polars

⚠️ Deprecations

• Deprecate use_pyarrow parameter for to_numpy methods (#16391)

✨ Enhancements

• Add field expression as selector with an struct scope (#16402)
• Field expansion renaming (#16397)
• Respect index order in DataFrame.to_numpy also for non-numeric frames (#16390)
• add Expr.interpolate_by (#16313)
• Implement Struct support for Series.to_numpy (#16383)

🐞 Bug fixes

• Fix struct arithmetic schema (#16396)
• Handle non-Sequence iterables in filter (#16254)
• Fix don't panic on chunked to_numpy conversion (#16393)
• Don't check nulls before conversion (#16392)
• Add support for generalized ufunc with different size input and output (#16336)
• Improve cursor close behaviour with respect to Oracle "thick mode" connections (#16380)
• Fix DataFrame.to_numpy for Array/Struct types (#16386)
• Handle ambiguous/nonexistent datetimes in Series construction (#16342)
• Fix DataFrame.to_numpy for Struct columns when structured=True (#16358)
• Use strings to expose ClosedInterval in expr IR (#16369)

📖 Documentation

• Expand docstrings for to_numpy methods (#16394)
• Add a not about index access on struct.field (#16389)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @alexander-beedie, @coastalwhite, @dangotbanned, @itamarst, @ritchie46, @stinodego and @wence-

💥 Breaking changes

• Remove incremental read based batched CSV reader (#16259)
• separate rolling_*_by from rolling_*(..., by=...) in Rust (#16102)
• Move CSV read options from CsvReader to CsvReadOptions (#16126)
• Rename all 'Chunk's to RecordBatch (#16063)
• prepare for join coalescing argument (#15418)
• Rename to CsvParserOptions to CsvReaderOptions, use in CsvReader (#15919)
• Add context trace to LazyFrame conversion errors (#15761)
• Move schema resolving of file scan to IR phase (#15739)
• Move schema resolving to IR phase. (#15714)
• Rename LogicalPlan and builders to reflect their uses better (#15712)

🚀 Performance improvements

• Use branchless uleb128 decoding for parquet (#16352)
• Reduce error bubbling in parquet hybrid_rle (#16348)
• use is_sorted in ewm_mean_by, deprecate check_sorted (#16335)
• Optimize is_sorted for numeric data (#16333)
• do not use pyo3-built (#16309)
• Faster bitpacking for Parquet writer (#16278)
• Avoid importing ctypes.util in CPU check script if possible (#16307)
• Don't rechunk when converting DataFrame to numpy/ndarray (#16288)
• use zeroed vec in ewm_mean_by for sorted fastpath (#16265)
• use zeroable_vec in ewm_mean_by (#16166)
• Improve cost of chunk_idx compute (#16154)
• Don't rechunk by default in concat (#16128)
• Ensure rechunk is parallel (#16127)
• Don't traverse deep datasets that we repr as union in CSE (#16096)
• Ensure better chunk sizes (#16071)
• Don't rechunk in parallel collection (#15907)
• Improve non-trivial list aggregations (#15888)
• Ensure we hit specialized gather for binary/strings (#15886)
• Limit the cache size for to_datetime (#15826)
• skip initial null items and don't recompute slope in interpolate (#15819)
• Fix quadratic in binview growable same source (#15734)

✨ Enhancements

• Raise when joining on the same keys twice (#16329)
• Don't require data to be sorted by by column in rolling_*_by operations (#16249)
• Add struct.field expansion (regex, wildcard, columns) (#16320)
• Faster bitpacking for Parquet writer (#16278)
• Add struct.with_fields (#16305)
• Handle implicit SQL string → temporal conversion in the BETWEEN clause (#16279)
• Add new index/range based selector cs.by_index, allow multiple indices for nth (#16217)
• Show warning if expressions are very deep (#16233)
• Native CSV file list reading (#16180)
• Register memory mapped files and raise when written to (#16208)
• Raise when encountering invalid supertype in functions during conversion (#16182)
• Add SQL support for GROUP BY ALL syntax and fix several issues with aliased group keys (#16179)
• Allow implicit string → temporal conversion in SQL comparisons (#15958)
• separate rolling_*_by from rolling_*(..., by=...) in Rust (#16102)
• Add run-length encoding to Parquet writer (#16125)
• add date pattern dd.mm.YYYY (#16045)
• Add RLE to RLE_DICTIONARY encoder (#15959)
• Support non-coalescing joins in default engine (#16036)
• Move diagonal & horizontal concat schema resolving to IR phase (#16034)
• raise more informative error messages in rolling_* aggregations instead of panicking (#15979)
• Convert concat during IR conversion (#16016)
• Improve dynamic supertypes (#16009)
• Additional uint datatype support for the SQL interface (#15993)
• Support Decimal read from IPC (#15965)
• Add typed collection from par iterators (#15961)
• Add by argument for Expr.top_k and Expr.bottom_k (#15468)
• Add option to disable globbing in csv (#15930)
• Add option to disable globbing in parquet (#15928)
• Rename to CsvParserOptions to CsvReaderOptions, use in CsvReader (#15919)
• Expressify dt.round (#15861)
• Improve error messages in context stack (#15881)
• Add dynamic literals to ensure schema correctness (#15832)
• dt.truncate supports broadcasting lhs (#15768)
• Expressify str.json_path_match (#15764)
• Support decimal float parsing in CSV (#15774)
• Add context trace to LazyFrame conversion errors (#15761)

