── HashJoin: LEFT OUTER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── HashJoin: LEFT OUTER
-        │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.web_returns (read rows: 679)
-        │       │       └── Probe
-        │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │   └── Scan: default.tpcds.date_dim (#5) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │           └── HashJoin: LEFT OUTER
+        │               ├── Build
+        │               │   └── Scan: default.tpcds.web_returns (#4) (read rows: 679)
+        │               └── Probe
+        │                   └── Scan: default.tpcds.web_sales (#3) (read rows: 7212)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
                 └── Probe
                     └── HashJoin: LEFT OUTER
                         ├── Build
-                        │   └── Scan: default.tpcds.store_returns (read rows: 2810)
+                        │   └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
                         └── Probe
-                            └── Scan: default.tpcds.store_sales (read rows: 28810)
+                            └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 
 # Q79
 query I
@@ -7844,17 +7841,17 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.store (read rows: 1)
+│       │       │   └── Scan: default.tpcds.store (#2) (read rows: 1)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │               │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │               └── Probe
-│       │                   └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │                   └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       └── Probe
-│           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│           └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
 └── Probe
-    └── Scan: default.tpcds.customer (read rows: 1000)
+    └── Scan: default.tpcds.customer (#4) (read rows: 1000)
 
 # Q80
 query I
@@ -7956,21 +7953,21 @@ UnionAll
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │       │       │       └── Probe
 │       │       │       │           └── HashJoin: INNER
 │       │       │       │               ├── Build
-│       │       │       │               │   └── Scan: default.tpcds.item (read rows: 180)
+│       │       │       │               │   └── Scan: default.tpcds.item (#4) (read rows: 180)
 │       │       │       │               └── Probe
 │       │       │       │                   └── HashJoin: LEFT OUTER
 │       │       │       │                       ├── Build
-│       │       │       │                       │   └── Scan: default.tpcds.store_returns (read rows: 2810)
+│       │       │       │                       │   └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
 │       │       │       │                       └── Probe
-│       │       │       │                           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │       │                           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.store (read rows: 1)
+│       │       │           └── Scan: default.tpcds.store (#3) (read rows: 1)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.promotion (read rows: 3)
+│       │           └── Scan: default.tpcds.promotion (#5) (read rows: 3)
 │       └── Right
 │           └── HashJoin: INNER
 │               ├── Build
@@ -7978,21 +7975,21 @@ UnionAll
 │               │       ├── Build
 │               │       │   └── HashJoin: INNER
 │               │       │       ├── Build
-│               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │       │   └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 │               │       │       └── Probe
 │               │       │           └── HashJoin: INNER
 │               │       │               ├── Build
-│               │       │               │   └── Scan: default.tpcds.item (read rows: 180)
+│               │       │               │   └── Scan: default.tpcds.item (#10) (read rows: 180)
 │               │       │               └── Probe
 │               │       │                   └── HashJoin: LEFT OUTER
 │               │       │                       ├── Build
-│               │       │                       │   └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│               │       │                       │   └── Scan: default.tpcds.catalog_returns (#7) (read rows: 1358)
 │               │       │                       └── Probe
-│               │       │                           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│               │       │                           └── Scan: default.tpcds.catalog_sales (#6) (read rows: 14313)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.catalog_page (read rows: 11718)
+│               │           └── Scan: default.tpcds.catalog_page (#9) (read rows: 11718)
 │               └── Probe
-│                   └── Scan: default.tpcds.promotion (read rows: 3)
+│                   └── Scan: default.tpcds.promotion (#11) (read rows: 3)
 └── Right
     └── HashJoin: INNER
         ├── Build
@@ -8000,21 +7997,21 @@ UnionAll
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │   └── Scan: default.tpcds.date_dim (#14) (read rows: 73049)
         │       │       └── Probe
         │       │           └── HashJoin: INNER
         │       │               ├── Build
-        │       │               │   └── Scan: default.tpcds.item (read rows: 180)
+        │       │               │   └── Scan: default.tpcds.item (#16) (read rows: 180)
         │       │               └── Probe
         │       │                   └── HashJoin: LEFT OUTER
         │       │                       ├── Build
-        │       │                       │   └── Scan: default.tpcds.web_returns (read rows: 679)
+        │       │                       │   └── Scan: default.tpcds.web_returns (#13) (read rows: 679)
         │       │                       └── Probe
-        │       │                           └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │       │                           └── Scan: default.tpcds.web_sales (#12) (read rows: 7212)
         │       └── Probe
-        │           └── Scan: default.tpcds.web_site (read rows: 1)
+        │           └── Scan: default.tpcds.web_site (#15) (read rows: 1)
         └── Probe
-            └── Scan: default.tpcds.promotion (read rows: 3)
+            └── Scan: default.tpcds.promotion (#17) (read rows: 3)
 
 # Q81
 query I
@@ -8075,35 +8072,35 @@ ORDER BY c_customer_id,
          ctr_total_return
 LIMIT 100;
 ----
-HashJoin: RIGHT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │   └── HashJoin: INNER
+│       │       ├── Build
+│       │       │   └── HashJoin: INNER
+│       │       │       ├── Build
+│       │       │       │   └── HashJoin: INNER
+│       │       │       │       ├── Build
+│       │       │       │       │   └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
+│       │       │       │       └── Probe
+│       │       │       │           └── Scan: default.tpcds.catalog_returns (#5) (read rows: 1358)
+│       │       │       └── Probe
+│       │       │           └── Scan: default.tpcds.customer_address (#7) (read rows: 500)
+│       │       └── Probe
+│       │           └── HashJoin: INNER
+│       │               ├── Build
+│       │               │   └── HashJoin: INNER
+│       │               │       ├── Build
+│       │               │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
+│       │               │       └── Probe
+│       │               │           └── Scan: default.tpcds.catalog_returns (#0) (read rows: 1358)
+│       │               └── Probe
+│       │                   └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 │       └── Probe
-│           └── HashJoin: INNER
-│               ├── Build
-│               │   └── HashJoin: INNER
-│               │       ├── Build
-│               │       │   └── HashJoin: INNER
-│               │       │       ├── Build
-│               │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-│               │       │       └── Probe
-│               │       │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
-│               │       └── Probe
-│               │           └── Scan: default.tpcds.customer_address (read rows: 500)
-│               └── Probe
-│                   └── Scan: default.tpcds.customer (read rows: 1000)
+│           └── Scan: default.tpcds.customer (#4) (read rows: 1000)
 └── Probe
-    └── HashJoin: INNER
-        ├── Build
-        │   └── HashJoin: INNER
-        │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        │       └── Probe
-        │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
-        └── Probe
-            └── Scan: default.tpcds.customer_address (read rows: 500)
+    └── Scan: default.tpcds.customer_address (#3) (read rows: 500)
 
