temp

run_sql_test.py

@@ -1,22 +1,41 @@
 import duckdb
 import json
 
-duckdb_file = "db.duckdb"
-duckdb_query = "queries/q1.sql"
+import os
 
-con = duckdb.connect(duckdb_file)
-with open(duckdb_query, "r") as sql_file:
+
+duckdb_file = "db_small.duckdb"
+duckdb_query_path = "queries"
+
+RUN_COUNT = 13
+WARMUP_COUNT = 2
+
+for query_file in os.listdir(duckdb_query_path):
+  print("Running {}", query_file)
+
+  con = duckdb.connect(duckdb_file)
+  duckdb_query = os.path.join(duckdb_query_path, query_file)
+  with open(duckdb_query, "r") as sql_file:
     sql_query = sql_file.read()
 
-con.sql("PRAGMA enable_profiling='json'")
-# print(sql_query)
-res = con.sql(sql_query).to_df()
-print(res['explain_value'][0])
-res = (res['explain_value'][0])
+  con.sql("PRAGMA enable_profiling='json'")
+  # Run the query a few times to warm up memory
+
+  print("Warmup: {}", sql_query)
+  for _ in range(WARMUP_COUNT):
+    con.sql(sql_query)
+
+  times = []
+
+  print("Running test: {}", sql_query)
+  for i in range(RUN_COUNT):
+    output = con.sql(sql_query).to_df()
+    res_str = output['explain_value'][0]
+    res_dict = json.loads(res_str)
+    times += [float(res_dict["timing"])]
 
-res = json.loads(res)
-total_time = res["timing"]
-# seq_time = res["children"]
-print(f"Total Time: {total_time}")
+  times = sorted(times)
+  print("Query file: {}, min_time: {}, max_time: {}, avg_time: {}, median_time: {}", query_file, times[0], times[-1], sum(times)/RUN_COUNT, times[RUN_COUNT/2])
+  con.close()
 
-con.close()
+#  print(res['explain_value'][0])