DuckDB 1.3.0 "Ossivalis"

This release of DuckDB is named "Ossivalis" after Bucephala Ossivalis, an ancestor of the Goldeneye duck that lived Millions of years ago.

Please also refer to the announcement blog post: https://duckdb.org/2025/05/21/announcing-duckdb-130

What's Changed

• V1.2 histrionicus by @Mytherin in #16070
• V1.2 histrionicus by @Mytherin in #16072
• unittests: clear test directory after every test by @Mytherin in #16053
• Benchmark runner: catch and log errors + add support for retry load N syntax by @Mytherin in #16054
• Throw an error when unsupported commands are used in concurrentloop by @Mytherin in #16009
• Remove extension definitions to prevent re-compilation of the entire system on commit by @Mytherin in #15955
• Display schema information of currently selected database only by @ashwaniYDV in #15815
• Issue #14366: Average Intervals by @hawkfish in #15864
• Internal #2176: Temporal AVG by @hawkfish in #15661
• discussions #15981: remove confusing comment in "duckdb/tools/shell/shell.cpp" by @komainu8 in #15984
• Fix #15466 Transform LIMIT or OFFSET first based on order specified in prepared statement by @ashwaniYDV in #15484
• Bitpacking mode info by @arjenpdevries in #15623
• Sniff Timestamp_TZ from CSV FIles by @pdet in #15730
• [no-op] Add documentation for filesystem read behavior by @dentiny in #15937
• Accept "Auto" as date/timestamp format by @pdet in #15808
• Parquet Reader Cleanup: Move ColumnReaders to separate files by @Mytherin in #16092
• Parquet Reader: Move decoding logic into separate Decoder classes by @Mytherin in #16100
• BundleStaticLibs to be also triggered by InvokeCI by @carlopi in #16107
• Parquet Reader: Split DeltaLengthByteArray decoder from DeltaByteArray, and read the strings in a streaming manner by @Mytherin in #16105
• Parquet Dictionary reader: set NULL values as the last value in the dictionary by @Mytherin in #16106
• Parquet Reader: Share ResizeableBuffers across decoders, and unify Plain/PlainReference by @Mytherin in #16113
• Using GitHub ARM runners for Linux CLI builds by @hannes in #16119
• Parquet Reader: Implement dedicated Skip method by @Mytherin in #16117
• Use ColumnSegment::FilterSelection and SelectionVector for filtering in Parquet scans by @Mytherin in #16126
• [Dev] Fix output (long lines > 333 characters) getting truncated in shell by @Tishj in #16128
• Adaptive table filter: initialize filter order based on heuristics by @Mytherin in #16127
• Feature #16044: TimeZone Offset Seconds by @hawkfish in #16048
• ATTACH OR REPLACE database to allow swapping of new data. by @xevix in #15355
• [Dev] Remove upsert_conflict_in_different_chunk.test by @Tishj in #15980
• [Dev] Fix issue related to unpacked columns and the NOT operator by @Tishj in #15534
• [Julia] Add support for named params in prepared statements by @tqml in #15621
• Use Adaptive Filters in the Parquet reader by @Mytherin in #16133
• Parquet reader: push table filters directly into dictionaries by @Mytherin in #16136
• Parquet reader: Plain templates - make CHECKED a template parameter, and use memcpy/bulk skip when reading/skipping without defines by @Mytherin in #16141
• Parquet reader: only set invalid entry in the dictionary when the column has defines by @Mytherin in #16144
• Add uniq_ptr_cast for interpreted benchmark. by @Tmonster in #16151
• Hopefully fixing ci runs by @hannes in #16150
• Removed the last CI job that used the Ubuntu 18 setup by @hannes in #16155
• Parquet Reader: Split CreateReader into two separate stages - ParseSchema and CreateReader by @Mytherin in #16161
• Have CSV Parellel tests on CI again by @pdet in #16164
• [Python][Dev] Bump the minimum pybind11 version from 2.6 to 2.9 by @Tishj in #16159
• Add StackTraces to FatalExceptions by @NiclasHaderer in #16158
• Rework invoke by @carlopi in #16108
• Adds pre-optimization hooks for DuckDB by @NiclasHaderer in #16115
• Unify behavior of range/generate_series with PostgreSQL by @kryonix in #15935
• [CI] Avoid Linux CLI jobs to fail-fast by @carlopi in #16173
• Parquet: Add dedicated Select method that can be used to push selection vectors into the read by @Mytherin in #16174
• Unvendor ICU by @m-kuhn in #16176
• Parquet reader: batch check if buffer is available in RLEBpDecoder by @Mytherin in #16185
• Parquet Reader: for DeltaLengthByteArray encoding, directly refer to strings from the block without copying by @Mytherin in #16186
• unified names for duckdb-extensions by @hmeriann in #16179
• Only delete test directory when --test-temp-dir is not specified by @Mytherin in #16192
• Fix #16163: COLUMNS should not treat identifiers as strings by @Mytherin in #16193
• Parquet reader: Avoid applying bloom filters if we are casting columns by @Mytherin in #16194
• Pretty print sniffer values by @pdet in #16182
• V1.2 histrionicus by @Mytherin in #16191
• Bump Julia by @Mytherin in #16199
• Ensure that dependent targets are present after find_package. by @BillyONeal in #16197
• Concurrency groups for R and Wasm by @hmeriann in #16201
• Parquet Writer Cleanup: Move ColumnWriters to separate files by @Mytherin in #16202
• [fix] Use bigobj when building with MSVC by @m-kuhn in #16200
• Improve performance of UNNEST/UNPIVOT by using selection vectors to unnest multiple lists at once by @Mytherin in #16210
• Add the TRY expression by @Tishj in #15939
• [Python][Dev] Replace the default connection when it's closed by @Tishj in #16160
• Use steady clock for profiler by @dentiny in #16198
• Add parallel memset when building hash join table by @hehezhou in #16172
• Avoid unnecessarily projecting the input columns of the UNPIVOT operator in the UNNEST by @Mytherin in #16221
• Left join push down optimization by @Damon07 in #15881
• Do In-Filter pushdown in PyArrow by @pdet in #16224
• Use _win32 with MSVC by @cfis in #16235
• Fix Python 3 executable name on Windows by @cfis in #16236
• Fix -std=c++11 by @cfis in #16237
• Issue #8265: AsOf Nested Loop by @hawkfish in #16218
• Include extension_util.hpp in libduckdb by @mlafeldt in #16255