 # Q82
 query I
@@ -8135,18 +8132,18 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │   └── Scan: default.tpcds.item (#0) (read rows: 180)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │               └── Probe
-│                   └── Scan: default.tpcds.inventory (read rows: 23490)
+│                   └── Scan: default.tpcds.inventory (#1) (read rows: 23490)
 └── Probe
-    └── Scan: default.tpcds.store_sales (read rows: 28810)
+    └── Scan: default.tpcds.store_sales (#3) (read rows: 28810)
 
 # Q83
-query I
+query
 explain join
 WITH sr_items AS
   (SELECT i_item_id item_id,
@@ -8227,55 +8224,55 @@ HashJoin: INNER
 │       │       ├── Build
 │       │       │   └── HashJoin: LEFT SEMI
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │   └── Scan: default.tpcds.date_dim (#14) (read rows: 73049)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │           └── Scan: default.tpcds.date_dim (#13) (read rows: 73049)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
 │       │               │   └── HashJoin: INNER
 │       │               │       ├── Build
-│       │               │       │   └── Scan: default.tpcds.item (read rows: 180)
+│       │               │       │   └── Scan: default.tpcds.item (#11) (read rows: 180)
 │       │               │       └── Probe
-│       │               │           └── Scan: default.tpcds.web_returns (read rows: 679)
+│       │               │           └── Scan: default.tpcds.web_returns (#10) (read rows: 679)
 │       │               └── Probe
-│       │                   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │                   └── Scan: default.tpcds.date_dim (#12) (read rows: 73049)
 │       └── Probe
 │           └── HashJoin: LEFT SEMI
 │               ├── Build
 │               │   └── HashJoin: LEFT SEMI
 │               │       ├── Build
-│               │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │       │   └── Scan: default.tpcds.date_dim (#9) (read rows: 73049)
 │               │       └── Probe
-│               │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│               │           └── Scan: default.tpcds.date_dim (#8) (read rows: 73049)
 │               └── Probe
 │                   └── HashJoin: INNER
 │                       ├── Build
 │                       │   └── HashJoin: INNER
 │                       │       ├── Build
-│                       │       │   └── Scan: default.tpcds.item (read rows: 180)
+│                       │       │   └── Scan: default.tpcds.item (#6) (read rows: 180)
 │                       │       └── Probe
-│                       │           └── Scan: default.tpcds.store_returns (read rows: 2810)
+│                       │           └── Scan: default.tpcds.catalog_returns (#5) (read rows: 1358)
 │                       └── Probe
-│                           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│                           └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 └── Probe
     └── HashJoin: LEFT SEMI
         ├── Build
         │   └── HashJoin: LEFT SEMI
         │       ├── Build
-        │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
         │       └── Probe
-        │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │           └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.item (read rows: 180)
+                │       │   └── Scan: default.tpcds.item (#1) (read rows: 180)
                 │       └── Probe
-                │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+                │           └── Scan: default.tpcds.store_returns (#0) (read rows: 2810)
                 └── Probe
-                    └── Scan: default.tpcds.date_dim (read rows: 73049)
+                    └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 
 # Q84
 query I
@@ -8309,17 +8306,17 @@ HashJoin: INNER
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │       │   └── Scan: default.tpcds.customer_address (#1) (read rows: 500)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │       │           └── Scan: default.tpcds.customer (#0) (read rows: 1000)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │       │           └── Scan: default.tpcds.household_demographics (#3) (read rows: 7200)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│       │           └── Scan: default.tpcds.income_band (#4) (read rows: 20)
 │       └── Probe
-│           └── Scan: default.tpcds.income_band (read rows: 20)
+│           └── Scan: default.tpcds.customer_demographics (#2) (read rows: 19208)
 └── Probe
-    └── Scan: default.tpcds.store_returns (read rows: 2810)
+    └── Scan: default.tpcds.store_returns (#5) (read rows: 2810)
 
 # Q85
 query I
@@ -8396,21 +8393,21 @@ HashJoin: INNER
 │       │       │       │       │       ├── Build
 │       │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       │       ├── Build
-│       │       │       │       │       │       │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │       │       │       │       │   └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 │       │       │       │       │       │       └── Probe
-│       │       │       │       │       │           └── Scan: default.tpcds.web_returns (read rows: 679)
+│       │       │       │       │       │           └── Scan: default.tpcds.web_returns (#1) (read rows: 679)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │       │       │       │           └── Scan: default.tpcds.customer_address (#5) (read rows: 500)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │       │           └── Scan: default.tpcds.date_dim (#6) (read rows: 73049)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.web_page (read rows: 1)
+│       │       │           └── Scan: default.tpcds.web_page (#2) (read rows: 1)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │           └── Scan: default.tpcds.customer_demographics (#3) (read rows: 19208)
 │       └── Probe
-│           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│           └── Scan: default.tpcds.customer_demographics (#4) (read rows: 19208)
 └── Probe
-    └── Scan: default.tpcds.reason (read rows: 1)
+    └── Scan: default.tpcds.reason (#7) (read rows: 1)
 
 # Q86
 query I
@@ -8442,11 +8439,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#2) (read rows: 180)
 
 # Q87
 query I
@@ -8488,11 +8485,11 @@ HashJoin: LEFT ANTI
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │   └── Scan: default.tpcds.date_dim (#7) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │           └── Scan: default.tpcds.web_sales (#6) (read rows: 7212)
 │       └── Probe
-│           └── Scan: default.tpcds.customer (read rows: 1000)
+│           └── Scan: default.tpcds.customer (#8) (read rows: 1000)
 └── Probe
     └── HashJoin: LEFT ANTI
         ├── Build
@@ -8500,21 +8497,21 @@ HashJoin: LEFT ANTI
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+        │       │           └── Scan: default.tpcds.catalog_sales (#3) (read rows: 14313)
         │       └── Probe
-        │           └── Scan: default.tpcds.customer (read rows: 1000)
+        │           └── Scan: default.tpcds.customer (#5) (read rows: 1000)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
                 │   └── HashJoin: INNER
                 │       ├── Build
-                │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │       │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
                 │       └── Probe
-                │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
                 └── Probe
-                    └── Scan: default.tpcds.customer (read rows: 1000)
+                    └── Scan: default.tpcds.customer (#2) (read rows: 1000)
 