🐞 Bug fixes

• correct AExpr.to_field for bitwise and logical and/or (#16360)
• cargo clippy for uleb128 safety comment (#16368)
• Infer CSV schema as supertype of all files (#16349)
• Address overflow combining u64 hashes in Debug builds (#16323)
• Don't exclude explicitly named columns in group-by context' expr expansion (#16318)
• Harden Series.reshape against invalid parameters (#16281)
• Fix list.sum dtype for boolean (#16290)
• Don't stackoverflow on all/any horizontal (#16287)
• compilation error when both lazy and ipc features are enabled (#16284)
• `rolling_*_by was throwing incorrect error when dataframe was sorted by contained multiple chunks (#16247)
• Clippy Error for CPUID (#16241)
• Reading CSV with low_memory gave no data (#16231)
• Empty unique (#16214)
• Fix empty drop nulls (#16213)
• Fix get expression group-by state (#16189)
• Fix rolling empty group OOB (#16186)
• offset=-0i was being treated differently to offset=0i in rolling (#16184)
• Fix panic on empty frame joins (#16181)
• Fix streaming glob slice (#16174)
• Fix CSV skip_rows_after_header for streaming (#16176)
• Flush parquet at end of batches tick (#16073)
• Check CSE name aliases for collisions. (#16149)
• Don't override CSV reader encoding with lossy UTF-8 (#16151)
• Add missing allow macros for windows (#16130)
• Ensure hex and bitstring literals work inside SQL IN clauses (#16101)
• Revert "Add RLE to RLE_DICTIONARY encoder" (#16113)
• Respect user passed 'reader_schema' in 'scan_csv' (#16080)
• Lazy csv + projection; respect null values arg (#16077)
• Materialize dtypes when converting to arrow (#16074)
• Fix casting decimal to decimal for high precision (#16049)
• Fix printing max scale decimals (#16048)
• Decimal supertype for dyn int (#16046)
• Do not set sorted flag on lexical sorting (#16032)
• properly handle nulls in DictionaryArray::iter_typed (#16013)
• Fix CSE case where upper plan has no projection (#16011)
• Crash/incorrect group_by/n_unique on categoricals created by (q)cut (#16006)
• Ternary supertype dynamics (#15995)
• Treat splitting by empty string as iterating over chars (#15922)
• Fix PartialEq for DataType::Unknown (#15992)
• Do not reverse null indices in descending arg_sort (#15974)
• Finish adding typed_lit to help schema determination in SQL "extract" func (#15955)
• do not panic when comparing against categorical with incompatible dtype (#15857)
• Join validation for multiple keys (#15947)
• Set default limit for String column display to 30 and fix edge cases (#15934)
• typo in add_half_life takes ln(negative) (#15932)
• Remove ffspec from parquet reader (#15927)
• avoid WRITE+EXEC for CPUID check (#15912)
• fix inconsistent decimal formatting (#15457)
• Preserve NULLs for is_not_nan (#15889)
• double projection check should only take the upstream projections into account (#15901)
• Ensure we don't create invalid frames when combining unit lit + … (#15903)
• Clear cached rename schema (#15902)
• Fix OOB in struct lit/agg aggregation (#15891)
• create (q)cut labels in fixed order (#15843)
• Tag shrink_dtype as non-streaming (#15828)
• drop-nulls edge case; remove drop-nulls special case (#15815)
• ewm_mean_by was skipping initial nulls when it was already sorted by "by" column (#15812)
• Consult cgroups to determine free memory (#15798)
• raise if index count like 2i is used when performing rolling, group_by_dynamic, upsample, or other temporal operatios (#15751)
• Don't deduplicate sort that has slice pushdown (#15784)
• Fix incorrect is_between pushdown to scan_pyarrow_dataset (#15769)
• Handle null index correctly for list take (#15737)
• Preserve lexical ordering on concat (#15753)
• Remove incorrect unsafe pointer cast for int -> enum (#15740)
• pass series name to apply for cut/qcut (#15715)
• count of null column shouldn't panic in agg context (#15710)

📖 Documentation

• Clarify arrow usage (#16152)
• Solve inconsistency between code and comment (#16135)
• add filter docstring examples to date and datetime (#15996)
• update the link to R API docs (#15973)
• Fix a typo in categorical section of the user guide (#15777)
• Fix incorrect column name in LazyFrame.sort doc example (#15658)

📦 Build system

• Update Rust nightly toolchain version (#16222)
• Don't import jemalloc (#15942)
• Use default allocator for lts-cpu (#15941)
• replace all macos-latest referrals with macos-13 (#15926)
• pin mimalloc and macos-13 (#15925)
• use jemalloc in lts-cpu (#15913)

🛠️ Other improvements

• simplify interpolate code, add test for rolling_*_by with nulls (#16334)
• Move expression expansion to conversion module (#16331)
• Add polars-expr README (#16316)
• Move physical expressions to new crate (#16306)
• Use cls (not self) in classmethods (#16303)
• conditionally print the CSEs (#16292)
• Rename ChunkedArray.chunk_id to chunk_lengths (#16273)
• Use Scalar instead of Series some aggregations (#16277)
• Use CsvReadOptions in LazyCsvReader (#16283)
• Do not hardcode bash path in Makefile (#16263)
• Add IR::Reduce (not yet implemented) (#16216)
• Remove incremental read based batched CSV reader (#16259)
• move all describe, describe_tree and dot-viz code to IR instead of DslPlan (#16237)
• move describe to IR instead of DSL (#16191)
• Use Duration.is_zero instead of comparing Duration.duration_ns to 0 (#16195)
• Remove unused code (#16175)
• Don't override CSV reader encoding with lossy UTF-8 (#16151)
• Move CSV read options from CsvReader to CsvReadOptions (#16126)
• Bump sccache action (#16088)
• Fix failures in test coverage workflow (#16083)
• Rename all 'Chunk's to RecordBatch (#16063)
• Use UnionArgs for DSL side (#16017)
• Add some comments (#16008)
• prepare for join coalescing argument (#15418)
• Pin coverage job to MacOS 13 for now (#15918)
• Reorganize from_iter and dispatch to collect_ca when possible (#15904)
• More polars-io cleanup (#15885)
• Improve type-coercion (#15879)
• Move type coercion to IR conversion phase (#15868)
• Reorganize polars_io::parquet module (#15860)
• Reorganize polars_io::csv module (#15831)
• Always expand horizontal_any/all (#15816)
• Rename decimal_float to decimal_comma (#15817)
• Move IO-related options structs to polars-io (#15806)
• Split coverage calculation (#15780)
• Update readme (#15787)
• Move schema resolving of file scan to IR phase (#15739)
• Factor out ensure_is_constant_duration (#15733)
• Move schema resolving to IR phase. (#15714)
• Rename LogicalPlan and builders to reflect their uses better (#15712)

Thank you to all our contributors for making this release possible!
@CanglongCl, @JulianCologne, @KDruzhkin, @MarcoGorelli, @NedJWestern, @NexVeridian, @NickCondron, @Robinsane, @ShivMunagala, @TobiasDummschat, @YichiZhang0613, @alexander-beedie, @avimallu, @bertiewooster, @brandon-b-miller, @c-peters, @coastalwhite, @dangotbanned, @datenzauberai, @deanm0000, @dependabot, @dependabot[bot], @eitsupi, @gasmith, @haocheng6, @ion-elgreco, @itamarst, @janpipek, @jr200, @jrycw, @jsarbach, @luke396, @marenwestermann, @max-muoto, @mbuhidar, @nameexhaustion, @orlp, @pydanny, @r-brink, @reswqa, @ritchie46, @stinodego, @thalassemia, @tharunsuresh-code, @twoertwein, @wence- and @wsyxbcl

💥 Breaking changes

• Remove incremental read based batched CSV reader (#16259)
• separate rolling_*_by from rolling_*(..., by=...) in Rust (#16102)
• Move CSV read options from CsvReader to CsvReadOptions (#16126)
• Rename all 'Chunk's to RecordBatch (#16063)
• prepare for join coalescing argument (#15418)
• Rename to CsvParserOptions to CsvReaderOptions, use in CsvReader (#15919)
• Add context trace to LazyFrame conversion errors (#15761)
• Move schema resolving of file scan to IR phase (#15739)
• Move schema resolving to IR phase. (#15714)
• Rename LogicalPlan and builders to reflect their uses better (#15712)

🚀 Performance improvements

• Use branchless uleb128 decoding for parquet (#16352)
• Reduce error bubbling in parquet hybrid_rle (#16348)
• use is_sorted in ewm_mean_by, deprecate check_sorted (#16335)
• Optimize is_sorted for numeric data (#16333)
• do not use pyo3-built (#16309)
• Faster bitpacking for Parquet writer (#16278)
• Avoid importing ctypes.util in CPU check script if possible (#16307)
• Don't rechunk when converting DataFrame to numpy/ndarray (#16288)
• use zeroed vec in ewm_mean_by for sorted fastpath (#16265)
• use zeroable_vec in ewm_mean_by (#16166)
• Improve cost of chunk_idx compute (#16154)
• Don't rechunk by default in concat (#16128)
• Ensure rechunk is parallel (#16127)
• Don't traverse deep datasets that we repr as union in CSE (#16096)
• Ensure better chunk sizes (#16071)
• Don't rechunk in parallel collection (#15907)
• Improve non-trivial list aggregations (#15888)
• Ensure we hit specialized gather for binary/strings (#15886)
• Limit the cache size for to_datetime (#15826)
• skip initial null items and don't recompute slope in interpolate (#15819)
• Fix quadratic in binview growable same source (#15734)