• Generalize rowid into the concept of virtual columns, and make filename a virtual column in the Parquet/CSV/JSON readers by @Mytherin in #16248
• Modify histogram test to more fuzzily check boundaries since the test can be inconsistent on different platforms by @Mytherin in #16261
• [Dev] Fix issue in TRY expression with dictionary_expression Vector verification by @Tishj in #16262
• [Python Dev] Add the correct variant of the -std=c++11 flag based on the compiler (MSVC or not) by @Tishj in #16267
• Fix extension install mode null by @samansmink in #16268
• A little cleanup. by @JasonPunyon in #16028
• Improve Parquet writer performance by @lnkuiper in #16243
• Merge v1.2-histrionicus into main by @Mytherin in #16284
• Many reclaim storage fixes by @taniabogatsch in #15825
• Arena allocator for MinMaxN and skip NULLs when creating enum by @lnkuiper in #16246
• Add pragma to truncate duckdb log storage by @samansmink in #16274
• Some more Parquet writer performance improvements by @lnkuiper in #16287
• Do duckdb_extract_statements to be able to execute pivot in ADBC by @pdet in #16162
• [Dev] Improve/Add handling of escapes in VARCHAR -> list/struct/map and align behavior by @Tishj in #15944
• make ValidityMask::RowIsValidUnsafe really unsafe by @xuke-hat in #16302
• Multi File Reader Rework: Add MultiFileReaderFunction that is used to wrap a single-file reader, and use it for the Parquet reader by @Mytherin in #16299
• [Python Dev] Add support for fully qualified references in .table(...) method by @Tishj in #16291
• [Dev] MultiFileReader - Add to the column_indexes for file_row_number by @Tishj in #16311
• Parquet reader performance by @lnkuiper in #16315
• Bump Julia FixedPointDecimals dependency version by @mbarbar in #16323
• Merge V1.2 histrionicus into main by @Mytherin in #16324
• Add new recursive semantics (USING KEY) by @cryoEncryp in #12430
• fix: add StringStats::SetMaxStringLength by @rustyconover in #16326
• Fix decorrelation of WITH USING KEY by @kryonix in #16330
• Issue #16250: Window Range Performance by @hawkfish in #16320
• Verify UTF-8 in DeltaLengthByteArrayDecoder and speed it up by @lnkuiper in #16328
• Add missing include by @Mytherin in #16342
• [chore] No ccache for OSX Python by @carlopi in #16348
• Linux CLI: override platform for ARM manylinux by @carlopi in #16347
• docs: tweak explanation of median for even cardinality inputs by @NickCrews in #13726
• [parquet] Fix implicit idx_t to int64_t conversion flagged by clang-tidy by @carlopi in #16368
• Improve error message on failure to install local extension by @carlopi in #16371
• MAIN_BRANCH_VERSIONING: main branch to get descriptors like v1.3.0-dev1234 instead of v1.2.1-dev1234 by @carlopi in #16366
• Parallel HT Zeroing: Set entries_per_task so that there are 4x more tasks than threads by @gropaul in #16301
• Internal #2176: SUMMARIZE Temporal Types by @hawkfish in #16095
• [SwiftRelease CI] fetch tags before checking there is already a tag with the same name by @hmeriann in #16376
• Push Select into ArrayColumnData to avoid scanning arrays that are not required for the query by @Mytherin in #16356
• Revert "Linux CLI: override platform for ARM manylinux" by @carlopi in #16374
• Rework CSV Reader: use the new MultiFileReaderFunction interface by @Mytherin in #16349
• Add connection and transaction identifiers by @samansmink in #16296
• Add parquet 'unknown' logical type by @hannes in #16378
• Internal #4287: INTERVAL Times DOUBLE by @hawkfish in #16386
• pb/compressed vector serialization by @peterboncz in #16066
• Fix issue #16377 by @kryonix in #16391
• Read support for Parquet Float16 by @hannes in #16395
• MAIN_BRANCH_VERSIONING: Adopt also for Python build and amalgamation by @carlopi in #16400
• Fuzzer Fix: Fix Avg for NULL cast to TIMESTAMP by @Tmonster in #16394
• [FriendlySQL] Expand functionality of the Unpacked COLUMNS expression by @Tishj in #16290
• Python Client: Faster Python Object Conversion by @Mytherin in #16431
• Fixup #16400 by correctly passing down SETUPTOOLS_SCM_PRETEND_VERSION by @carlopi in #16435
• Issue #16250: Window Range Performance by @hawkfish in #16438
• Merge v1.2-histrionicus into main by @Mytherin in #16439
• MAIN_BRANCH_VERSIONING: Add also prefix_version by @carlopi in #16441
• [no-op] Remove unused function GetValueRefUnsafe by @dentiny in #16440
• Filter Combiner Clean-up: move filter pushdown to separate functions, remove old commented out code by @Mytherin in #16443
• [Python] Add the SQLExpression method to the Expression API by @Tishj in #16424
• [Dev] Mention the problematic type in UNNEST BinderException by @Tishj in #16429
• Merge v1.2 into main again by @Mytherin in #16447
• Filter Combiner: Allow rowid pushdown for IN/OR filters and pushdown for temporal types by @Mytherin in #16450
• Parquet: always launch max threads if we are scanning multiple files by @Mytherin in #16457
• fix documents of C functions by @yiyuanliu in #16357
• Add a TableFilterState for execution of table filters by @Mytherin in #16461
• Mirror discussions to the internal repository by @szarnyasg in #16464
• Rework JSON Reader: use the new MultiFileReaderFunction interface by @Mytherin in #16477
• Speed-up contains by using memchr on every iteration by @Mytherin in #16484
• Fix error cases by @Y-- in #16494
• Prevent external joins (if possible) by @lnkuiper in #16430
• Merge v1.2 into main by @Mytherin in #16517
• Optimize FSST decoding by @lnkuiper in #16508
• Extract subsystem by name by @dentiny in #16226
• Avoid throwing an exception (that is then swallowed) when computing compressed materialization over stats that are not set by @Mytherin in #16532
• Checksum backward compatibility by @lnkuiper in #16505
• Prefetch Parquet page header by @lnkuiper in #16507
• Let GitHub render *.test files as SQL by @mlafeldt in #16534
• Fix ADBC to properly quote table and schema names by @CurtHagenlocher in #16526
• Pass ClientContext to catalog initialize, and postpone index binding when replaying the WAL by @Mytherin in #16536
• Allow UNITTEST_ROOT_DIRECTORY to be configured through CMake by @Mytherin in #16540