 # Q88
 query I
@@ -8666,13 +8663,13 @@ HashJoin: CROSS
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+│       │       │       │   └── Scan: default.tpcds.store (#31) (read rows: 0)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │       │           └── Scan: default.tpcds.store_sales (#28) (read rows: 28810)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│       │           └── Scan: default.tpcds.household_demographics (#29) (read rows: 7200)
 │       └── Probe
-│           └── Scan: default.tpcds.time_dim (read rows: 86400)
+│           └── Scan: default.tpcds.time_dim (#30) (read rows: 86400)
 └── Probe
     └── HashJoin: CROSS
         ├── Build
@@ -8682,13 +8679,13 @@ HashJoin: CROSS
         │       │       ├── Build
         │       │       │   └── HashJoin: INNER
         │       │       │       ├── Build
-        │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+        │       │       │       │   └── Scan: default.tpcds.store (#27) (read rows: 0)
         │       │       │       └── Probe
-        │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+        │       │       │           └── Scan: default.tpcds.store_sales (#24) (read rows: 28810)
         │       │       └── Probe
-        │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+        │       │           └── Scan: default.tpcds.household_demographics (#25) (read rows: 7200)
         │       └── Probe
-        │           └── Scan: default.tpcds.time_dim (read rows: 86400)
+        │           └── Scan: default.tpcds.time_dim (#26) (read rows: 86400)
         └── Probe
             └── HashJoin: CROSS
                 ├── Build
@@ -8698,13 +8695,13 @@ HashJoin: CROSS
                 │       │       ├── Build
                 │       │       │   └── HashJoin: INNER
                 │       │       │       ├── Build
-                │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+                │       │       │       │   └── Scan: default.tpcds.store (#23) (read rows: 0)
                 │       │       │       └── Probe
-                │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                │       │       │           └── Scan: default.tpcds.store_sales (#20) (read rows: 28810)
                 │       │       └── Probe
-                │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+                │       │           └── Scan: default.tpcds.household_demographics (#21) (read rows: 7200)
                 │       └── Probe
-                │           └── Scan: default.tpcds.time_dim (read rows: 86400)
+                │           └── Scan: default.tpcds.time_dim (#22) (read rows: 86400)
                 └── Probe
                     └── HashJoin: CROSS
                         ├── Build
@@ -8714,13 +8711,13 @@ HashJoin: CROSS
                         │       │       ├── Build
                         │       │       │   └── HashJoin: INNER
                         │       │       │       ├── Build
-                        │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+                        │       │       │       │   └── Scan: default.tpcds.store (#19) (read rows: 0)
                         │       │       │       └── Probe
-                        │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                        │       │       │           └── Scan: default.tpcds.store_sales (#16) (read rows: 28810)
                         │       │       └── Probe
-                        │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+                        │       │           └── Scan: default.tpcds.household_demographics (#17) (read rows: 7200)
                         │       └── Probe
-                        │           └── Scan: default.tpcds.time_dim (read rows: 86400)
+                        │           └── Scan: default.tpcds.time_dim (#18) (read rows: 86400)
                         └── Probe
                             └── HashJoin: CROSS
                                 ├── Build
@@ -8730,13 +8727,13 @@ HashJoin: CROSS
                                 │       │       ├── Build
                                 │       │       │   └── HashJoin: INNER
                                 │       │       │       ├── Build
-                                │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+                                │       │       │       │   └── Scan: default.tpcds.store (#15) (read rows: 0)
                                 │       │       │       └── Probe
-                                │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                │       │       │           └── Scan: default.tpcds.store_sales (#12) (read rows: 28810)
                                 │       │       └── Probe
-                                │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+                                │       │           └── Scan: default.tpcds.household_demographics (#13) (read rows: 7200)
                                 │       └── Probe
-                                │           └── Scan: default.tpcds.time_dim (read rows: 86400)
+                                │           └── Scan: default.tpcds.time_dim (#14) (read rows: 86400)
                                 └── Probe
                                     └── HashJoin: CROSS
                                         ├── Build
@@ -8746,13 +8743,13 @@ HashJoin: CROSS
                                         │       │       ├── Build
                                         │       │       │   └── HashJoin: INNER
                                         │       │       │       ├── Build
-                                        │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+                                        │       │       │       │   └── Scan: default.tpcds.store (#11) (read rows: 0)
                                         │       │       │       └── Probe
-                                        │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                        │       │       │           └── Scan: default.tpcds.store_sales (#8) (read rows: 28810)
                                         │       │       └── Probe
-                                        │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+                                        │       │           └── Scan: default.tpcds.household_demographics (#9) (read rows: 7200)
                                         │       └── Probe
-                                        │           └── Scan: default.tpcds.time_dim (read rows: 86400)
+                                        │           └── Scan: default.tpcds.time_dim (#10) (read rows: 86400)
                                         └── Probe
                                             └── HashJoin: CROSS
                                                 ├── Build
@@ -8762,13 +8759,13 @@ HashJoin: CROSS
                                                 │       │       ├── Build
                                                 │       │       │   └── HashJoin: INNER
                                                 │       │       │       ├── Build
-                                                │       │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+                                                │       │       │       │   └── Scan: default.tpcds.store (#7) (read rows: 0)
                                                 │       │       │       └── Probe
-                                                │       │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                │       │       │           └── Scan: default.tpcds.store_sales (#4) (read rows: 28810)
                                                 │       │       └── Probe
-                                                │       │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+                                                │       │           └── Scan: default.tpcds.household_demographics (#5) (read rows: 7200)
                                                 │       └── Probe
-                                                │           └── Scan: default.tpcds.time_dim (read rows: 86400)
+                                                │           └── Scan: default.tpcds.time_dim (#6) (read rows: 86400)
                                                 └── Probe
                                                     └── HashJoin: INNER
                                                         ├── Build
@@ -8776,13 +8773,13 @@ HashJoin: CROSS
                                                         │       ├── Build
                                                         │       │   └── HashJoin: INNER
                                                         │       │       ├── Build
-                                                        │       │       │   └── Scan: default.tpcds.store (read rows: 0)
+                                                        │       │       │   └── Scan: default.tpcds.store (#3) (read rows: 0)
                                                         │       │       └── Probe
-                                                        │       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+                                                        │       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
                                                         │       └── Probe
-                                                        │           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+                                                        │           └── Scan: default.tpcds.household_demographics (#1) (read rows: 7200)
                                                         └── Probe
-                                                            └── Scan: default.tpcds.time_dim (read rows: 86400)
+                                                            └── Scan: default.tpcds.time_dim (#2) (read rows: 86400)
 