✨ Enhancements

• Raise when joining on the same keys twice (#16329)
• Don't require data to be sorted by by column in rolling_*_by operations (#16249)
• Add struct.field expansion (regex, wildcard, columns) (#16320)
• Faster bitpacking for Parquet writer (#16278)
• Add struct.with_fields (#16305)
• Handle implicit SQL string → temporal conversion in the BETWEEN clause (#16279)
• Add new index/range based selector cs.by_index, allow multiple indices for nth (#16217)
• Show warning if expressions are very deep (#16233)
• Native CSV file list reading (#16180)
• Register memory mapped files and raise when written to (#16208)
• Raise when encountering invalid supertype in functions during conversion (#16182)
• Add SQL support for GROUP BY ALL syntax and fix several issues with aliased group keys (#16179)
• Allow implicit string → temporal conversion in SQL comparisons (#15958)
• separate rolling_*_by from rolling_*(..., by=...) in Rust (#16102)
• Add run-length encoding to Parquet writer (#16125)
• add date pattern dd.mm.YYYY (#16045)
• Add RLE to RLE_DICTIONARY encoder (#15959)
• Support non-coalescing joins in default engine (#16036)
• Move diagonal & horizontal concat schema resolving to IR phase (#16034)
• raise more informative error messages in rolling_* aggregations instead of panicking (#15979)
• Convert concat during IR conversion (#16016)
• Improve dynamic supertypes (#16009)
• Additional uint datatype support for the SQL interface (#15993)
• Support Decimal read from IPC (#15965)
• Add typed collection from par iterators (#15961)
• Add by argument for Expr.top_k and Expr.bottom_k (#15468)
• Add option to disable globbing in csv (#15930)
• Add option to disable globbing in parquet (#15928)
• Rename to CsvParserOptions to CsvReaderOptions, use in CsvReader (#15919)
• Expressify dt.round (#15861)
• Improve error messages in context stack (#15881)
• Add dynamic literals to ensure schema correctness (#15832)
• dt.truncate supports broadcasting lhs (#15768)
• Expressify str.json_path_match (#15764)
• Support decimal float parsing in CSV (#15774)
• Add context trace to LazyFrame conversion errors (#15761)

🐞 Bug fixes

• correct AExpr.to_field for bitwise and logical and/or (#16360)
• cargo clippy for uleb128 safety comment (#16368)
• Infer CSV schema as supertype of all files (#16349)
• Address overflow combining u64 hashes in Debug builds (#16323)
• Don't exclude explicitly named columns in group-by context' expr expansion (#16318)
• Harden Series.reshape against invalid parameters (#16281)
• Fix list.sum dtype for boolean (#16290)
• Don't stackoverflow on all/any horizontal (#16287)
• compilation error when both lazy and ipc features are enabled (#16284)
• `rolling_*_by was throwing incorrect error when dataframe was sorted by contained multiple chunks (#16247)
• Clippy Error for CPUID (#16241)
• Reading CSV with low_memory gave no data (#16231)
• Empty unique (#16214)
• Fix empty drop nulls (#16213)
• Fix get expression group-by state (#16189)
• Fix rolling empty group OOB (#16186)
• offset=-0i was being treated differently to offset=0i in rolling (#16184)
• Fix panic on empty frame joins (#16181)
• Fix streaming glob slice (#16174)
• Fix CSV skip_rows_after_header for streaming (#16176)
• Flush parquet at end of batches tick (#16073)
• Check CSE name aliases for collisions. (#16149)
• Don't override CSV reader encoding with lossy UTF-8 (#16151)
• Add missing allow macros for windows (#16130)
• Ensure hex and bitstring literals work inside SQL IN clauses (#16101)
• Revert "Add RLE to RLE_DICTIONARY encoder" (#16113)
• Respect user passed 'reader_schema' in 'scan_csv' (#16080)
• Lazy csv + projection; respect null values arg (#16077)
• Materialize dtypes when converting to arrow (#16074)
• Fix casting decimal to decimal for high precision (#16049)
• Fix printing max scale decimals (#16048)
• Decimal supertype for dyn int (#16046)
• Do not set sorted flag on lexical sorting (#16032)
• properly handle nulls in DictionaryArray::iter_typed (#16013)
• Fix CSE case where upper plan has no projection (#16011)
• Crash/incorrect group_by/n_unique on categoricals created by (q)cut (#16006)
• Ternary supertype dynamics (#15995)
• Treat splitting by empty string as iterating over chars (#15922)
• Fix PartialEq for DataType::Unknown (#15992)
• Do not reverse null indices in descending arg_sort (#15974)
• Finish adding typed_lit to help schema determination in SQL "extract" func (#15955)
• do not panic when comparing against categorical with incompatible dtype (#15857)
• Join validation for multiple keys (#15947)
• Set default limit for String column display to 30 and fix edge cases (#15934)
• typo in add_half_life takes ln(negative) (#15932)
• Remove ffspec from parquet reader (#15927)
• avoid WRITE+EXEC for CPUID check (#15912)
• fix inconsistent decimal formatting (#15457)
• Preserve NULLs for is_not_nan (#15889)
• double projection check should only take the upstream projections into account (#15901)
• Ensure we don't create invalid frames when combining unit lit + … (#15903)
• Clear cached rename schema (#15902)
• Fix OOB in struct lit/agg aggregation (#15891)
• create (q)cut labels in fixed order (#15843)
• Tag shrink_dtype as non-streaming (#15828)
• drop-nulls edge case; remove drop-nulls special case (#15815)
• ewm_mean_by was skipping initial nulls when it was already sorted by "by" column (#15812)
• Consult cgroups to determine free memory (#15798)
• raise if index count like 2i is used when performing rolling, group_by_dynamic, upsample, or other temporal operatios (#15751)
• Don't deduplicate sort that has slice pushdown (#15784)
• Fix incorrect is_between pushdown to scan_pyarrow_dataset (#15769)
• Handle null index correctly for list take (#15737)
• Preserve lexical ordering on concat (#15753)
• Remove incorrect unsafe pointer cast for int -> enum (#15740)
• pass series name to apply for cut/qcut (#15715)
• count of null column shouldn't panic in agg context (#15710)

📖 Documentation

• Clarify arrow usage (#16152)
• Solve inconsistency between code and comment (#16135)
• add filter docstring examples to date and datetime (#15996)
• update the link to R API docs (#15973)
• Fix a typo in categorical section of the user guide (#15777)
• Fix incorrect column name in LazyFrame.sort doc example (#15658)

📦 Build system

• Update Rust nightly toolchain version (#16222)
• Don't import jemalloc (#15942)
• Use default allocator for lts-cpu (#15941)
• replace all macos-latest referrals with macos-13 (#15926)
• pin mimalloc and macos-13 (#15925)
• use jemalloc in lts-cpu (#15913)