• Internal #4347: ISO Year Week by @hawkfish in #16567
• throw() -> noexcept in skiplist by @r-barnes in #16548
• Fix test/sql/aggregate/aggregates/histogram_table_function.test to pass the Linux CLI (arm64) CI by @hmeriann in #16538
• feat: move GRANT from reserved to unreserved keyword by @stephaniewang526 in #16546
• Python test runner: Avoid enabling profiling when executing restart command by @Flogex in #16547
• Add duckdb_prepared_statements by @Tishj in #16541
• [minor] Keep bit type sanity check consistent by @dentiny in #16575
• Support CREATE TABLE AS ... WITH NO DATA by @hannes in #16586
• Parquet FLOAT16 - fix cast by @hannes in #16580
• remove invalid tokens from nanosecond example by @hamilton in #16577
• CrossVersion.yml: Add v1.2.1, v1.2-histrionicus and main by @carlopi in #16576
• Fix #16524: DEPENDENT_JOIN may not flatten by @flashmouse in #16537
• [Julia] Add support for appending duckdb List types by @era127 in #16512
• [PySpark] - Add expr function by @mariotaddeucci in #16468
• regex_replace no longer swallows regex errors by @hannes in #16380
• Parquet Writer Clean-up: Split CreateWriterRecursive into two methods, and use ParquetColumnData for writer as well by @Mytherin in #16592
• Bump Julia to 1.2.1 by @Mytherin in #16593
• Improved appender error message by @NiclasHaderer in #16599
• Change static variables to be on the stack instead by @Y-- in #16597
• Add support for RETURN_STATS to COPY by @Mytherin in #16595
• Better error messages for the CSV Scanner by @pdet in #16585
• Support Enum types in read_csv - Python by @pdet in #15710
• Fix CI Tidy by @pdet in #16610
• Add some minor helper functions (QueryResultIterator::IsNull and casts to MultiFileList/Reader by @Mytherin in #16611
• Add support for ALTER TABLE tbl SET PARTITIONED BY (key1, key2, ...) in the grammar by @Mytherin in #16612
• Issue template: direct UI issues to the UI repository by @szarnyasg in #16619
• [Dev] Make the various mappings in MultiFileReaderData typesafe by @Tishj in #16596
• Bump mbedtls to 3.6.2 and re-apply patches by @hannes in #16485
• Read and Write Complex Json from Arrow Types by @pdet in #16385
• Add Docker support for RISC-V CI with appropriate build commands by @mocusez in #16549
• Fix missing **kwargs in adbc_driver_duckdb.dbapi.connect() by @davlee1972 in #16637
• [Dev] Clean up and fix the CGroup memory/cpu limit discovery logic by @Tishj in #16608
• Expose Value::ToSQLString() in C API by @mt-caret in #16471
• Add the missing binding for json_serialize_sql by @liznear in #16666
• Do not create validity mask for non-null const vector by @xuke-hat in #16669
• Fix #16665: fix parquet multi_reader bloom_probe logic error by @flashmouse in #16677
• Add alias to catalog by @c-herrewijn in #16600
• Decouple physical operator ownership from operators by @taniabogatsch in #16545
• cmake: fix external icu by @autoantwort in #16676
• Character length and date functions by @hannes in #16653
• [Dev] Don't try to include third_party/mbedtls/VERSION with package_build.py by @Tishj in #16683
• Add -ui to CLI help text by @akx in #16626
• Fix alias of column reference lost in ReplaceProjectionBindings by @Damon07 in #16686
• Merge v1.2-histrionicus into main by @Mytherin in #16687
• Fix for GCC-4.8 by @Mytherin in #16690
• JSON Reader: make read_position atomic so this can be read by the progress bar while processing the JSON file by @Mytherin in #16692
• [Julia] support binding for vectors by @slwu89 in #16701
• Make CSV Parser strict_mode=True fail on a mix of new line delimiters. by @pdet in #15959
• [pypi] Fix cleanup logic for multiple branches by @hmeriann in #16634
• Add support for ALTER TABLE tbl SET SORTED BY (key1 DESC, key2, ...) in the grammar by @Mytherin in #16714
• RETURN_STATS: remove footer_offset, and emit written partition keys by @Mytherin in #16715
• In case all rows of a CSV batch are errors, we continue processing by @pdet in #16713
• add workaround for patching httpfs ext by @samansmink in #16722
• Implement UUID v7 by @dentiny in #15819
• Fix roundtripping of stringified nested types by @Tishj in #16304
• Add Notify External Repositories Workflow by @maiadegraaf in #16730
• Expose a selection vector and the Slice method to the C API by @joseph-isaacs in #16696
• Add support for tracking column_size_bytes and contains_nan in RETURN_STATS by @Mytherin in #16731
• Add support for WRITE_EMPTY_FILE option to COPY - which allows skipping of writing empty files by @Mytherin in #16737
• Parquet Writer: Truncate string stats for large strings, instead of bailing on writing stats by @Mytherin in #16736
• RLE compression - memset alignment bytes to zero when aligning the counts by @Mytherin in #16735
• Write UUID stats to Parquet files and support reading uuid stats by @Mytherin in #16744
• Add an initial value to list_reduce by @maiadegraaf in #16602
• shell: make -bail work for more errors by @mlafeldt in #16594
• Call Notify External Repositories from Invoke CI by @maiadegraaf in #16747
• JSON bugfixes by @lnkuiper in #16729
• Add support for dynamically providing extra info post-execution in table functions, and use this to emit the total number of files read by the MultiFileReader by @Mytherin in #16749
• [Python Dev] Fix the versioning of the nightly python builds by @Tishj in #16739
• shell: fix sometimes-uninitialized error by @mlafeldt in #16761
• Issue #16250: Window Range Performance by @hawkfish in #16765
• Avoid building Python 3.7 wheels also for Linux by @carlopi in #16769
• Pyodide 0.27.2: conditionally skip tests by @carlopi in #16772
• Push catalog lookups through an extensible EntryLookupInfo struct by @Mytherin in #16764
• Fix two minor problems with NotifyExternalRepositories / odbc by @carlopi in #16776
• update expected results reflecting the changes brought ups with Fix roundtripping of stringified nested types PR by @hmeriann in #16775
• Merge V1.2 -> Main by @pdet in #16751
• Add support for time travel syntax in the FROM clause by @Mytherin in #16774
• Python docs: List all join types by @szarnyasg in #16789