 # Q89
 query I
@@ -8809,17 +8806,17 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.store (read rows: 1)
+│   └── Scan: default.tpcds.store (#3) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.item (read rows: 180)
+        │   └── Scan: default.tpcds.item (#0) (read rows: 180)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+                │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
                 └── Probe
-                    └── Scan: default.tpcds.store_sales (read rows: 28810)
+                    └── Scan: default.tpcds.store_sales (#1) (read rows: 28810)
 
 # Q90
 query I
@@ -8857,29 +8854,29 @@ HashJoin: CROSS
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.web_page (read rows: 0)
+│       │       │   └── Scan: default.tpcds.web_page (#7) (read rows: 0)
 │       │       └── Probe
 │       │           └── HashJoin: INNER
 │       │               ├── Build
-│       │               │   └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│       │               │   └── Scan: default.tpcds.household_demographics (#5) (read rows: 7200)
 │       │               └── Probe
-│       │                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │                   └── Scan: default.tpcds.web_sales (#4) (read rows: 7212)
 │       └── Probe
-│           └── Scan: default.tpcds.time_dim (read rows: 86400)
+│           └── Scan: default.tpcds.time_dim (#6) (read rows: 86400)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.web_page (read rows: 0)
+        │       │   └── Scan: default.tpcds.web_page (#3) (read rows: 0)
         │       └── Probe
         │           └── HashJoin: INNER
         │               ├── Build
-        │               │   └── Scan: default.tpcds.household_demographics (read rows: 7200)
+        │               │   └── Scan: default.tpcds.household_demographics (#1) (read rows: 7200)
         │               └── Probe
-        │                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+        │                   └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
         └── Probe
-            └── Scan: default.tpcds.time_dim (read rows: 86400)
+            └── Scan: default.tpcds.time_dim (#2) (read rows: 86400)
 
 # Q91
 query I
@@ -8927,19 +8924,19 @@ HashJoin: INNER
 │       │       │       │       ├── Build
 │       │       │       │       │   └── HashJoin: INNER
 │       │       │       │       │       ├── Build
-│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (read rows: 0)
+│       │       │       │       │       │   └── Scan: default.tpcds.customer_address (#4) (read rows: 0)
 │       │       │       │       │       └── Probe
-│       │       │       │       │           └── Scan: default.tpcds.customer (read rows: 1000)
+│       │       │       │       │           └── Scan: default.tpcds.customer (#3) (read rows: 1000)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.catalog_returns (read rows: 1358)
+│       │       │       │           └── Scan: default.tpcds.catalog_returns (#1) (read rows: 1358)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │           └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer_demographics (read rows: 19208)
+│       │           └── Scan: default.tpcds.customer_demographics (#5) (read rows: 19208)
 │       └── Probe
-│           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│           └── Scan: default.tpcds.household_demographics (#6) (read rows: 7200)
 └── Probe
-    └── Scan: default.tpcds.call_center (read rows: 1)
+    └── Scan: default.tpcds.call_center (#0) (read rows: 1)
 
 # Q92
 query I
@@ -8962,23 +8959,23 @@ WHERE i_manufact_id = 350
 ORDER BY sum(ws_ext_discount_amt)
 LIMIT 100;
 ----
-HashJoin: RIGHT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │   └── HashJoin: INNER
+│       │       │       ├── Build
+│       │       │       │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
+│       │       │       └── Probe
+│       │       │           └── Scan: default.tpcds.web_sales (#3) (read rows: 7212)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │           └── Scan: default.tpcds.item (#1) (read rows: 180)
 │       └── Probe
-│           └── Scan: default.tpcds.item (read rows: 180)
+│           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 └── Probe
-    └── HashJoin: INNER
-        ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
-        └── Probe
-            └── Scan: default.tpcds.web_sales (read rows: 7212)
+    └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 
 # Q93
 query I
@@ -9005,13 +9002,13 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.reason (read rows: 0)
+│   └── Scan: default.tpcds.reason (#2) (read rows: 0)
 └── Probe
     └── HashJoin: LEFT OUTER
         ├── Build
-        │   └── Scan: default.tpcds.store_returns (read rows: 2810)
+        │   └── Scan: default.tpcds.store_returns (#1) (read rows: 2810)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 
 # Q94
 query I
@@ -9051,17 +9048,17 @@ HashJoin: RIGHT ANTI
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.web_site (read rows: 0)
+│       │       │       │       │   └── Scan: default.tpcds.web_site (#3) (read rows: 0)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │       │           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │           └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │           └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 │       └── Probe
-│           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│           └── Scan: default.tpcds.web_sales (#4) (read rows: 7212)
 └── Probe
-    └── Scan: default.tpcds.web_returns (read rows: 679)
+    └── Scan: default.tpcds.web_returns (#5) (read rows: 679)
 
 # Q95
 query I
@@ -9108,29 +9105,29 @@ HashJoin: RIGHT SEMI
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpcds.web_site (read rows: 0)
+│       │       │       │       │   └── Scan: default.tpcds.web_site (#3) (read rows: 0)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpcds.web_sales (read rows: 7212)
+│       │       │       │           └── Scan: default.tpcds.web_sales (#0) (read rows: 7212)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │       │           └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.customer_address (read rows: 500)
+│       │           └── Scan: default.tpcds.customer_address (#2) (read rows: 500)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│               │   └── Scan: default.tpcds.web_sales (#5) (read rows: 7212)
 │               └── Probe
-│                   └── Scan: default.tpcds.web_sales (read rows: 7212)
+│                   └── Scan: default.tpcds.web_sales (#4) (read rows: 7212)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.web_returns (read rows: 679)
+        │   └── Scan: default.tpcds.web_returns (#6) (read rows: 679)
         └── Probe
             └── HashJoin: INNER
                 ├── Build
-                │   └── Scan: default.tpcds.web_sales (read rows: 7212)
+                │   └── Scan: default.tpcds.web_sales (#8) (read rows: 7212)
                 └── Probe
-                    └── Scan: default.tpcds.web_sales (read rows: 7212)
+                    └── Scan: default.tpcds.web_sales (#7) (read rows: 7212)
 
