duckdb 0.9.1

0.9.1 Bugfix Release

on Python PyPI

This is a bug fix release for various issues discovered after we released 0.9.0. There are no new features, just bug fixes. Database files created by DuckDB v0.9.0 can be read by DuckDB v0.9.1 (i.e. v0.9.1 is backwards compatible with v0.9.0 and vice versa).

What's Changed

• Fix missing header in nightly GCC build by @Maxxen in #9109
• Add scripts for cancelling/rerunning workflows by @Mytherin in #9120
• Update R-CMD-Check workflow by @Tmonster in #9128
• CSV error message fix by @pdet in #9132
• Throw SerializationException instead of InternalException by @Maxxen in #9130
• Radix HT: Get max_threads from config instead of system by @lnkuiper in #9129
• Wrapping CSV State Machine by @pdet in #9131
• [CSV] Properly skip new lines mid csv file by @pdet in #9139
• [PythonDev] Rework duckdb_cursor fixture in pytest by @Tishj in #9140
• Update logo in README by @szarnyasg in #9147
• Bump uncovered lines by @carlopi in #9102
• Python: Uniform DuckDB SQL version to v0.X.Y by @carlopi in #9095
• [PythonDev] Don't dereference None when creating pandas dataframe by @Tishj in #9127
• fix: fix some typo by @shizuocheng in #9156
• [Swift] fix row count and index mapping by @tcldr in #9149
• Run Single-Threaded CSV Parser when only one thread is available by @pdet in #9148
• [ODBC] Add missing fields to SQLColAttribute by @maiadegraaf in #9146
• Merge main into feature by @Mause in #9163
• [Python] Pyarrow integer filter pushdown bug fix by @Tishj in #9155
• [CSV] Be sure to run header detection even if a header is defined by the user by @pdet in #9168
• Avoid (Try)AutoLoad logic if extension already loaded by @carlopi in #9177
• Fix fuzzer exclusion by @Tmonster in #9103
• [PythonDev] Fix #duckdb-internal/418 by @Tishj in #9184
• chore: skip arrow union tests on pyarrow<11 by @gforsyth in #9186
• Fix issue related to map() and Dictionary Vectors. by @Tishj in #9046
• Issue #9183: Arbitrary ASOF Conditions by @hawkfish in #9188
• Add instructions for jdk installed via brew by @nicku33 in #9192
• Fix nits in API comments by @szarnyasg in #8768
• Internal #445: Destruct All Aggregates by @hawkfish in #9202
• Don't initialize data chunk in ExpressionExecutor if it's just a BoundReferenceExpression by @lnkuiper in #9213
• [ODBC] Add info to fatal error in compiler by @maiadegraaf in #9218
• Rewrite filter remover arrow test by @lnkuiper in #9217
• [ODBC] Add Unittests to Windows CI Run by @maiadegraaf in #9216
• Add missing parameter defaults for create_function in duckdb-stubs by @earwig in #9224
• Several fixes related to vector_size=2 nightly build by @pdet in #9166
• Avoid keeping read-only transactions stored in old_transactions by @Mytherin in #9236
• Fix reusable workflow for OOTE building by @samansmink in #9229
• [Parquet] Implement reading byte stream split encoded data by @adamreeve in #9240
• Bug report template: ask for imports to be included by @szarnyasg in #9197
• Remove bundled sqlite, does not seem to be used anywhere by @hannes in #9245
• update azure extension by @samansmink in #9243
• Fix #8413: Avoid pushing collations to non-varchar columns, and cleanup around PushCollation by @Mytherin in #9242
• Fix #8624 - allow changing of schema and search_path parameters even if configuration is locked by @Mytherin in #9244
• [ODBC] SQLColAttribte: More tests by @maiadegraaf in #9200
• Fix for creating ephemeral schema on readonly by @nicku33 in #9261
• Wasm extensions Makefile changes by @carlopi in #9269
• Fix #9241 - avoid pushing filters through DISTINCT ON by @Mytherin in #9270
• Fix #7880 part 1 - rebind expression during alias replacement instead of copying the already bound expression by @Mytherin in #9274
• ICU: Remove unused variables by @carlopi in #9275
• Fix #9252 - avoid overwriting start value as part of INCREMENT BY in sequence creation by @Mytherin in #9272
• [Python] Add back value_counts by @Tishj in #9278
• [SparkAPI] Add read.json and read.parquet + some unhappy path testing by @Tishj in #9276
• [Arrow][Dev] Make each produced Array own its own memory. by @Tishj in #9254
• Properly register all JSON cast functions so the binder can disambiguate by @lnkuiper in #9209
• [Dev] Fix test_map_vector_types.test by @Tishj in #9277
• Issue #9280: Parquet TIME_TZ Reading by @hawkfish in #9283
• Update iceberg extension by @samansmink in #9284
• Fix prepare statement issue by @taniabogatsch in #9288
• [Arrow] Fix issue with scanning of UNION of STRUCTS by @Tishj in #9291
• Remove nodejs client by @hannes in #9287
• Clear cached buffers before emitting next chunk by @pdet in #9295
• [Dev] Restore _extension_client_tests.yml syntax by @carlopi in #9300
• Fix MacOS exception catching by @carlopi in #9286
• Unify replacement scan filename parsing by @lnkuiper in #9273
• Extension install fixes by @carlopi in #9294
• Re-add windows extension builds for R by @hannes in #9302