GeezerVison

Pull requests

• Hold one connection claim from submission to settling in execute, executemany and the DB-API Arrow fetch

  #765·el-marvin opened Nov 10, 2026Collaborator·

  ·
  9

• Read a qualified name in create_view and to_view, and lift the three part limit on relation writes

  #764·el-marvin opened Nov 10, 2026Collaborator·

  ·
  1

  14

• Render the grouped profile a 2.x engine writes in duckdb.query_graph

  #763·el-marvin opened Nov 10, 2026Collaborator·

  ·
  14

• Keep two tests stable under xdist and machine load, and judge the GIL scan tests one way

  #760·el-marvin opened Nov 10, 2026Collaborator·

  ·
  5

• Add site-packages/.duckdb_extensions to every connection's extension search path

  #759·el-marvin opened Nov 10, 2026Collaborator·

  ·
  15

• Release the type constants a held module kept alive, and silence nanobind's exit leak report in builds without assertions

  #758·el-marvin opened Nov 10, 2026Collaborator·

  ·
  28

• Own Python UDF callables per database and let the collector reclaim their cycles

  #743·evertlammerts opened Nov 6, 2026Member·

  ·
  11

• Refuse relational calls on a CONNECT-ed connection before they run

  #742·evertlammerts opened Nov 6, 2026Member·

  ·
  7

• Read a KeyError or None from an fsspec modified() as an unavailable modification time, which read_blob reports as NULL

  #735·PeterVanHolland opened Nov 4, 2026·

  ·

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

  #637·breken-ai opened Sep 26, 2026Contributor·

  ·

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

  #636·breken-ai opened Sep 26, 2026Contributor·

  ·

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

  #635·carlopi opened Sep 24, 2026Member·

  ·
  1

• 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·

  ·

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

  #602·abhibesu2018 opened Aug 23, 2026·

  ·

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

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

  ·
  1

  1

• 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

• [codex] Use non-view registry for Python register

  #470·jprafael opened May 25, 2026·

  ·

• [spark] Add 8 aggregate functions (count_distinct, collect_set, count_if, max_by/min_by, bool_and/or, kurtosis)

  #450·aaron-ang opened May 11, 2026·2 tasks done·

  ·
  1

• feat: Support replacing existing table relations

  #430·Zirak opened Apr 20, 2026·

  ·

• Further develop the Catalog API to match Spark better

  #427·joaoe opened Apr 15, 2026·

  ·

• [PySpark] feat: add explode and explode_outer to experimental PySpark API

  #415·tinovyatkin opened Apr 3, 2026·

  ·

• feat: implement templating system based on tstrings

  #409·NickCrews opened Mar 29, 2026Contributor·

  ·
  6