duckdb 1.1.1

v1.1.1 Bugfix Release

on Python PyPI

This is a bug fix release for various issues discovered after we released 1.1.0. There are no new major features, just bug fixes. Database files created by DuckDB since v0.9.* can be read by DuckDB v1.1.1.

What's Changed

• [Python] Fix a crash related to handling of the f_locals by @Tishj in #13821
• [Dev] Remove unnecessary parameter from BufferHandle constructor by @Tishj in #13823
• When vacuuming, immediately checkpoint the vacuumed row groups instead of scheduling a checkpoint task by @Mytherin in #13825
• Check for unquoted errors after finishing up a buffer. by @pdet in #13826
• Issue #13813: TIMETZ Uninvertible Casts by @hawkfish in #13820
• add cardinality for cross product and propositional join by @jeewonhh in #13818
• Add requested_schema argument to PyCapsule interface by @WillAyd in #13802
• Throw at double/float cast to decimal if it does not fit by @pdet in #13745
• [Fix] C API - Correct type comparison in MAP value functions by @taniabogatsch in #13844
• [C-API Dev] Fix up "Test DataChunk populate ListVector in C API" test by @Tishj in #13839
• [Dev] Minor cleanup to BufferManager and BlockHandle API by @Tishj in #13838
• Small fixes for prompt of sniff_csv by @pdet in #13843
• [BufferManager] Fix duckdb_memory() reporting wrong size for temporary_storage_bytes by @Tishj in #13837
• Fix for internal error when using rejects tables and adding implicit cast for boolean values. by @pdet in #13849
• [Fix ART] Correct prefix transformation by @taniabogatsch in #13858
• [Python] Fix issue causing an exception when creating a duckdb.StarExpression without an exclude_list by @Tishj in #13881
• Fix issue with cgroups/slurm variables: skip if memory limit cannot be parsed, and only run this on Linux by @Mytherin in #13879
• Explicit windows-2019 instead of windows-latest by @carlopi in #13883
• CI/Windows: Drop redundant package by @szarnyasg in #13874
• Preserve operator in BindWithCTE by @lnkuiper in #13873
• Update description of 'max_temp_directory_size' by @szarnyasg in #13724
• 13810 unnest cross join error by @Tmonster in #13878
• Tweak allocation purging by @lnkuiper in #13877
• Copy From Database - create a balanced UNION ALL tree instead of a depth-first union all tree to avoid stack overflows by @Mytherin in #13889
• [Python] Fix issue related to the GIL when using execute with multiple statements by @Tishj in #13893
• Fix #13880: correctly name http_proxy_password setting by @Mytherin in #13890
• Fix #13872: duckdb_result_return_type is not deprecated, and group together deprecated functions by @Mytherin in #13891
• [Python] Add python_scan_all_frames to opt-in to scanning all frames (< 1.1 behavior) by @Tishj in #13896
• Improve error on enums by @carlopi in #13885
• Handle extension ABI mismatches in a forward & backward compatible way by @carlopi in #13894
• Fix #13824: min() max() varchar column use default collation by @flashmouse in #13909
• Fix issue in casting 2 byte BIT -> BLOB by @Tishj in #13908
• add missing azure secret providers for autoloading by @samansmink in #13902
• Remove buffer_manager_allocate.patch and bump spatial by @carlopi in #13895
• [Python] Improve install_extension to support repository/repository_url and version by @Tishj in #13876
• fix REGEX not supported anymore for valid queries (only statement error) #2889 by @hmeriann in #13633
• [CI] Invert operations on OSX.yml, deploying nightly artifacts before test by @carlopi in #13927
• [Python][Jupyter] Don't use ExplainFormat::HTML for explain('analyze') by @Tishj in #13928
• Only bind to SQL value functions if there is no alias with this name present we can bind to instead by @Mytherin in #13925
• Improve logic for remote extension install on Windows by @carlopi in #13929
• CI: Trigger actions for labeled discussions by @szarnyasg in #13937
• [Swift] Update README.md in Swift repo by @tcldr in #13955
• [Dev] Add exclusion for pybind11 internal _pybind11_conduit_v1_ method by @Tishj in #13961