🛠️ Other improvements

• simplify interpolate code, add test for rolling_*_by with nulls (#16334)
• Move expression expansion to conversion module (#16331)
• Add polars-expr README (#16316)
• Move physical expressions to new crate (#16306)
• Use cls (not self) in classmethods (#16303)
• conditionally print the CSEs (#16292)
• Rename ChunkedArray.chunk_id to chunk_lengths (#16273)
• Use Scalar instead of Series some aggregations (#16277)
• Use CsvReadOptions in LazyCsvReader (#16283)
• Do not hardcode bash path in Makefile (#16263)
• Add IR::Reduce (not yet implemented) (#16216)
• Remove incremental read based batched CSV reader (#16259)
• move all describe, describe_tree and dot-viz code to IR instead of DslPlan (#16237)
• move describe to IR instead of DSL (#16191)
• Use Duration.is_zero instead of comparing Duration.duration_ns to 0 (#16195)
• Remove unused code (#16175)
• Don't override CSV reader encoding with lossy UTF-8 (#16151)
• Move CSV read options from CsvReader to CsvReadOptions (#16126)
• Bump sccache action (#16088)
• Fix failures in test coverage workflow (#16083)
• Rename all 'Chunk's to RecordBatch (#16063)
• Use UnionArgs for DSL side (#16017)
• Add some comments (#16008)
• prepare for join coalescing argument (#15418)
• Pin coverage job to MacOS 13 for now (#15918)
• Reorganize from_iter and dispatch to collect_ca when possible (#15904)
• More polars-io cleanup (#15885)
• Improve type-coercion (#15879)
• Move type coercion to IR conversion phase (#15868)
• Reorganize polars_io::parquet module (#15860)
• Reorganize polars_io::csv module (#15831)
• Always expand horizontal_any/all (#15816)
• Rename decimal_float to decimal_comma (#15817)
• Move IO-related options structs to polars-io (#15806)
• Split coverage calculation (#15780)
• Update readme (#15787)
• Move schema resolving of file scan to IR phase (#15739)
• Factor out ensure_is_constant_duration (#15733)
• Move schema resolving to IR phase. (#15714)
• Rename LogicalPlan and builders to reflect their uses better (#15712)

Thank you to all our contributors for making this release possible!
@CanglongCl, @JulianCologne, @KDruzhkin, @MarcoGorelli, @NedJWestern, @NexVeridian, @NickCondron, @Robinsane, @ShivMunagala, @TobiasDummschat, @YichiZhang0613, @alexander-beedie, @avimallu, @bertiewooster, @brandon-b-miller, @c-peters, @coastalwhite, @dangotbanned, @datenzauberai, @deanm0000, @dependabot, @dependabot[bot], @eitsupi, @gasmith, @haocheng6, @ion-elgreco, @itamarst, @janpipek, @jr200, @jrycw, @jsarbach, @luke396, @marenwestermann, @max-muoto, @mbuhidar, @nameexhaustion, @orlp, @pydanny, @r-brink, @reswqa, @ritchie46, @stinodego, @thalassemia, @tharunsuresh-code, @twoertwein, @wence- and @wsyxbcl

⚠️ Deprecations

• Change parameter chunked to allow_chunks in parametric testing strategies (#16264)

🚀 Performance improvements

• Use branchless uleb128 decoding for parquet (#16352)
• Reduce error bubbling in parquet hybrid_rle (#16348)
• use is_sorted in ewm_mean_by, deprecate check_sorted (#16335)
• Optimize is_sorted for numeric data (#16333)
• do not use pyo3-built (#16309)
• Faster bitpacking for Parquet writer (#16278)
• Improve Series.to_numpy performance for chunked Series that would otherwise be zero-copy (#16301)
• Further optimise initial polars import (#16308)
• Avoid importing ctypes.util in CPU check script if possible (#16307)
• Don't rechunk when converting DataFrame to numpy/ndarray (#16288)
• use zeroed vec in ewm_mean_by for sorted fastpath (#16265)

✨ Enhancements

• expose BooleanFunction in expr IR (#16355)
• Allow read_excel to handle bytes/BytesIO directly when using the "calamine" (fastexcel) engine (#16344)
• Raise when joining on the same keys twice (#16329)
• Don't require data to be sorted by by column in rolling_*_by operations (#16249)
• Support List types in Series.to_numpy (#16315)
• Add to_jax methods to support Jax Array export from DataFrame and Series (#16294)
• Enable generating data with time zones in parametric testing (#16298)
• Add struct.field expansion (regex, wildcard, columns) (#16320)
• Add new alpha, alphanumeric and digit selectors (#16310)
• Faster bitpacking for Parquet writer (#16278)
• Add require_all parameter to the by_name column selector (#15028)
• Start updating BytecodeParser for Python 3.13 (#16304)
• Add struct.with_fields (#16305)
• Handle implicit SQL string → temporal conversion in the BETWEEN clause (#16279)
• Expose string expression nodes to python (#16221)
• Add new index/range based selector cs.by_index, allow multiple indices for nth (#16217)
• Show warning if expressions are very deep (#16233)
• Fix some issues in parametric testing with nested dtypes (#16211)

🐞 Bug fixes

• pick a consistent order for the sort options in PyIR (#16350)
• Infer CSV schema as supertype of all files (#16349)
• Fix issue in parametric testing where excluded_dtypes list would grow indefinitely (#16340)
• Address overflow combining u64 hashes in Debug builds (#16323)
• Don't exclude explicitly named columns in group-by context' expr expansion (#16318)
• Improve map_elements typing (#16257)
• Harden Series.reshape against invalid parameters (#16281)
• Fix list.sum dtype for boolean (#16290)
• Don't stackoverflow on all/any horizontal (#16287)
• Fix Series.to_numpy for Array types with nulls and nested Arrays (#16230)
• `rolling_*_by was throwing incorrect error when dataframe was sorted by contained multiple chunks (#16247)
• Don't allow passing missing data to generalized ufuncs (#16198)
• Address overly-permissive expand_selectors function, minor fixes (#16250)
• Add missing support for parsing instantiated Object dtypes Object() (#16260)
• Reading CSV with low_memory gave no data (#16231)
• Add missing read_database overload (#16229)
• Fix a rounding error in parametric test datetimes generation (#16228)
• Fix some issues in parametric testing with nested dtypes (#16211)

📖 Documentation

• Add missing word in join docstring (#16299)
• document that month_start/month_end preserve the current time (#16293)
• Add example for separator parameter in pivot (#15957)

📦 Build system

• Fix allocator features (#16365)
• Update Rust nightly toolchain version (#16222)

🛠️ Other improvements

• Move DataFrame.to_numpy implementation to Rust side (#16354)
• Organize PyO3 NumPy code into interop::numpy module (#16346)
• simplify interpolate code, add test for rolling_*_by with nulls (#16334)
• Very minor refactor of DataFrame.to_numpy code (#16325)
• InterchangeDataFrame.version should be a ClassVar (not a property) (#16312)
• Add polars-expr README (#16316)
• Raise import timing test threshold (#16302)
• Use cls (not self) in classmethods (#16303)
• Use Scalar instead of Series some aggregations (#16277)
• Do not hardcode bash path in Makefile (#16263)
• Add IR::Reduce (not yet implemented) (#16216)
• move all describe, describe_tree and dot-viz code to IR instead of DslPlan (#16237)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @NickCondron, @ShivMunagala, @alexander-beedie, @brandon-b-miller, @coastalwhite, @datenzauberai, @itamarst, @jsarbach, @max-muoto, @nameexhaustion, @orlp, @r-brink, @ritchie46, @stinodego, @thalassemia, @twoertwein and @wence-