• [chore] NotifyExternalRepositories.yml: Fix endpoint to be pinged by @carlopi in #16793
• Remove delta from extensions built on a nightly basis (vs main branch) by @carlopi in #16795
• OSX.yml & Windows.yml: remove repository_dispatch, already handled by InvokeCI by @carlopi in #16796
• Make extensions be linked privitally into duckdb by @JAicewizard in #16726
• Add additional iterations to avoid assertion failure in TemporaryMemoryManager by @lnkuiper in #16801
• Change the STANDARD_MASK_SIZE calculation to use size of template type. by @sebastiaan-dev in #16807
• Fix nightly table sample error by @Tmonster in #16811
• Fix tidy by @pdet in #16805
• support 'categories' label in function catalog by @c-herrewijn in #15654
• regenerate function headers by @c-herrewijn in #16822
• Internal #4490: Window Jump Reset by @hawkfish in #16816
• Regression.yml: Actually checkout proper base.sha commit by @carlopi in #16824
• fix: drop useless python import by @yihong0618 in #16808
• NightlyTests.yml: Inline env variables into build command by @carlopi in #16817
• Benchmark runner summary by @hmeriann in #16759
• Add storage_version 66 for version 1.3.0 by @carlopi in #16800
• Revert "fix: drop useless python import" by @Mytherin in #16834
• [MultiFileReader] Rework MultiFileReader::FinalizeChunk to use Expressions by @Tishj in #16630
• Merge v1.2 into main by @Mytherin in #16832
• Fix NULL key handling in mark join by @xuke-hat in #16825
• compressed vector serialization fixes by @peterboncz in #16648
• really sorry about this by @peterboncz in #16840
• Fix Python docstrings for unique by @szarnyasg in #16845
• [MultiFileReader] Create "local" filters to hand to underlying readers by @Tishj in #16838
• Revert "Regression.yml: Actually checkout proper base.sha commit" by @Mytherin in #16860
• [ART] Immediately erase empty fixed-size buffers by @taniabogatsch in #16727
• Resolve defaults and column index map by pushing a Projection (instead of executing in the insert itself) by @Mytherin in #16867
• Fix issue with sorting dev versions in pypi_cleanup.py script to keep on PyPi the most recent dev versions by @hmeriann in #16873
• Allow filters to be pushed through joins if there are projection maps by @lnkuiper in #16871
• Expressions in create secret by @samansmink in #15801
• Python - Arrow IPC support in from_arrow by @pdet in #16821
• [ART] Introduce a new ARTScanner and make InitMerge and Vacuum iterative by @taniabogatsch in #16861
• Do not pushdown filters which bindings only match the right side of the left join by @Damon07 in #16880
• MultiFileReader Rework (part 17) - remove MultiFileReaderData - and move as much as possible out of the file readers by @Mytherin in #16882
• ICU: Unify TimeZone accessing code by @Mytherin in #16887
• Rework ICU age computation to convert to a timestamp and use the regular interval age computation by @Mytherin in #16889
• Reduce allocations during aggregations by @lnkuiper in #16849
• CI: Prevent marking issues as 'stale' if they have the 'no stale' label by @szarnyasg in #16903
• Add field name to log line which fails Parquet spec by @jsbali in #16862
• Internal #4490: Window Threading Cleanup by @hawkfish in #16879
• Adding gzip version of shell for linux/osx install script by @hannes in #16116
• Fix USING KEY reference error by @kryonix in #16906
• [Nested] Enable Varargs in LIST_CONCAT by @maiadegraaf in #16870
• Fix several issues with vsize=2, and move vsize=2 tests to Main.yml by @Mytherin in #16918
• C API comments: Fix a/an typos by @szarnyasg in #16925
• Reduce locking with FILE_SIZE_BYTES/ROW_GROUPS_PER_FILE in Parquet writer by @lnkuiper in #16928
• [Python] Fix annotation of condition argument in join so it accepts Expression by @MarcoGorelli in #16933
• Fix GCC 4.8 and add it back to Main workflow by @Mytherin in #16937
• Merge v1.2 into main again by @Mytherin in #16939
• MultiFileReader - Perform nested remapping of field indexes instead of relying on casts by @Mytherin in #16941
• Internal #4552: Short Circuit CSE by @hawkfish in #16931
• Add back manylinux extensions by @carlopi in #16944
• Run CI on merge group by @Mytherin in #16945
• Internal #4516: Interval BIGINT Variants by @hawkfish in #16904
• Split query string for multi-statement queries by @Mytherin in #16955
• Vector Verification: Rework to run based on env variable DUCKDB_DEBUG_VERIFY_VECTOR and move to Main.yml by @Mytherin in #16957
• Move the no string inline/alternative verify workflow to Main.yml by @Mytherin in #16958
• [Python] Tighten type annotations on shape and columns by @MarcoGorelli in #16948
• Pass down CMAKE_POLICY_VERSION_MINIMUM and fix for local development by @carlopi in #16953
• [ART] Use the ARTScanner for VerifyAllocations (make it iterative) by @taniabogatsch in #16946
• Move ThreadSanitizer test from nightly test to Main, and fix locking issue by @Mytherin in #16960
• Re-enable workflows to run on PRs by @Mytherin in #16961
• Fix for selecting NaN values from Parquet files by @Mytherin in #16962
• Move LatestStorage tests to NightlyRelease - and fix issue with overflow string blocks not being cleaned up correctly by @Mytherin in #16972
• Arena-allocate physical operators by @taniabogatsch in #16911
• Make file_row_number a virtual column, and support per-file virtual columns in the MultiFileReader by @Mytherin in #16979
• Add a setting scheduler_process_partial that allows partial scheduling of tasks in the background threads by @Mytherin in #16973
• Clean up format script, gather all files then run concurrently instead of running concurrently per directory by @Mytherin in #16988
• Add support for altering struct columns (adding fields, dropping fields, renaming fields) by @Mytherin in #17003
• Fix CSV fuzzer tests by @pdet in #16994
• [Fix] Keep original expression for macro + lambda's with subqueries by @taniabogatsch in #17020
• Detect when tables have been dropped or altered, and prevent deletes in this scenario by @Mytherin in #17018
• Update links pointing to duckdb.org by @szarnyasg in #16999
• Fix for joining on floating columns #16901 by @nickzoic in #16965
