v2.0.0b0

Highlight

• 🚀 87x performance boost query on Pandas DataFrame, See Benchmark

• ⬆️ Upgrade ClickHouse engine to 24.5

• 🐍 Query on Pandas DataFrame, ArrowTable, Dict or Any Python object Directly!

import chdb
import pandas as pd
df = pd.DataFrame(
    {
        "a": [1, 2, 3, 4, 5, 6],
        "b": ["tom", "jerry", "auxten", "tom", "jerry", "auxten"],
    }
)

chdb.query("SELECT b, sum(a) FROM Python(df) GROUP BY b ORDER BY b").show()

• For more: test_query_py.py.

Changes

• Use in-process instead of embedded by @auxten in #222
• 226 crash on mergetree background task cleanup by @auxten in #227
• Replace icu with utf8proc by @auxten in #228
• Start agg step stream according to agg_col_cost and group_by_keys_cost by @auxten in #230
• Fix various issues like 190, 229 by @auxten in #231
• Before 2.0 release Patch #1 by @auxten in #232
• Fix python version check in CMakeLists.txt by @auxten in #233
• Fix mistake config.h by @auxten in #234
• Add tests on query python objects by @auxten in #235
• Add docs for Python Table Engine by @auxten in #238

Full Changelog: v1.4.1...v2.0.0b0