⚠️ Deprecations

• Deprecate allow_infinities and null_probability args to parametric test strategies (#16183)

🚀 Performance improvements

• Avoid needless copy when converting chunked Series to NumPy (#16178)
• use zeroable_vec in ewm_mean_by (#16166)
• Improve cost of chunk_idx compute (#16154)
• Don't rechunk by default in concat (#16128)
• Ensure rechunk is parallel (#16127)

✨ Enhancements

• Clarify to_torch "features" and "label" parameter behaviour when return type is not "dataset" (#16218)
• Native CSV file list reading (#16180)
• Register memory mapped files and raise when written to (#16208)
• Implement support for Struct types in parametric tests (#16197)
• Enable Null datatype and null values by default in parametric testing (#16192)
• Support Enum types in parametric testing (#16188)
• Raise when encountering invalid supertype in functions during conversion (#16182)
• Add SQL support for GROUP BY ALL syntax and fix several issues with aliased group keys (#16179)
• Overhaul parametric test implementations and update Hypothesis to latest version (#16062)
• Avoid an extra copy when converting Boolean Series to writable NumPy array (#16164)
• Allow implicit string → temporal conversion in SQL comparisons (#15958)
• Add run-length encoding to Parquet writer (#16125)
• Support passing instantiated adbc/alchemy connection objects to write_database (#16099)
• Add top-level nth(n) method, to go with existing first and last (#16112)

🐞 Bug fixes

• Empty unique (#16214)
• Fix empty drop nulls (#16213)
• Fix get expression group-by state (#16189)
• Fix rolling empty group OOB (#16186)
• offset=-0i was being treated differently to offset=0i in rolling (#16184)
• Fix panic on empty frame joins (#16181)
• Fix streaming glob slice (#16174)
• Fix CSV skip_rows_after_header for streaming (#16176)
• Flush parquet at end of batches tick (#16073)
• Check CSE name aliases for collisions. (#16149)
• Don't override CSV reader encoding with lossy UTF-8 (#16151)
• Respect dtype and strict in pl.Series's constructor for pyarrow arrays, numpy arrays, and pyarrow-backed pandas (#15962)
• Ensure hex and bitstring literals work inside SQL IN clauses (#16101)

📖 Documentation

• Add examples for multiple Series functions (#16172)
• Add deprecated messages to cumfold and cumreduce (#16173)
• StringNameSpace.replace_all docstring (#16169)
• Explain how Polars floor division differs from Python (#16054)
• Clarify arrow usage (#16152)
• Add example to polars.Series.flags (#16123)

🛠️ Other improvements

• Switch over some of the custom Python date/time conversions to native PyO3 conversions (#16203)
• move describe to IR instead of DSL (#16191)
• More PyO3 0.21 Bound<> APIs, and finally disable gil-refs backwards compat feature on pyo3 crate (#16143)
• Move hypothesis tests into unit test module (#16185)
• Remove unused code (#16175)
• Update plugin example to PyO3 0.21 (#16157)
• Return correct temporal type from Rust in to_numpy (#14353)
• Create NumPy view for Datetime/Duration Series on the Rust side (#16148)
• Don't override CSV reader encoding with lossy UTF-8 (#16151)
• Continue converting to PyO3 0.21 Bound<> APIs (#16081)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @YichiZhang0613, @alexander-beedie, @bertiewooster, @coastalwhite, @dangotbanned, @itamarst, @janpipek, @jrycw, @luke396, @nameexhaustion, @pydanny, @ritchie46, @stinodego, @thalassemia and @tharunsuresh-code

🐞 Bug fixes

• Revert "Add RLE to RLE_DICTIONARY encoder"
• Improve error handling of ParameterCollisionError in read_excel (#16100)

Thank you to all our contributors for making this release possible!
@nameexhaustion, @ritchie46 and @wsyxbcl

[!WARNING]
This release was yanked. Please use the 0.20.25 release instead.

🏆 Highlights

• Support pytorch Tensor and Dataset export with new to_torch DataFrame/Series method (#15931)

🚀 Performance improvements

• Don't traverse deep datasets that we repr as union in CSE (#16096)
• Ensure better chunk sizes (#16071)

✨ Enhancements

• split out rolling_*(..., by='foo') into rolling_*_by('foo', ...) (#16059)
• add date pattern dd.mm.YYYY (#16045)
• split Expr.top_k and Expr.top_k_by into separate functions (#16041)
• Support non-coalescing joins in default engine (#16036)
• Support pytorch Tensor and Dataset export with new to_torch DataFrame/Series method (#15931)
• Minor DB type inference updates (#16030)
• Move diagonal & horizontal concat schema resolving to IR phase (#16034)
• raise more informative error messages in rolling_* aggregations instead of panicking (#15979)
• Convert concat during IR conversion (#16016)
• Improve dynamic supertypes (#16009)
• Additional uint datatype support for the SQL interface (#15993)
• Add post-optimization callback (#15972)
• Support Decimal read from IPC (#15965)
• Expose plan and expression nodes through NodeTraverser to Python (#15776)
• Add typed collection from par iterators (#15961)
• Add by argument for Expr.top_k and Expr.bottom_k (#15468)

🐞 Bug fixes

• Respect user passed 'reader_schema' in 'scan_csv' (#16080)
• Lazy csv + projection; respect null values arg (#16077)
• Materialize dtypes when converting to arrow (#16074)
• Fix casting decimal to decimal for high precision (#16049)
• Fix Series constructor failure for Array types for large integers (#16050)
• Fix printing max scale decimals (#16048)
• Decimal supertype for dyn int (#16046)
• Correctly handle large timedelta objects in Series constructor (#16043)
• Do not close connection just because we're not returning Arrow data in batches (#16031)
• properly handle nulls in DictionaryArray::iter_typed (#16013)
• Fix CSE case where upper plan has no projection (#16011)
• Crash/incorrect group_by/n_unique on categoricals created by (q)cut (#16006)
• converting from numpy datetime64 and overriding dtype with a different resolution was returning incorrect results (#15994)
• Ternary supertype dynamics (#15995)
• Fix PartialEq for DataType::Unknown (#15992)
• Finish adding typed_lit to help schema determination in SQL "extract" func (#15955)
• Fix dtype parameter in pandas_to_pyseries function (#15948)
• do not panic when comparing against categorical with incompatible dtype (#15857)
• Join validation for multiple keys (#15947)
• Add missing "truncate_ragged_lines" parameter to read_csv_batched (#15944)

📖 Documentation

• Ensure consistent docstring warning in fill_nan methods (pointing out that nan isn't null) (#16061)
• add filter docstring examples to date and datetime (#15996)
• Fix docstring mistake for polars.concat_str (#15937)
• Update reference to apply (#15982)
• Remove unwanted linebreaks from docstrings (#16002)
• correct default in rolling_* function examples (#16000)
• Improve user-guide doc of UDF (#15923)
• update the link to R API docs (#15973)

🛠️ Other improvements

• Bump sccache action (#16088)
• Fix failures in test coverage workflow (#16083)
• Update benchmarks/coverage jobs with "requirements-ci" (#16072)
• Add TypeGuard to is_polars_dtype util (#16065)
• Clean up hypothesis decimal strategy (#16056)
• split Expr.top_k and Expr.top_k_by into separate functions (#16041)
• Use UnionArgs for DSL side (#16017)
• Add some comments (#16008)
• Improve hypothesis strategy for decimals (#16001)
• Set up TPC-H benchmark tests (#15908)
• Even more Pyo3 0.21 Bound<> APIs (#15914)
• Fix failing test (#15936)