 # Q96
 query I
@@ -9156,13 +9153,13 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpcds.store (read rows: 0)
+│       │       │   └── Scan: default.tpcds.store (#3) (read rows: 0)
 │       │       └── Probe
-│       │           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│       │           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 │       └── Probe
-│           └── Scan: default.tpcds.household_demographics (read rows: 7200)
+│           └── Scan: default.tpcds.household_demographics (#1) (read rows: 7200)
 └── Probe
-    └── Scan: default.tpcds.time_dim (read rows: 86400)
+    └── Scan: default.tpcds.time_dim (#2) (read rows: 86400)
 
 # Q97
 query I
@@ -9206,15 +9203,15 @@ HashJoin: FULL OUTER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#3) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+│           └── Scan: default.tpcds.catalog_sales (#2) (read rows: 14313)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │   └── Scan: default.tpcds.date_dim (#1) (read rows: 73049)
         └── Probe
-            └── Scan: default.tpcds.store_sales (read rows: 28810)
+            └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 
 # Q98
 query I
@@ -9250,11 +9247,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+│       │   └── Scan: default.tpcds.date_dim (#2) (read rows: 73049)
 │       └── Probe
-│           └── Scan: default.tpcds.store_sales (read rows: 28810)
+│           └── Scan: default.tpcds.store_sales (#0) (read rows: 28810)
 └── Probe
-    └── Scan: default.tpcds.item (read rows: 180)
+    └── Scan: default.tpcds.item (#1) (read rows: 180)
 
 # Q99
 query I
@@ -9306,22 +9303,18 @@ LIMIT 100;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpcds.call_center (read rows: 1)
+│   └── Scan: default.tpcds.call_center (#3) (read rows: 1)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpcds.warehouse (read rows: 1)
+        │       │   └── Scan: default.tpcds.warehouse (#1) (read rows: 1)
         │       └── Probe
         │           └── HashJoin: INNER
         │               ├── Build
-        │               │   └── Scan: default.tpcds.date_dim (read rows: 73049)
+        │               │   └── Scan: default.tpcds.date_dim (#4) (read rows: 73049)
         │               └── Probe
-        │                   └── Scan: default.tpcds.catalog_sales (read rows: 14313)
+        │                   └── Scan: default.tpcds.catalog_sales (#0) (read rows: 14313)
         └── Probe
-            └── Scan: default.tpcds.ship_mode (read rows: 20)
-
-
-statement ok
-set enable_runtime_filter = 0;
+            └── Scan: default.tpcds.ship_mode (#2) (read rows: 20)
\ No newline at end of file
diff --git a/tests/sqllogictests/suites/tpch/join.test b/tests/sqllogictests/suites/tpch/join.test
index d82496001369..6beb1d25c877 100644
--- a/tests/sqllogictests/suites/tpch/join.test
+++ b/tests/sqllogictests/suites/tpch/join.test
@@ -4,9 +4,6 @@ set sandbox_tenant = 'test_tenant';
 statement ok
 use tpch_test;
 
-statement ok
-set enable_runtime_filter = 1;
-
 query I
 select
     c_custkey, count(o_orderkey) as c_count
@@ -332,5 +329,28 @@ select l_orderkey from (select * from lineitem order by l_orderkey limit 5000) a
 3
 3
 
-statement ok
-set enable_runtime_filter = 0;
\ No newline at end of file
+# LEFT OUTER / LEFT SINGEL / FULL
+query I
+select l_orderkey, o_orderdate, o_shippriority from lineitem left join orders on l_orderkey = o_orderkey  and o_orderdate < to_date('1995-03-15') order by o_orderdate limit 5;
+----
+571586 1992-01-01 0
+190656 1992-01-01 0
+359170 1992-01-01 0
+414725 1992-01-01 0
+45697 1992-01-01 0
+
+# LEFT ANTI
+query I
+select o_custkey from orders where not exists (select * from customer where substring(c_phone from 1 for 2) in ('13', '31', '23', '29', '30', '18', '17') and o_custkey = c_custkey) order by o_custkey limit 10;
+----
+1
+1
+1
+1
+1
+1
+1
+1
+1
+4
+
diff --git a/tests/sqllogictests/suites/tpch/queries.test b/tests/sqllogictests/suites/tpch/queries.test
index ff5b0e1923d4..f73b0ec40d14 100644
--- a/tests/sqllogictests/suites/tpch/queries.test
+++ b/tests/sqllogictests/suites/tpch/queries.test
@@ -1,9 +1,6 @@
 statement ok
 set sandbox_tenant = 'test_tenant';
 
-statement ok
-set enable_runtime_filter = 1;
-
 statement ok
 use tpch_test;
 
@@ -1175,7 +1172,7 @@ order by
     l_returnflag,
     l_linestatus;
 ----
-Scan: default.tpch_test.lineitem (read rows: 600572)
+Scan: default.tpch_test.lineitem (#0) (read rows: 600572)
 