• fix: remove ununsed stream struct member from ArrowScanLocalState by @rustyconover in #17023
• [Dev] Use UnifiedVectorFormat instead of a flattened Vector in UpdateSegment::Update by @Tishj in #16974
• Remove Arrow Extenson from core extensions by @pdet in #17027
• Correctly propagate ClientContext to TaskExecutor by @ywelsch in #17026
• Issue #17001: AsOf memory Management by @hawkfish in #17028
• [MultiFileReader] Make it possible for the multi file reader to add a DeleteFilter to the BaseFileReader by @Tishj in #17032
• Add optional OVERRIDE_NEW_DELETE build parameter by @lnkuiper in #17035
• Clean-up virtual columns and make MultiFileReader::InitializeReader virtual by @Mytherin in #17038
• Allow a table to define their own row-id columns for delete/update, instead of assuming it is always COLUMN_IDENTIFIER_ROW_ID by @Mytherin in #17039
• Handle Parquet with compressed empty DataPage v2 by @EnricoMi in #17031
• Combine small row groups in Parquet writer by @lnkuiper in #17036
• Merge v1.2.2 into main by @carlopi in #17037
• implement function so I can send a patch to httpfs by @lnkuiper in #17048
• FORCE_ASYNC_SINK_SOURCE: pass also to unittester by @carlopi in #17053
• If a Max Line Size Error happens on all CSV dialect candidates, throw a max line size error. by @pdet in #16935
• Expose BindExtraColumns as a public function by @Mytherin in #17060
• trigger .github/workflows/NightlyBuildsCheck.yml from external repo by @hmeriann in #16949
• Minor parquet crypto clean-up: allow footer key to be passed in directly, and avoid constantly re-reading the key from the config by @Mytherin in #17070
• update julia to v1.2.2 by @Maxxen in #17074
• MultiFileReader Rework (part 18): Replace file path with OpenFileInfo struct by @Mytherin in #17071
• Fix httpfs patches: avoid git log since might contain unsanitised error word by @carlopi in #17075
• Re-enable Avro on core by @Tishj in #17072
• [Nested] Optimize List Type in list_value by @maiadegraaf in #17063
• Grow string dictionary dynamically in Parquet writer by @lnkuiper in #17061
• Add extended file info to OpenFileInfo, and use this to pass encryption keys and footer size to Parquet reader by @Mytherin in #17085
• [Dev] Automatically re-execute when calling __arrow_c_stream__ on an already-consumed-result by @Tishj in #17087
• fsst: Avoid to propagate alignment information in FSST_UNALIGNED_STORE by @carlopi in #17094
• Fix sqlite3 api wrapper link + remove R-CMD-check + add more nightly tests by @carlopi in #17095
• support large dictionary value and constant vector creation in the C API by @joseph-isaacs in #17064
• Add missing lock to UpdateSegment::FetchRow, and cleanup API to require the lock by @Mytherin in #17100
• Valgrind requires tpch by @carlopi in #17101
• Switch to manylinux_2_28 by @hannes in #16956
• Changing mbedtls encryption API by @ccfelius in #16196
• Pull OpenFileExtended through the opener and virtual file system layers by @Mytherin in #17102
• Fix an issue in upserts where the local append state was not correctly flushed by @Mytherin in #17109
• Always parallelize read_json schema detection by @lnkuiper in #17106
• Move transaction cleanup outside of the transaction lock by @taniabogatsch in #17034
• Remove R_CMD_CHECK.yml, now handled by duckdb/duckdb-r repo by @carlopi in #17127
• JSON Bugfixes by @lnkuiper in #17119
• Refactor relassert runs, adding some variations in compiler / statically linked extensions by @carlopi in #17104
• extension-upload-from-nightly.sh: Add --region by @carlopi in #17120
• MultiFileReader: several fixes for virtual column handling and make virtual column handling extensible by @Mytherin in #17123
• Remove misleading lock comment in data table by @taniabogatsch in #17125
• [Dev] Add "registries" to vcpkg.json, add script to list the packages of the registry. by @Tishj in #17124
• External File Cache by @lnkuiper in #16463
• Notify nightly build status by @hmeriann in #17108
• Strict UUID cast by @lnkuiper in #17138
• Copy To File: avoid calling Combine for threads that have not written any rows by @Mytherin in #17142
• Add file_index virtual column to the multi file reader that returns the file index of the read file by @Mytherin in #17144
• MultiFileReader: simplify constant handling, and allow virtual columns returned by the multi file reader to be constant by @Mytherin in #17149
• Changes to encodings to make them more flexible to replacement maps. by @pdet in #17146
• Optimize large Top N queries by @lnkuiper in #17141
• Only trigger TopN rewrite relatively small limits compared to the table size. by @Tmonster in #17140
• platform.hpp: Propagate DUCKDB_EXPLICIT_PLATFORM, avoid early return by @carlopi in #17137
• Keeping the filters which do not remove NULL values by @Damon07 in #17045
• Improve FileSync call on unix platform by @dentiny in #16893
• README: Fix to building link by @szarnyasg in #17161
• [InvokeCI] Add missing pipe to run instruction by @hmeriann in #17163
• Internal #4667: 2025b TimeZone Data by @hawkfish in #17160
• Unify function list by @c-herrewijn in #17168
• [Dev] Generate the EXTENSION_SECRET_TYPES instead of hardcoding them by @Tishj in #17183
• Fix grouping feature with interval type by @handstuyennn in #17181
• Add filename to GZIP stream error by @marcoslot in #17166
• Issue #17115: TimeTZ Approximate Quantile by @hawkfish in #17162
• Issue #17046: AsOf Left Predicates by @hawkfish in #17159
• [Fix] Pass delete indexes when committing updates by @taniabogatsch in #17176
• Python.yml: Add back logic to perform fast-fail on Python 3.10 by @carlopi in #17107
• Notify JDBC repo to run Vendor.yml workflow by @staticlibs in #17099
• Issue #17049: ICU Date Cast by @hawkfish in #17067
• Add bind_operator callback to TableFunction - allowing table functions to directly emit a LogicalOperator by @Mytherin in #17196
• [ENCRYPTION] Make block header size adaptive by @ccfelius in #17118
• Issue #16839: Disable TIMESTAMP Casts by @hawkfish in #16899
• Add support for an explicit PRESERVE_ORDER flag for copy to file by @Mytherin in #17199