Thank you to all our contributors for making this release possible!
@CanglongCl, @JulianCologne, @KDruzhkin, @MarcoGorelli, @alexander-beedie, @avimallu, @bertiewooster, @c-peters, @dependabot, @dependabot[bot], @eitsupi, @haocheng6, @itamarst, @luke396, @marenwestermann, @nameexhaustion, @orlp, @ritchie46, @stinodego, @thalassemia, @wence- and @wsyxbcl

🏆 Highlights

• Add a low-friction sql method for DataFrame and LazyFrame (#15783)

🚀 Performance improvements

• Don't rechunk in parallel collection (#15907)
• Improve non-trivial list aggregations (#15888)
• Ensure we hit specialized gather for binary/strings (#15886)
• Limit the cache size for to_datetime (#15826)
• skip initial null items and don't recompute slope in interpolate (#15819)

✨ Enhancements

• don't require pyarrow for converting pandas to Polars if all columns have simple numpy-backed datatypes (#15933)
• Add option to disable globbing in csv (#15930)
• Add option to disable globbing in parquet (#15928)
• Expressify dt.round (#15861)
• Improve error messages in context stack (#15881)
• Add dynamic literals to ensure schema correctness (#15832)
• add timestamp time travel in delta scan/read (#15813)

🐞 Bug fixes

• Set default limit for String column display to 30 and fix edge cases (#15934)
• Change recognition of numba ufunc (#15916)
• series.search_sorted could support more types of input (#15940)
• Remove ffspec from parquet reader (#15927)
• avoid WRITE+EXEC for CPUID check (#15912)
• fix inconsistent decimal formatting (#15457)
• Preserve NULLs for is_not_nan (#15889)
• double projection check should only take the upstream projections into account (#15901)
• Ensure we don't create invalid frames when combining unit lit + … (#15903)
• Clear cached rename schema (#15902)
• Fix OOB in struct lit/agg aggregation (#15891)
• Refine interaction of "schema_overrides" with read_excel when using "calamine" engine (#15827)
• Don't modify user-supplied storage_options dict (take a shallow-copy) (#15859)
• create (q)cut labels in fixed order (#15843)
• Tag shrink_dtype as non-streaming (#15828)

📖 Documentation

• improve graphviz install documentation/error message (#15791)
• Extend docstring examples for asof_join (#15810)

📦 Build system

• Don't import jemalloc (#15942)
• Use default allocator for lts-cpu (#15941)
• replace all macos-latest referrals with macos-13 (#15926)
• pin mimalloc and macos-13 (#15925)
• use jemalloc in lts-cpu (#15913)
• Update Cargo.lock (#15865)
• Bump ruff version and improve make clean on the Python side (#15858)
• Exclude rust-toolchain.toml from wheels (#15840)

🛠️ Other improvements

• Replace copy/paste import handling with import_optional utility function (#15906)
• Reorganize from_iter and dispatch to collect_ca when possible (#15904)
• More PyO3 0.21 bound APIs (#15872)
• Improve type-coercion (#15879)
• Move type coercion to IR conversion phase (#15868)
• Fix Python test coverage upload (#15853)
• More upgrades to PyO3 0.21 Bound<> APIs (#15790)
• Use uv for installing Python dependencies in CI (#15848)
• Update benchmark tests (#15825)

Thank you to all our contributors for making this release possible!
@JulianCologne, @MarcoGorelli, @NedJWestern, @NexVeridian, @alexander-beedie, @deanm0000, @dependabot, @dependabot[bot], @ion-elgreco, @itamarst, @jr200, @nameexhaustion, @orlp, @reswqa, @ritchie46 and @stinodego

🚀 Performance improvements

• Improved type-inference for read_excel and read_ods, use calamine engine for read_ods (#15808)
• Fix quadratic in binview growable same source (#15734)
• use two binary searches for equality mask when data is sorted (#15702)
• improve filter parallelism (#15686)

✨ Enhancements

• Minor type-inference update for read_database (#15809)
• Improved type-inference for read_excel and read_ods, use calamine engine for read_ods (#15808)
• dt.truncate supports broadcasting lhs (#15768)
• Expressify str.json_path_match (#15764)
• raise if storage_options is passed to read_csv but fsspec isnt available (#15778)
• Support decimal float parsing in CSV (#15774)
• Add context trace to LazyFrame conversion errors (#15761)
• Improve error message when passing invalid input to lit (#15718)
• Remove outdated join validation checks (#15701)

🐞 Bug fixes

• drop-nulls edge case; remove drop-nulls special case (#15815)
• ewm_mean_by was skipping initial nulls when it was already sorted by "by" column (#15812)
• Consult cgroups to determine free memory (#15798)
• raise if index count like 2i is used when performing rolling, group_by_dynamic, upsample, or other temporal operatios (#15751)
• Don't deduplicate sort that has slice pushdown (#15784)
• Allow passing files opened by fsspec in read_parquet (#15770)
• Fix incorrect is_between pushdown to scan_pyarrow_dataset (#15769)
• Handle null index correctly for list take (#15737)
• Preserve lexical ordering on concat (#15753)
• Remove incorrect unsafe pointer cast for int -> enum (#15740)
• pass series name to apply for cut/qcut (#15715)
• count of null column shouldn't panic in agg context (#15710)
• manual cache (#15711)
• Ensure we don't hold onto Mutex when grabbing join tuples (#15704)
• allow null dtypes in UDFs if they match the schema (#15699)
• Respect join_null argument for semi/anti joins (#15696)
• Ensure we don't hold RwLock when spawning group parallelism in w… (#15697)
• Ensure empty with_columns is a no-op (#15694)
• Include predicate in cache state union (#15693)
• Add the missing feature flag for ewm_mean_by (#15687)
• 8/16-bits int could also apply in place for log expr (#15680)
• prepare_expression_for_context shouldn't panic if exceptions raised from optimizer (#15681)

📖 Documentation

• Add docstring examples for datetimes (#13161) (#15804)
• Fix a typo in categorical section of the user guide (#15777)
• Fix a docstring mistake for DataType.is_float (#15773)
• Remove incorrect "1i (1 index count)" from some docs methods (#15750)
• Add example for Config.set_tbl_width_chars (#15566)
• Align docstring phrasing in Series/Expr.dt.truncate/round (#15698)
• Various deprecation docstring improvements (#15648)

🛠️ Other improvements

• Always expand horizontal_any/all (#15816)
• Rename decimal_float to decimal_comma (#15817)
• Split coverage calculation (#15780)
• Update readme (#15787)
• Start at using new Bound<> API from PyO3 (#15752)
• Make json_path_match expr non-anonymous (#15682)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @NedJWestern, @Robinsane, @TobiasDummschat, @alexander-beedie, @c-peters, @dependabot, @dependabot[bot], @gasmith, @henryharbeck, @itamarst, @kszlim, @mbuhidar, @nameexhaustion, @orlp, @reswqa, @ritchie46, @stinodego and @wsyxbcl

🚀 Performance improvements

• use two binary searches for equality mask when data is sorted (#15702)
• improve filter parallelism (#15686)

✨ Enhancements

• Remove outdated join validation checks (#15701)