 # Q2
 query I
@@ -1223,7 +1220,7 @@ order by
     s_name,
     p_partkey;
 ----
-HashJoin: RIGHT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
@@ -1231,17 +1228,17 @@ HashJoin: RIGHT SINGLE
 │       │       ├── Build
 │       │       │   └── HashJoin: INNER
 │       │       │       ├── Build
-│       │       │       │   └── Scan: default.tpch_test.region (read rows: 5)
+│       │       │       │   └── Scan: default.tpch_test.region (#4) (read rows: 5)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpch_test.nation (read rows: 25)
+│       │       │           └── Scan: default.tpch_test.nation (#3) (read rows: 25)
 │       │       └── Probe
-│       │           └── Scan: default.tpch_test.supplier (read rows: 1000)
+│       │           └── Scan: default.tpch_test.supplier (#1) (read rows: 1000)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpch_test.part (read rows: 20000)
+│               │   └── Scan: default.tpch_test.part (#0) (read rows: 20000)
 │               └── Probe
-│                   └── Scan: default.tpch_test.partsupp (read rows: 80000)
+│                   └── Scan: default.tpch_test.partsupp (#2) (read rows: 80000)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -1249,13 +1246,13 @@ HashJoin: RIGHT SINGLE
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpch_test.region (read rows: 5)
+        │       │       │   └── Scan: default.tpch_test.region (#8) (read rows: 5)
         │       │       └── Probe
-        │       │           └── Scan: default.tpch_test.nation (read rows: 25)
+        │       │           └── Scan: default.tpch_test.nation (#7) (read rows: 25)
         │       └── Probe
-        │           └── Scan: default.tpch_test.supplier (read rows: 1000)
+        │           └── Scan: default.tpch_test.supplier (#6) (read rows: 1000)
         └── Probe
-            └── Scan: default.tpch_test.partsupp (read rows: 80000)
+            └── Scan: default.tpch_test.partsupp (#5) (read rows: 80000)
 
 # Q3
 query I
@@ -1287,11 +1284,11 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpch_test.customer (read rows: 15000)
+│       │   └── Scan: default.tpch_test.customer (#0) (read rows: 15000)
 │       └── Probe
-│           └── Scan: default.tpch_test.orders (read rows: 150000)
+│           └── Scan: default.tpch_test.orders (#1) (read rows: 150000)
 └── Probe
-    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+    └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
 
 # Q4
 query I
@@ -1319,9 +1316,9 @@ order by
 ----
 HashJoin: RIGHT SEMI
 ├── Build
-│   └── Scan: default.tpch_test.orders (read rows: 150000)
+│   └── Scan: default.tpch_test.orders (#0) (read rows: 150000)
 └── Probe
-    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+    └── Scan: default.tpch_test.lineitem (#1) (read rows: 600572)
 
 # Q5
 query I
@@ -1352,25 +1349,25 @@ order by
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.customer (read rows: 15000)
+│   └── Scan: default.tpch_test.supplier (#3) (read rows: 1000)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpch_test.orders (read rows: 150000)
+        │   └── HashJoin: INNER
+        │       ├── Build
+        │       │   └── HashJoin: INNER
+        │       │       ├── Build
+        │       │       │   └── HashJoin: INNER
+        │       │       │       ├── Build
+        │       │       │       │   └── Scan: default.tpch_test.region (#5) (read rows: 5)
+        │       │       │       └── Probe
+        │       │       │           └── Scan: default.tpch_test.nation (#4) (read rows: 25)
+        │       │       └── Probe
+        │       │           └── Scan: default.tpch_test.customer (#0) (read rows: 15000)
+        │       └── Probe
+        │           └── Scan: default.tpch_test.orders (#1) (read rows: 150000)
         └── Probe
-            └── HashJoin: INNER
-                ├── Build
-                │   └── HashJoin: INNER
-                │       ├── Build
-                │       │   └── HashJoin: INNER
-                │       │       ├── Build
-                │       │       │   └── Scan: default.tpch_test.region (read rows: 5)
-                │       │       └── Probe
-                │       │           └── Scan: default.tpch_test.nation (read rows: 25)
-                │       └── Probe
-                │           └── Scan: default.tpch_test.supplier (read rows: 1000)
-                └── Probe
-                    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+            └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
 
 # Q6
 query I
@@ -1384,7 +1381,7 @@ where
    and l_discount between 0.05 and 0.07
    and l_quantity < 24;
 ----
-Scan: default.tpch_test.lineitem (read rows: 600572)
+Scan: default.tpch_test.lineitem (#0) (read rows: 600572)
 
 # Q7
 query I
@@ -1432,9 +1429,9 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpch_test.nation (read rows: 25)
+│       │   └── Scan: default.tpch_test.nation (#4) (read rows: 25)
 │       └── Probe
-│           └── Scan: default.tpch_test.supplier (read rows: 1000)
+│           └── Scan: default.tpch_test.supplier (#0) (read rows: 1000)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -1442,13 +1439,13 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpch_test.nation (read rows: 25)
+        │       │       │   └── Scan: default.tpch_test.nation (#5) (read rows: 25)
         │       │       └── Probe
-        │       │           └── Scan: default.tpch_test.customer (read rows: 15000)
+        │       │           └── Scan: default.tpch_test.customer (#3) (read rows: 15000)
         │       └── Probe
-        │           └── Scan: default.tpch_test.orders (read rows: 150000)
+        │           └── Scan: default.tpch_test.orders (#2) (read rows: 150000)
         └── Probe
-            └── Scan: default.tpch_test.lineitem (read rows: 600572)
+            └── Scan: default.tpch_test.lineitem (#1) (read rows: 600572)
 
 # Q8
 query I
@@ -1492,7 +1489,7 @@ order by
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.nation (read rows: 25)
+│   └── Scan: default.tpch_test.nation (#6) (read rows: 25)
 └── Probe
     └── HashJoin: INNER
         ├── Build
@@ -1500,9 +1497,9 @@ HashJoin: INNER
         │       ├── Build
         │       │   └── HashJoin: INNER
         │       │       ├── Build
-        │       │       │   └── Scan: default.tpch_test.region (read rows: 5)
+        │       │       │   └── Scan: default.tpch_test.region (#7) (read rows: 5)
         │       │       └── Probe
-        │       │           └── Scan: default.tpch_test.nation (read rows: 25)
+        │       │           └── Scan: default.tpch_test.nation (#5) (read rows: 25)
         │       └── Probe
         │           └── HashJoin: INNER
         │               ├── Build
@@ -1510,15 +1507,15 @@ HashJoin: INNER
         │               │       ├── Build
         │               │       │   └── HashJoin: INNER
         │               │       │       ├── Build
-        │               │       │       │   └── Scan: default.tpch_test.part (read rows: 20000)
+        │               │       │       │   └── Scan: default.tpch_test.part (#0) (read rows: 20000)
         │               │       │       └── Probe
-        │               │       │           └── Scan: default.tpch_test.lineitem (read rows: 600572)
+        │               │       │           └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
         │               │       └── Probe
-        │               │           └── Scan: default.tpch_test.orders (read rows: 150000)
+        │               │           └── Scan: default.tpch_test.orders (#3) (read rows: 150000)
         │               └── Probe
-        │                   └── Scan: default.tpch_test.customer (read rows: 15000)
+        │                   └── Scan: default.tpch_test.customer (#4) (read rows: 15000)
         └── Probe
-            └── Scan: default.tpch_test.supplier (read rows: 1000)
+            └── Scan: default.tpch_test.supplier (#1) (read rows: 1000)
 