• Create a balanced union tree, also for export by @Mytherin in #13956
• Increment julia version by @Mytherin in #13966
• Fix #13585 - transform from or select first based on order specified by the user by @Mytherin in #13959
• Fix Cross Product Cardinality by @jeewonhh in #13954
• Do not run the date/timestamp format sniffer if they are set by the user by @pdet in #13960
• Modify create_art_varchar.benchmark so it passes weekly regressions by @Tmonster in #13968
• Fix data race when writing GeoParquet by @Maxxen in #13962
• Check vector type in GetVectorScanType to avoid concurrent race when updating validity by @Mytherin in #13965
• Parser Keyword Category Search by @pdet in #13875
• Escape quotes in FTS by @lnkuiper in #13917
• Fix #13941: fix error message in appender by @Mytherin in #13957
• fix: remove http prefix from proxy value when present by @dylanspag-lmco in #13973
• Use defaults when serializing copy to file by @Mytherin in #13978
• Fix #13933: disable join filter pushdown when a join is performed over collated columns by @Mytherin in #13958
• Fix partitions on wide tables by @piever in #13988
• [Fix] Throw exception for UNNEST in lambdas by @taniabogatsch in #13969
• Fixing some parquet issues found by fuzzing by @hannes in #13964
• fix julia partitions for streaming result by @piever in #14000
• More descriptive Parquet created_by with version and source hash by @hannes in #13992
• Decimal downcast limit check by @pdet in #13922
• [C API] Add SQLNULL to the duckdb types by @taniabogatsch in #13999
• Fix crash in the shell caused by printing blobs that failed to cast by @Maxxen in #13983
• [Binding] Always try binding with the schema of the UserType first if it's set by @Tishj in #13995
• [Arrow] Only produce 'arrow.json' Extension types when arrow_lossless_conversion is enabled. by @Tishj in #13989
• Bugfixes by @lnkuiper in #13982
• Disable swift linux tests by @Mytherin in #14019
• fix minor typos in comments of aggregate function tests by @era127 in #14007
• [CSV Sniffer] Slight change of rules for dialect detection by @pdet in #14013
• [Test Fix] Add noforcestorage to in-memory tests by @taniabogatsch in #14016
• Fix #14020: fix off-by-one in RLE compression: avoid flushing when last_seen_count == 0 which can happen if a column has exactly 2^16 (65535) repeated values by @Mytherin in #14030
• Adds Julia support for scalar UDFs by @drizk1 in #14024
• Proper NULL handling in special json extraction functions by @lnkuiper in #14032
• Fix #13993 - avoid disabling optimizers for SET VARIABLE by @Mytherin in #14028
• [Arrow] Make unknown Arrow extensions throw at scan instead of bind by @pdet in #14015
• Improve assertion macros by @c-herrewijn in #14033
• [Arrow] Move ArrowUtil to its own file by @Tishj in #13990
• [CSV Sniffer] Verify validity of header before value access by @pdet in #14040
• Fix #14026: all TIMESTAMP_xx cannot cast to TIME directly by @flashmouse in #14045
• Only merge distinct stats if both sides have distinct stats available by @Mytherin in #14046
• Avoid http-redirect in README by @carlopi in #14029
• fix: make ArrowArrayWrapper::GetNextChunk() virtual by @rustyconover in #14003
• Issue #13655: MEDIAN Even Tests by @hawkfish in #13722
• Fix #13934: use CreateSortKeyWithValidity to correctly handle NULL in all calls to arg_max by @Mytherin in #14053
• Upgrade MySQL/Postgres extensions by @Mytherin in #14054
• [Union Reader] Early-out on readers of files that do not have data by @pdet in #14050
• Issue #13899: AsOf Unrelated Pushdown by @hawkfish in #13975
• Feature #3128: 2024b Time Zones by @hawkfish in #14061
• bump spatial by @Maxxen in #14062
• Fixup StagedUpload invocation via workflow_call by @carlopi in #14065
• Skip polars test in 3.7 due to missing PanicException by @carlopi in #14066
• Add building, testing and distributing for Python 3.13 by @carlopi in #14067
• bump sqlsmith and azure versions by @samansmink in #14069
• bump substrait by @pdet in #14055
• CIBW_SKIP also musllinux on Python 3.7 or Python 3.8 by @carlopi in #14074

Full Changelog: v1.1.0...v1.1.1