All pull requests

Pull requests

• Fix #638: run module cleanup via atexit when duckdb leaves sys.modules

  #647·AnnasMazhar opened Oct 5, 2026·

  ·
  1

• Add DuckDBPyRelation.positional_join()

  #646·teaguesterling opened Oct 3, 2026·

  ·

• Add connection keyword to read_json and write_csv stubs

  #643·tojacob03 opened Sep 29, 2026·

  ·
  1

• Honor overwrite=False for existing parquet/csv files

  #642·00200200 opened Sep 29, 2026·

  ·
  2

• [duckdb-labs bot] Bump DuckDB submodule

  submodule-bump
  #641·duckdblabs-bot opened Sep 28, 2026Collaborator·

  ·

• Run PRAGMA after earlier statements in one Python query

  #640·t-kalinowski opened Sep 26, 2026·

  ·

• [duckdb-labs bot] Bump DuckDB submodule

  submodule-bump
  #639·duckdblabs-bot opened Sep 26, 2026Collaborator·

  ·

• Do not push down inf/nan float literals in the polars lazy filter

  #637·breken-ai opened Sep 26, 2026·

  ·

• Respect time unit and time zone in polars datetime filter pushdown

  #636·breken-ai opened Sep 26, 2026·

  ·

• Rework linking of libduckdb.a AND lib_<name>_extension.a, moving away from building duckdb in-tree

  #635·carlopi opened Sep 24, 2026Member·

  ·
  1

• Quote the path in DataFrameReader.load when no format is given

  #626·Raaif-Yousuf opened Sep 20, 2026·

  ·

• Add row_groups_per_file to to_parquet/write_parquet

  #620·joaquinhuigomez opened Sep 15, 2026·

  ·

• Populate the HTTPException body

  #614·afeld opened Sep 9, 2026·

  ·

• CI: Python 3.15 builds

  #605·paultiq opened Aug 27, 2026Contributor·

  ·
  3

• feat(python): implement runtime connection context manager hooks #3152

  #602·abhibesu2018 opened Aug 23, 2026·

  ·

• Don't touch Python state from destructors that outlive the interpreter

  #601·alexander-beedie opened Aug 21, 2026·

  ·
  1

• Add query appender

  #587·daviddallakyan2005 opened Aug 16, 2026·

  ·

• Add Appender.append_chunk and nested type support

  #586·daviddallakyan2005 opened Aug 16, 2026·

  ·

• Add Python Appender for scalar create, append, flush and close

  #585·daviddallakyan2005 opened Aug 16, 2026·

  ·

• Return real row counts from cursor.rowcount for CHANGED_ROWS statements

  #578·colin-k-rogers opened Aug 4, 2026·

  ·
  1

• fix(typing): read_json's stub only allowed a single path, not a list

  #577·aleks-drozy opened Aug 2, 2026·

  ·
  1

  1

• Fix and conform Spark API's show() and UDF register().

  #552·Shooter23 opened Jul 18, 2026·

  ·

• Type SQLExpression and spark expr() as LiteralString

  #540·NickCrews opened Jul 13, 2026Contributor·

  ·

• Only require numpy for the vectorized UDF path

  #538·yvaishalirao opened Jul 9, 2026·

  ·
  1

• Add support for path-like file name in to_csv and to_parquet

  #527·leostimpfle opened Jul 2, 2026Contributor·

  ·
  6