 # Q9
 query I
@@ -1557,25 +1554,25 @@ limit 5;
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.partsupp (read rows: 80000)
+│   └── HashJoin: INNER
+│       ├── Build
+│       │   └── HashJoin: INNER
+│       │       ├── Build
+│       │       │   └── Scan: default.tpch_test.nation (#5) (read rows: 25)
+│       │       └── Probe
+│       │           └── Scan: default.tpch_test.supplier (#1) (read rows: 1000)
+│       └── Probe
+│           └── Scan: default.tpch_test.partsupp (#3) (read rows: 80000)
 └── Probe
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpch_test.nation (read rows: 25)
+        │       │   └── Scan: default.tpch_test.part (#0) (read rows: 20000)
         │       └── Probe
-        │           └── Scan: default.tpch_test.supplier (read rows: 1000)
+        │           └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
         └── Probe
-            └── HashJoin: INNER
-                ├── Build
-                │   └── HashJoin: INNER
-                │       ├── Build
-                │       │   └── Scan: default.tpch_test.part (read rows: 20000)
-                │       └── Probe
-                │           └── Scan: default.tpch_test.lineitem (read rows: 600572)
-                └── Probe
-                    └── Scan: default.tpch_test.orders (read rows: 150000)
+            └── Scan: default.tpch_test.orders (#4) (read rows: 150000)
 
 # Q10
 query I
@@ -1615,15 +1612,15 @@ HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpch_test.nation (read rows: 25)
+│       │   └── Scan: default.tpch_test.nation (#3) (read rows: 25)
 │       └── Probe
 │           └── HashJoin: INNER
 │               ├── Build
-│               │   └── Scan: default.tpch_test.orders (read rows: 150000)
+│               │   └── Scan: default.tpch_test.orders (#1) (read rows: 150000)
 │               └── Probe
-│                   └── Scan: default.tpch_test.customer (read rows: 15000)
+│                   └── Scan: default.tpch_test.customer (#0) (read rows: 15000)
 └── Probe
-    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+    └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
 
 # Q11
 query I
@@ -1661,21 +1658,21 @@ RangeJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpch_test.nation (read rows: 25)
+│       │       │   └── Scan: default.tpch_test.nation (#5) (read rows: 25)
 │       │       └── Probe
-│       │           └── Scan: default.tpch_test.supplier (read rows: 1000)
+│       │           └── Scan: default.tpch_test.supplier (#4) (read rows: 1000)
 │       └── Probe
-│           └── Scan: default.tpch_test.partsupp (read rows: 80000)
+│           └── Scan: default.tpch_test.partsupp (#3) (read rows: 80000)
 └── Right
     └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: INNER
         │       ├── Build
-        │       │   └── Scan: default.tpch_test.nation (read rows: 25)
+        │       │   └── Scan: default.tpch_test.nation (#2) (read rows: 25)
         │       └── Probe
-        │           └── Scan: default.tpch_test.supplier (read rows: 1000)
+        │           └── Scan: default.tpch_test.supplier (#1) (read rows: 1000)
         └── Probe
-            └── Scan: default.tpch_test.partsupp (read rows: 80000)
+            └── Scan: default.tpch_test.partsupp (#0) (read rows: 80000)
 
 # Q12
 query I
@@ -1710,9 +1707,9 @@ order by
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│   └── Scan: default.tpch_test.lineitem (#1) (read rows: 600572)
 └── Probe
-    └── Scan: default.tpch_test.orders (read rows: 150000)
+    └── Scan: default.tpch_test.orders (#0) (read rows: 150000)
 
 # Q13
 query I
@@ -1742,9 +1739,9 @@ order by
 ----
 HashJoin: RIGHT OUTER
 ├── Build
-│   └── Scan: default.tpch_test.customer (read rows: 15000)
+│   └── Scan: default.tpch_test.customer (#0) (read rows: 15000)
 └── Probe
-    └── Scan: default.tpch_test.orders (read rows: 150000)
+    └── Scan: default.tpch_test.orders (#1) (read rows: 150000)
 
 # Q14
 query I
@@ -1764,9 +1761,9 @@ where
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│   └── Scan: default.tpch_test.lineitem (#0) (read rows: 600572)
 └── Probe
-    └── Scan: default.tpch_test.part (read rows: 20000)
+    └── Scan: default.tpch_test.part (#1) (read rows: 20000)
 
 # Q15
 query T
@@ -1803,13 +1800,13 @@ order by
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│   └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+        │   └── Scan: default.tpch_test.lineitem (#1) (read rows: 600572)
         └── Probe
-            └── Scan: default.tpch_test.supplier (read rows: 1000)
+            └── Scan: default.tpch_test.supplier (#0) (read rows: 1000)
 
 # Q15
 query T
@@ -1846,7 +1843,7 @@ order by
 ----
 MaterializedCte: 0
 ├── Left
-│   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│   └── Scan: default.tpch_test.lineitem (#1) (read rows: 600572)
 └── Right
     └── HashJoin: INNER
         ├── Build
@@ -1860,7 +1857,7 @@ MaterializedCte: 0
                 │       ├── CTE index: 0, sub index: 1
                 │       └── estimated rows: 81.00
                 └── Probe
-                    └── Scan: default.tpch_test.supplier (read rows: 1000)
+                    └── Scan: default.tpch_test.supplier (#0) (read rows: 1000)
 
 # Q16
 query I
@@ -1898,13 +1895,13 @@ limit 20;
 ----
 HashJoin: RIGHT MARK
 ├── Build
-│   └── Scan: default.tpch_test.supplier (read rows: 1000)
+│   └── Scan: default.tpch_test.supplier (#2) (read rows: 1000)
 └── Probe
     └── HashJoin: INNER
         ├── Build
-        │   └── Scan: default.tpch_test.part (read rows: 20000)
+        │   └── Scan: default.tpch_test.part (#1) (read rows: 20000)
         └── Probe
-            └── Scan: default.tpch_test.partsupp (read rows: 80000)
+            └── Scan: default.tpch_test.partsupp (#0) (read rows: 80000)
 