• Add SYSTEM_PEAK_BUFFER_MANAGER_MEMORY and SYSTEM_PEAK_TEMP_DIRECTORY_SIZE to profiler by @lnkuiper in #17164
• Fix [InvokeCI / NotifyExternalRepository] Unexpected value 'true' by @hmeriann in #17212
• Add support for the cast_to_type function, that allows generating a cast from an expression to the type of another column by @Mytherin in #17209
• Better cardinality estimates for inequality joins/grouped aggregations by @lnkuiper in #17139
• Add ExternalFileCache validation as option for ExtendedOpenFileInfo by @lnkuiper in #17205
• Explicitly flush the thread-local optimistic writer in PhysicalBatchInsert when finalizing by @Mytherin in #17214
• Pushdown arbitrary expressions into scans by @Mytherin in #17213
• Fix #17170: sort selection result in OR expression by @flashmouse in #17180
• [Dev] Re-enable Iceberg, Bump Avro, fix generate_extension_functions.py for dependencies between extensions by @Tishj in #17204
• Change Invalid Unicode Error to Invalid Encoding by @pdet in #17208
• Direct IO for temp files by @lnkuiper in #17219
• Fix [InvokeCI / NotifyExternalRepository] GitHub Actions has encountered an internal error when running your job. by @hmeriann in #17218
• Add "thousands" option to CSV Reader by @pdet in #17220
• add capi functions to create map and union values by @jraymakers in #17227
• Only notify JDBC when all runs are successful by @staticlibs in #17233
• Update Friendlier SQL link.md by @hfrifkin in #17248
• Implement reading concatenated GZIP members by @lnkuiper in #17255
• Return invalid BufferHandle upon loading a destroyed BlockHandle by @lnkuiper in #17249
• Internal #4772: Timestamp Error Parameter by @hawkfish in #17283
• BUGFIX: do not perform unused columns optimization in presence of multiple grouping sets by @Tmonster in #17259
• Internal #4532: 13 Month Intervals by @hawkfish in #17303
• Dont try to load extension if storage type is already registered by @Maxxen in #17241
• Adapt size of hash table during aggregation using HyperLogLog by @lnkuiper in #17236
• Switch to always using list identifier instead of array by @J-Meyers in #17242
• Add root's query_location also to TransformInterval by @carlopi in #17271
• Histogram table function test by @hmeriann in #17276
• Guess Parquet footer size by @lnkuiper in #17300
• Issue #16563: FLOAT to DECIMAL by @hawkfish in #17302
• Feature #15873: Windowed ORDER BYs by @hawkfish in #17304
• Switch from Bottom-Up to Top-Down Decorrelation Strategy by @kryonix in #17294
• Generating random data for mbedtls without key by @ccfelius in #17309
• Fix CI by @Mytherin in #17319
• [Arrow] Implement support to consuming and producing Decimal 32 and 64. by @pdet in #17314
• take the column ids from the logical get, don't require a LogicalGet … by @Tishj in #17315
• Allow installing extensions with external access allowlist by @samansmink in #17316
• Implement ARTMerger replacing the recursive ART merge algorithm by @taniabogatsch in #17243
• Share null mask with constant null arg vector by @iceTTTT in #17234
• Fix #17311: correctly check for presence of recursive keys in transformer by @Mytherin in #17320
• [CSV Reader] Simplify Quote/Escape detection code, make it more robust and decouple comment and skip_rows option. by @pdet in #17284
• Fix try_cast from NaN double to decimal by @lnkuiper in #17322
• Add serialization for new TableColumn type by @Mytherin in #17321
• Extract expressions from nested conjunction AND for index scan by @lnkuiper in #17297
• Support late materialization in the Parquet reader, and handle COUNT(*) directly in the multi file reader by @Mytherin in #17325
• Implement ARTOperator replacing Lookup and the recursive Insert by @taniabogatsch in #17327
• Internal #4723: Inequality Condition Pushdown by @hawkfish in #17317
• Properly format strings when throw JSON errors by @lnkuiper in #17331
• Fix potential vulnerable cloned function by @npt-1707 in #17340
• Fix potential vulnerable cloned function by @npt-1707 in #17339
• Revert "Skip MinGW, currently failing on main" by @carlopi in #17342
• Unify Parquet Metadata cache invalidation logic with Cached File System cache invalidation by @Mytherin in #17334
• Fix issue with empty ranges by @kryonix in #17332
• Internal #4797: Timestamp Range Cardinality by @hawkfish in #17330
• Some nitpicking fixes by @szarnyasg in #17337
• Issue #17299: Integer Rounding by @hawkfish in #17328
• Parquet Reader: emit partition stats for any files that have cached metadata, and implement ListFilesExtended that adds extra info to files globbed by @Mytherin in #17344
• Add support for UUID v7 to Filename Pattern - and clean it up so that it correctly supports composite patterns by @Mytherin in #17345
• Add support for the HIVE_FILE_PATTERN option - that allows partitioned files to be written without writing them to a hive-style directory structure by @Mytherin in #17346
• Add an OnDetach callback to the catalog that is triggered when the user detaches a catalog by @Mytherin in #17347
• Pass commit ID to NotifyExternalRepositories.yml by @staticlibs in #17333
• Add support for BENCHMARK_ROOT_DIRECTORY cmake option to change benchmark runner root directory, and add support for cache_file and reload options to enable better caching for non-DuckDB databases by @Mytherin in #17355
• Support --directories option in format.py by @Mytherin in #17354
• Handle both ENCRYPTION_KEY and STORAGE_VERSION passed as options by @carlopi in #17357
• Fix internal exception from assigning invalid index to optional_idx query_id; by @Tishj in #17359
• Fixup amalgamation: reqlen is only used with assert enabled by @carlopi in #17361
• md5_number: return UHUGEINT by @szarnyasg in #17336
• Skip emitting partition stats if "has_deletes" is set in the file info by @Mytherin in #17365
• Benchmark runner: add argument, include and load_only options - and make ClickBench run the original benchmark instead of a subset by @Mytherin in #17367
• Fix two off-by-one errors in row estimate of range and generate_series by @JelteF in #17373