🐞 Bug fixes

• manual cache (#15711)
• Ensure we don't hold onto Mutex when grabbing join tuples (#15704)
• allow null dtypes in UDFs if they match the schema (#15699)
• Respect join_null argument for semi/anti joins (#15696)
• Ensure we don't hold RwLock when spawning group parallelism in w… (#15697)
• Ensure empty with_columns is a no-op (#15694)
• Include predicate in cache state union (#15693)
• Add the missing feature flag for ewm_mean_by (#15687)
• 8/16-bits int could also apply in place for log expr (#15680)
• prepare_expression_for_context shouldn't panic if exceptions raised from optimizer (#15681)

🛠️ Other improvements

• Make json_path_match expr non-anonymous (#15682)

Thank you to all our contributors for making this release possible!
@henryharbeck, @kszlim, @orlp, @reswqa and @ritchie46

🚀 Performance improvements

• improve filter parallelism (#15686)

🐞 Bug fixes

• Add the missing feature flag for ewm_mean_by (#15687)
• 8/16-bits int could also apply in place for log expr (#15680)
• prepare_expression_for_context shouldn't panic if exceptions raised from optimizer (#15681)

📖 Documentation

• Various deprecation docstring improvements (#15648)

🛠️ Other improvements

• Make json_path_match expr non-anonymous (#15682)

Thank you to all our contributors for making this release possible!
@henryharbeck, @reswqa and @ritchie46

🚀 Performance improvements

• Fix regression that led to using only a single thread (#15667)

✨ Enhancements

• add ewm_mean_by (#15638)

🐞 Bug fixes

• Ensure profile of simple-projection only take own runtime (#15671)
• Panic if invalid array in object (#15664)
• Ensure 'CachedSchema' doesn't get synced between plans (#15661)
• group_by multiple null columns produce phantom row (#15659)
• rolling_* aggs were behaving as if they return scalars in group-by (#15657)
• Correct the unsoundness slice range of arr.min/max (#15654)
• list.mean fast path shouldn't produce NaN (#15652)
• Fix Display implementation of Duration (#15647)

📖 Documentation

• Fix typo in legacy install instructions (#15662)
• Include prelude import in the example (#15633)

🛠️ Other improvements

• Remove the remaining usage of deprecated numpy crate APIs (#15668)
• make Duration.is_constant_duration less strict for non-timezone-aware case (#15639)
• Fix some typos in comments (#15665)
• remove unnecessary unsafe in list mean/sum (#15660)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @Priyansh4444, @StevenMia, @itamarst, @mcrumiller, @orlp, @reswqa, @ritchie46 and @stinodego

🚀 Performance improvements

• Fix regression that led to using only a single thread (#15667)

✨ Enhancements

• add ewm_mean_by (#15638)

🐞 Bug fixes

• Ensure profile of simple-projection only take own runtime (#15671)
• Panic if invalid array in object (#15664)
• Ensure 'CachedSchema' doesn't get synced between plans (#15661)
• group_by multiple null columns produce phantom row (#15659)
• rolling_* aggs were behaving as if they return scalars in group-by (#15657)
• Correct the unsoundness slice range of arr.min/max (#15654)
• list.mean fast path shouldn't produce NaN (#15652)

📖 Documentation

• Add missing deprecation warning to DataFrame.replace (#15612)
• Fix typo in legacy install instructions (#15662)

🛠️ Other improvements

• make Duration.is_constant_duration less strict for non-timezone-aware case (#15639)
• Fix some typos in comments (#15665)
• remove unnecessary unsafe in list mean/sum (#15660)
• fixup failing test due to offset deprecation in upsample (#15636)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @Priyansh4444, @StevenMia, @eitsupi, @itamarst, @mcrumiller, @orlp, @reswqa, @ritchie46 and @stinodego

🏆 Highlights

• Full plan CSE (#15264)

💥 Breaking changes

• rename memmap -> memory_map as like Python (#15642)
• pref(rust!, python): Unify sort with SortOptions and SortMultipleOptions (#15590)
• Update the argument name from dims to dimensions in reshape (#15561)
• Allow specifying Hive schema in read/scan_parquet (#15434)
• Raise error when schema_overrides contains nonexistent column name (#15290)
• Rename Chunk to RecordBatch (#15298)
• Refactor AnyValue supertype logic (#15280)
• Rename group_by_rolling to rolling and improve related error messages (#14765)
• Rename ChunkedArray.try_apply to try_apply_values (#14947)
• Rename parameter by to group_by in DataFrame.upsample/group_by_dynamic/rolling (#14840)

🚀 Performance improvements

• Fix cross join batch size when one of the DataFrames is tiny (#14347)
• Fix binview growable complexity O(n*m) -> O(n) (#15628)
• Remove extra thread spawn from row group fetcher (#15626)
• Use vertical parallelism if input is chunked for Filter,Select,WithColumns (#15608)
• read_ipc memory usage tests, and writing fix (#15599)
• Refactor CSV serialization to not go thorough AnyValue (#15576)
• don't use dynamic dispatch in visitors (#15607)
• Improve Bitmap construction performance (#15570)
• join by row-encoding (#15559)
• Replace std::thread spawn with tokio block_in_place (#15517)
• speed up offset_by when a single offset is passed (#15493)
• Avoid allocation in the hot path for struct JSON serialization (#15449)
• avoid double-allocation in rolling_apply_agg_window (#15423)
• Make LogicalPlan immutable (#15416)
• Add non-order preserving variable row-encoding (#15414)
• Use row-encoding for multiple key group by (#15392)
• load bits one word at a time for BitmapIter (#15333)
• Ipc exec multiple paths (#15040)
• add SIMD support for if-then-else kernels (#15131)

✨ Enhancements

• add Expr.dt.add_business_days and Series.dt.add_business_days (#15595)
• Add str.head and str.tail (#14425)
• Extended BytecodeParser to handle additional math functions, and imports from the global namespace (#15627)
• Push down is_between expressions to Arrow (#15180)
• add holidays argument to business_day_count (#15580)
• change default to write parquet statistics (#15597)
• Expressify to_integer (#15604)
• Optimizer; remove double SORT and redundant projections (#15573)
• Add null_on_oob parameter to expr.array.get (#15426)
• support weekend argument in business_day_count (#15544)
• Enable is_first/last_distinct for not nested non-numeric list (#15552)
• Turn off cse if cache node found (#15554)
• Tag concat list as elementwise (#15545)
• Support list group-by of non numeric lists (#15540)
• add business_day_count function (#15512)
• Add SQL support for MEDIAN aggfunc (#15519)
• Implement string, boolean and binary dtype in top_k (#15488)
• Add SQL support for TRUNCATE TABLE command (#15513)
• Add SQL support for GREATEST and LEAST (#15511)
• Allow specifying Hive schema in read/scan_parquet (#15434)
• Implements agg_list for NullChunked (#15439)
• Supports explode_by_offsets for decimal (#15417)
• Add null_on_oob parameter to expr.list.get (#15395)
• CSV-writer escape carriage return (#15399)
• Remove 'FileCacher' optimization (#15357)
• check input type in entropy (#15351)
• Implements arr.n_unique (#15296)
• CSE don't scan share if predicate pushdown predicates don't match (#15328)
• Remove cached nodes when finished (#15310)
• Full plan CSE (#15264)
• Add IR for expressions. (#15168)
• Warn if map_elements is called without return_dtype specified (#15188)
• Rename group_by_rolling to rolling and improve related error messages (#14765)
• Rename ChunkedArray.try_apply to try_apply_values (#14947)
• Implement strict AnyValue construction for temporal types (#15146)