 #Q17
 query I
@@ -1926,15 +1923,15 @@ where
             l_partkey = p_partkey
 );
 ----
-HashJoin: RIGHT SINGLE
+HashJoin: INNER
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpch_test.part (read rows: 20000)
+│       │   └── Scan: default.tpch_test.part (#1) (read rows: 20000)
 │       └── Probe
-│           └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│           └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
 └── Probe
-    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+    └── Scan: default.tpch_test.lineitem (#0) (read rows: 600572)
 
 #Q18
 query I
@@ -1977,13 +1974,13 @@ HashJoin: INNER
 │       ├── Build
 │       │   └── HashJoin: INNER
 │       │       ├── Build
-│       │       │   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│       │       │   └── Scan: default.tpch_test.lineitem (#3) (read rows: 600572)
 │       │       └── Probe
-│       │           └── Scan: default.tpch_test.orders (read rows: 150000)
+│       │           └── Scan: default.tpch_test.orders (#1) (read rows: 150000)
 │       └── Probe
-│           └── Scan: default.tpch_test.customer (read rows: 15000)
+│           └── Scan: default.tpch_test.customer (#0) (read rows: 15000)
 └── Probe
-    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+    └── Scan: default.tpch_test.lineitem (#2) (read rows: 600572)
 
 # Q19
 query I
@@ -2059,9 +2056,9 @@ where
 ----
 HashJoin: INNER
 ├── Build
-│   └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│   └── Scan: default.tpch_test.lineitem (#0) (read rows: 600572)
 └── Probe
-    └── Scan: default.tpch_test.part (read rows: 20000)
+    └── Scan: default.tpch_test.part (#1) (read rows: 20000)
 
 # Q20
 query I
@@ -2107,19 +2104,19 @@ HashJoin: RIGHT SEMI
 ├── Build
 │   └── HashJoin: INNER
 │       ├── Build
-│       │   └── Scan: default.tpch_test.nation (read rows: 25)
+│       │   └── Scan: default.tpch_test.nation (#1) (read rows: 25)
 │       └── Probe
-│           └── Scan: default.tpch_test.supplier (read rows: 1000)
+│           └── Scan: default.tpch_test.supplier (#0) (read rows: 1000)
 └── Probe
-    └── HashJoin: RIGHT SINGLE
+    └── HashJoin: INNER
         ├── Build
         │   └── HashJoin: LEFT SEMI
         │       ├── Build
-        │       │   └── Scan: default.tpch_test.part (read rows: 20000)
+        │       │   └── Scan: default.tpch_test.part (#3) (read rows: 20000)
         │       └── Probe
-        │           └── Scan: default.tpch_test.partsupp (read rows: 80000)
+        │           └── Scan: default.tpch_test.partsupp (#2) (read rows: 80000)
         └── Probe
-            └── Scan: default.tpch_test.lineitem (read rows: 600572)
+            └── Scan: default.tpch_test.lineitem (#4) (read rows: 600572)
 
 # Q21
 query I
@@ -2173,17 +2170,17 @@ HashJoin: RIGHT ANTI
 │       │       │       ├── Build
 │       │       │       │   └── HashJoin: INNER
 │       │       │       │       ├── Build
-│       │       │       │       │   └── Scan: default.tpch_test.nation (read rows: 25)
+│       │       │       │       │   └── Scan: default.tpch_test.nation (#3) (read rows: 25)
 │       │       │       │       └── Probe
-│       │       │       │           └── Scan: default.tpch_test.supplier (read rows: 1000)
+│       │       │       │           └── Scan: default.tpch_test.supplier (#0) (read rows: 1000)
 │       │       │       └── Probe
-│       │       │           └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│       │       │           └── Scan: default.tpch_test.lineitem (#1) (read rows: 600572)
 │       │       └── Probe
-│       │           └── Scan: default.tpch_test.orders (read rows: 150000)
+│       │           └── Scan: default.tpch_test.orders (#2) (read rows: 150000)
 │       └── Probe
-│           └── Scan: default.tpch_test.lineitem (read rows: 600572)
+│           └── Scan: default.tpch_test.lineitem (#4) (read rows: 600572)
 └── Probe
-    └── Scan: default.tpch_test.lineitem (read rows: 600572)
+    └── Scan: default.tpch_test.lineitem (#5) (read rows: 600572)
 
 # Q22
 query I
@@ -2229,12 +2226,8 @@ HashJoin: RIGHT ANTI
 ├── Build
 │   └── RangeJoin: INNER
 │       ├── Left
-│       │   └── Scan: default.tpch_test.customer (read rows: 15000)
+│       │   └── Scan: default.tpch_test.customer (#1) (read rows: 15000)
 │       └── Right
-│           └── Scan: default.tpch_test.customer (read rows: 15000)
+│           └── Scan: default.tpch_test.customer (#0) (read rows: 15000)
 └── Probe
-    └── Scan: default.tpch_test.orders (read rows: 150000)
-
-
-statement ok
-set enable_runtime_filter = 0;
+    └── Scan: default.tpch_test.orders (#2) (read rows: 150000)
diff --git a/tests/suites/0_stateless/00_dummy/00_0002_dummy_select_py.py b/tests/suites/0_stateless/00_dummy/00_0002_dummy_select_py.py
index 7c96bb83c511..e6057f5beb17 100755
--- a/tests/suites/0_stateless/00_dummy/00_0002_dummy_select_py.py
+++ b/tests/suites/0_stateless/00_dummy/00_0002_dummy_select_py.py
@@ -5,7 +5,7 @@
 import signal
 
 import boto3
-from moto import mock_s3
+from moto import mock_aws
 
 CURDIR = os.path.dirname(os.path.realpath(__file__))
 sys.path.insert(0, os.path.join(CURDIR, "../../../helpers"))
@@ -40,7 +40,7 @@ def save(self):
         s3.put_object(Bucket="mybucket", Key=self.name, Body=self.value)
 
 
-@mock_s3
+@mock_aws
 def test_my_model_save():
     conn = boto3.resource("s3", region_name="us-east-1")
     # We need to create the bucket since this is all in Moto's 'virtual' AWS account