• [Nested] Fix: 16489 - Find NULLs in lists using list_position by @maiadegraaf in #17080
• fix #17258: Allow to open database in readonly mode within cli by @jjballano in #17375
• Join Hash Table Probing Optimization: Optional Probing Selection Vector by @gropaul in #17062
• Remove bundled TPCH & TPCDS in Python wheels by @carlopi in #15923
• [Compression] Introduce DICT_FSST compression method by @Tishj in #15637
• Deprecate lambda arrow (->) and replace it with LAMBDA x : x + 1 by @taniabogatsch in #17235
• fix not setting nested validity when map_extract returns null by @Maxxen in #17379
• Function chaining: report missing column instead of missing function if function exists by @Mytherin in #17383
• Improve error messages in UPDATE ... SET by @Mytherin in #17384
• Add candidates suggestion when COLUMNS regex does not match any columns by @Mytherin in #17385
• add step to clean up the disc space to fix No space left on device by @hmeriann in #17390
• Fix issue in string -> hugeint conversion with decimals and exponents by @Mytherin in #17388
• Improve error message reporting for cast failures by @Mytherin in #17382
• Fix Python CI: pin virtualenv to previous version by @Mytherin in #17386
• Improve error reporting for missing qualified columns by @Mytherin in #17397
• Issue #17266: Lead Lag Nulls by @hawkfish in #17391
• Fix #17266：the result of lad/lead when the offset is null by @ditdb in https://github.com/duckdb/duckdb/pull/17268
• VirtualFileSystem to take an input, allowing to customize behaviour by @carlopi in https://github.com/duckdb/duckdb/pull/17393
• [Dev] Add QualifiedName::ParseComponents, add input to the error messages by @Tishj in https://github.com/duckdb/duckdb/pull/17403
• Provide suggestions and a link to the documentation for OOM errors by @Mytherin in https://github.com/duckdb/duckdb/pull/17402
• [Dev] Flatten any deeper children vectors, when the top level is a FLAT vector by @Tishj in https://github.com/duckdb/duckdb/pull/17387
• Minor fixes for the CLI by @Mytherin in https://github.com/duckdb/duckdb/pull/17405
• Add support for CREATE OR REPLACE TYPE, CREATE TYPE IF NOT EXISTS and CREATE TEMPORARY TYPE by @Mytherin in https://github.com/duckdb/duckdb/pull/17404
• Use an insertion order preserving map in Value::MAP by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17389
• Implement json_each/json_tree by @lnkuiper in https://github.com/duckdb/duckdb/pull/17406
• Fix #16552: adjust join condition sequence by @flashmouse in https://github.com/duckdb/duckdb/pull/16943
• WAL replay index fixes by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17409
• ZSTD: use a high penalty when min size is exceeded instead of disabling compression to allow force compression to work by @Mytherin in https://github.com/duckdb/duckdb/pull/17412
• Internal #4723: PWMJ Inequality Pushdown by @hawkfish in https://github.com/duckdb/duckdb/pull/17400
• Move all httplib code to HTTPUtil class by @Mytherin in https://github.com/duckdb/duckdb/pull/17420
• Avoid generating default views and macros in the temporary catalog by @Mytherin in https://github.com/duckdb/duckdb/pull/17408
• unittest: improve detection of whether or not we can run --force-restart tests by @Mytherin in https://github.com/duckdb/duckdb/pull/17419
• Give tasks a TaskType with a name by @Mytherin in https://github.com/duckdb/duckdb/pull/17421
• Use argparse in scripts/format.py by @adsharma in https://github.com/duckdb/duckdb/pull/17360
• Add missing commas by @szarnyasg in https://github.com/duckdb/duckdb/pull/17424
• Internal #4830: IEJoin Inequality Pushdown by @hawkfish in https://github.com/duckdb/duckdb/pull/17422
• Add conn.query_progress() method by @nickzoic in https://github.com/duckdb/duckdb/pull/16927
• Fixes filter pruning use the statistics updated by the same filter by @Damon07 in https://github.com/duckdb/duckdb/pull/17425
• Fix JSON extension compilation on Ubuntu 22.04 by @staticlibs in https://github.com/duckdb/duckdb/pull/17434
• Use pytest in SQLLogic Python test runner by @Flogex in https://github.com/duckdb/duckdb/pull/16685
• On COPY TO/FROM check the format during binding. by @pdet in https://github.com/duckdb/duckdb/pull/17381
• BUGFIX: DELIM_JOINS should reflect functionality of NULL filtering conditions in joins with DELIM_GETS by @Tmonster in https://github.com/duckdb/duckdb/pull/16910
• Allow directly attaching of Parquet/CSV/JSON files by @Mytherin in https://github.com/duckdb/duckdb/pull/17415
• Force errors when trying lines as early as possible by @pdet in https://github.com/duckdb/duckdb/pull/17427
• Enable SYSTEM_PEAK_BUFFER_MEMORY and SYSTEM_PEAK_TEMP_DIR_SIZE profiling by default by @lnkuiper in https://github.com/duckdb/duckdb/pull/17407
• [C API] Expose the client context, connection id and scalar function bind data by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17449
• [CSV Sniffer] Proper type replacement in header only files by @pdet in https://github.com/duckdb/duckdb/pull/17447
• Recurse into MAP and LIST with the remap_struct and the MFR ColumnMapper by @Tishj in https://github.com/duckdb/duckdb/pull/17448
• Fix: pyproject.toml does not contain a tool.setuptools_scm section by @YUKI2eN3e in https://github.com/duckdb/duckdb/pull/17443
• [Fix] Macro binding with unknown parameters in list_has_all and some other code tidying by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17450
• Generalize HTTP interface and use the new HTTP interface in httpfs by @Mytherin in https://github.com/duckdb/duckdb/pull/17464
• [Fix] Switch between constant and flat vector in C API by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17465
• Fix TIMETZ cast in example by @szarnyasg in https://github.com/duckdb/duckdb/pull/17468
• Remove duplicated arrow fetch test by @emmanuel-ferdman in https://github.com/duckdb/duckdb/pull/17476
• Multi File Reader Rework (Part 19): Make MultiFileReaderInterface virtual, and move reading methods to the BaseFileReader by @Mytherin in https://github.com/duckdb/duckdb/pull/17475
• [Serializer] Lambda Compatibilty Fix by @maiadegraaf in https://github.com/duckdb/duckdb/pull/17428
• fix parsing bool values in JSON by @ccfelius in https://github.com/duckdb/duckdb/pull/17460
• Emit dictionary vectors with unaligned start index by @OmidAfroozeh in https://github.com/duckdb/duckdb/pull/17471
• Add release version by @hannes in https://github.com/duckdb/duckdb/pull/17479
• Expose qualified table names in GetTableNames and add duckdb_get_table_names to C API by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17472
• Bump avro, httpfs, mysql, postgres and sqlite by @Mytherin in https://github.com/duckdb/duckdb/pull/17482
• Fix GeoParquet ExpressionColumnReader schema by @Maxxen in https://github.com/duckdb/duckdb/pull/17481
• add regression_threshold_seconds argument to regression/test_runner.py by @hmeriann in https://github.com/duckdb/duckdb/pull/17485
• DROP of missing entry should fail in binding by @jeewonhh in https://github.com/duckdb/duckdb/pull/17474
• HTTPFS Parameters fix by @Mytherin in https://github.com/duckdb/duckdb/pull/17486
• HTTPUtil Fix: correctly pass in on_retry by @Mytherin in https://github.com/duckdb/duckdb/pull/17494
• Bump spatial & vss by @Maxxen in https://github.com/duckdb/duckdb/pull/17492
• Add support for altering structs (drop, add, rename field) inside LIST and MAP columns. by @Tishj in https://github.com/duckdb/duckdb/pull/17462
• [Python Dev] Guard against python exceptions when interacting with the currentframe object by @Tishj in https://github.com/duckdb/duckdb/pull/17490
• If distinct count from stats is 0, do not use it in Join Order Optimizer by @Tmonster in https://github.com/duckdb/duckdb/pull/17466
• Make the encodings extension a core extension, and make it auto-loadable. by @pdet in https://github.com/duckdb/duckdb/pull/17206
• Allow passing down rc-style version also via OVERRIDE_GIT_DESCRIBE by @carlopi in https://github.com/duckdb/duckdb/pull/17501
• Allow DUCKDB_EXPLICIT_VERSION to be propagated by @carlopi in https://github.com/duckdb/duckdb/pull/17498
• Minor nightly fixes by @Mytherin in https://github.com/duckdb/duckdb/pull/17500
• Add FileSystem::TryRemoveFile - that only removes a file if it exists by @Mytherin in https://github.com/duckdb/duckdb/pull/17502
• Add OperatorFinalize callback to operators - which is called after a pipeline is finished by @Mytherin in https://github.com/duckdb/duckdb/pull/17503
• Apply dynamic filter pushdown of TopN optimizer also to existing TopN nodes by @Mytherin in https://github.com/duckdb/duckdb/pull/17504
• Fix: Optional Probe Selection by @gropaul in https://github.com/duckdb/duckdb/pull/17505
• FileHandle Logging by @samansmink in https://github.com/duckdb/duckdb/pull/16758
• Fix typos by @szarnyasg in https://github.com/duckdb/duckdb/pull/17478
• Remove spatial from OSX Relassert by @carlopi in https://github.com/duckdb/duckdb/pull/17509
• Update more extensions by @Maxxen in https://github.com/duckdb/duckdb/pull/17510
• Bump HTTPFS again by @Mytherin in https://github.com/duckdb/duckdb/pull/17511
• feat: include catalog and schema names in function serialization by @rustyconover in https://github.com/duckdb/duckdb/pull/17512
• Fix encodings by @carlopi in https://github.com/duckdb/duckdb/pull/17514
• Fix python nightly build by @Tishj in https://github.com/duckdb/duckdb/pull/17515
• Use Catalog::TryAutoLoad for encodings extension by @pdet in https://github.com/duckdb/duckdb/pull/17520
• [Python Dev] Using reinterpret_steal breaks the refcount of the passed-in object by @Tishj in https://github.com/duckdb/duckdb/pull/17525
• Fix update extensions by @carlopi in https://github.com/duckdb/duckdb/pull/17527
• Minor fixes to exception error messages by @carlopi in https://github.com/duckdb/duckdb/pull/17528
• [Python Dev] Fix failing tests for the Python SQLLogicTester by @Tishj in https://github.com/duckdb/duckdb/pull/17529
• Resolve GitHub workflow set-output deprecation warnings by @kurtmckee in https://github.com/duckdb/duckdb/pull/17516
• [CSV Reader] Detect SQLNULL types for schema merging, use schema merging in csv relations, add files_to_sniff option. by @pdet in https://github.com/duckdb/duckdb/pull/17467
• Fix extension test by @carlopi in https://github.com/duckdb/duckdb/pull/17536
• [Dev] Fix crash when describing a table with a virtual column by @Tishj in https://github.com/duckdb/duckdb/pull/17544
• [HTTPUtil] Let requests made through the HTTPUtil interface accept URI's without a scheme. by @Tishj in https://github.com/duckdb/duckdb/pull/17545
• Attach after setting database type by @Mytherin in https://github.com/duckdb/duckdb/pull/17546
• Pass MultiFileGlobalState to InitializeReader, and pass file list to CreateMapping instead of eagerly getting the first file by @Mytherin in https://github.com/duckdb/duckdb/pull/17553
• [Dev] Fix allowed_directories crash by @Tishj in https://github.com/duckdb/duckdb/pull/17548
• [Fix] duplicate filters during index scans by @taniabogatsch in https://github.com/duckdb/duckdb/pull/17547
• Generate data for tpch sf100 in steps by @Tmonster in https://github.com/duckdb/duckdb/pull/17539
• Issue #17537: Fractional Second Padding by @hawkfish in https://github.com/duckdb/duckdb/pull/17556
• Make MultiFileList::Copy a virtual method by @Mytherin in https://github.com/duckdb/duckdb/pull/17566
• [Dev] Can't use USING COMPRESSION with a deprecated compression type by @Tishj in https://github.com/duckdb/duckdb/pull/17542
• Add (de)serialization for ExtraOperatorInfo by @NiclasHaderer in https://github.com/duckdb/duckdb/pull/17563
• Fix issue with ExternalFileCache when data is evicted by @lnkuiper in https://github.com/duckdb/duckdb/pull/17567
• Remote Reads: allocate correct buffer size for prefetch by @Mytherin in https://github.com/duckdb/duckdb/pull/17557
• Remove patch and bump httpfs by @carlopi in https://github.com/duckdb/duckdb/pull/17558
• [Dev] Fix Arrow fixed size binary reading by @Tishj in https://github.com/duckdb/duckdb/pull/17573
• Fix setup.py to correctly handle OVERRIDE_GIT_DESCRIBE by @carlopi in https://github.com/duckdb/duckdb/pull/17580

Full Changelog: v1.2.2...v1.3.0