🐞 Bug fixes

• Return appropriate data type for time mean and median (#14471)
• Support index upsampling (#13621)
• Fix issue in write_excel that could lead to incorrect spanning range determination (#15631)
• Output correct dtype for mean_horizontal on a single column (#15118)
• Recompute RowIndex schema after projection pd (#15625)
• Mean of boolean in streaming group_by incorrectly always gave NULL (#15616)
• Include cloud creds in cache key (#15609)
• Fix elementwise-apply if any input is AggregatedScalar (#15606)
• Explode list should take validity into account (#15572)
• use larger recursive stack in debug mode (#15593)
• SQL interface "off-by-one' indexing error with GROUP BY clauses that use position ordinals (#15584)
• Enable missing features in polars-time (#15558)
• Handle quoted identifiers when registering CTEs in the SQL engine (#15564)
• Decompress moved out of schema initialization (#15550)
• Turn off cse if cache node found (#15554)
• Resolve function names and prune all aliases. (#15522)
• list.get should take validity into account (#15516)
• block decimal in streaming (#15520)
• group_by partitioned with literal Series panic (#15487)
• Initialize validity for GroupsProxy::Slice windows (#15509)
• Fix struct name resolving (#15507)
• pow return type evaluation (#15506)
• Allow selectors inside frame-level .filter() (#15445)
• Don't prune alias in AnonymousFunction subtree (#15453)
• Fix deadlock in async parquet scan (#15440)
• datetime operations (e.g. .dt.year) were raising when null values were backed by out-of-range integers (#15420)
• Ensure Binary -> Binview cast doesn't overflow the buffer size (#15408)
• Don't prune alias in function subtree (#15406)
• Return 0 for n_unique() in group-by context when group is empty (#15289)
• Unset UpdateGroups after group-sensitive expression (#15400)
• to_any_value should supports all LiteralValue type (#15387)
• Hash failure combining hash of two numeric columns containing equal values (#15397)
• Add FixedSizeBinary to arrow field conversion (#15389)
• Conversion of expr_ir in partition fast path (#15388)
• sort for series with unsupported dtype should raise instead of panic (#15385)
• Return correct dtype for s.clear() when dtype is Object (#15315)
• ensure first datapoint is always included in group_by_dynamic (#15312)
• Non-exhaustive patterns: arrow-schema::DataType in polars-arrow (#15250)
• use dynamic stacks for problematic recursive functions (#15355)
• Raise error when schema_overrides contains nonexistent column name (#15290)
• Fix cache dot visualization (#15311)
• Properly propagate strict flag when constructing a Struct Series from any values (#15302)
• ensure eq for BinaryViewArray checks all elements (#15268)
• Raise when join projects name with suffix that doesn't exist (#15256)
• fix kurtosis/skew (#15137)
• Ensure ooc_start is set (#15255)
• Fix bug where rolling operations were ignoring check_sorted in some cases (#15227)
• Fix lazy schema for rle expression (#15248)
• incorrect negative offset in multi-byte string slicing (#15140)
• do not clamp negative offsets to start of array prematurely (#15242)
• allow null index in list.get and array.get (#15239)
• properly support nulls_last + descending (#15212)
• Block rounding/truncating to negative durations (#15175)
• Make parse_url work on windows with object_store (#15191)
• divide by zero in download speed computation (#15182)

📖 Documentation

• Add legacy CPU install instructions in user guide (#13676)
• Various minor updates to User Guide's SQL intro section (#15557)
• Add outer_coalesce join strategy in the user guide (#15405)
• Improve docs for Series::new with AnyValue input (#15306)
• Fix formatting in Series::from_any_values_and_dtype docs (#15244)
• Correct the definition of an expression in the user guide (#14750)

📦 Build system

• Fix a feature gate for lz4 compression in polars-parquet (#15565)
• Update Rust toolchain (#15353)

🛠️ Other improvements

• rename memmap -> memory_map as like Python (#15642)
• fixup failing test due to offset deprecation in upsample (#15636)
• use bound api (#15630)
• Don't run streaming group-by in partitionable gb (#15611)
• pref(rust!, python): Unify sort with SortOptions and SortMultipleOptions (#15590)
• remove try_binary_elementwise_values (#15592)
• remove raw pointers from visitors. (#15579)
• rename to IR (#15571)
• Update the argument name from dims to dimensions in reshape (#15561)
• Rename ALogicalPlan to FullAccessIR (#15553)
• Set up CodSpeed (#15537)
• make dsl immutable and cheap to clone (#15394)
• use recursive crate, add missing recursive tag (#15393)
• Update CODEOWNERS (polars-sql) (#15384)
• Update Rust toolchain (#15353)
• Update CODEOWNERS (#15352)
• remove try_apply_values (#15336)
• always use non-legacy float_sum for mean (#15343)
• remove legacy bitmap module (#15335)
• More clippy in Makefile (#15340)
• Rename Cache[count] to Cache[cache_hits] (#15300)
• Cleanup file_caching optimization call (#15299)
• Rename Chunk to RecordBatch (#15298)
• Refactor AnyValue supertype logic (#15280)
• reuse message parsing in IPC (#15265)
• remove 'fast-projection' node (#15253)
• cleanup column names in optimizer (#15252)
• remove left_most_input_name from expr ir (#15251)
• add AlignedBitmapSlice (#15171)
• Refactor AnyValue construction for Categorical/Enum dtype (#15220)
• Move ConsecutiveCountState into support module (#15186)
• Run non-benchmark tests in benchmark workflow (#15207)
• Add wrapping_abs to arithmetic kernel (#15210)
• remove raw buffers from BinViewArray (#15206)
• Enable RUST_BACKTRACE=1 in the CI test suite (#15204)
• Rename parameter by to group_by in DataFrame.upsample/group_by_dynamic/rolling (#14840)
• Set dual license for polars-arrow and polars-parquet (#15173)
• remove parts of legacy bit_util (#15169)
• remove legacy arrow compute (#15164)

Thank you to all our contributors for making this release possible!
@CanglongCl, @ChayimFriedman2, @Fokko, @JamesCE2001, @MarcoGorelli, @NedJWestern, @Sol-Hee, @TrevorWinstral, @alexander-beedie, @braaannigan, @c-peters, @cmdlineluser, @cojmeister, @deanm0000, @dependabot, @dependabot[bot], @douglas-raillard-arm, @eitsupi, @filabrazilska, @henryharbeck, @i-aki-y, @itamarst, @kszlim, @leoforney, @mbuhidar, @mcrumiller, @mickvangelderen, @nameexhaustion, @orlp, @ozgrakkurt, @petrosbar, @reswqa, @ritchie46, @rob-sil, @sportfloh, @stinodego, @thomaslin2020 and @yutannihilation

🚀 Performance improvements

• Fix cross join batch size when one of the DataFrames is tiny (#14347)
• Fix binview growable complexity O(n*m) -> O(n) (#15628)
• Remove extra thread spawn from row group fetcher (#15626)
• Use vertical parallelism if input is chunked for Filter,Select,WithColumns (#15608)
• Refactor CSV serialization to not go thorough AnyValue (#15576)
• don't use dynamic dispatch in visitors (#15607)
• Improve Bitmap construction performance (#15570)
• join by row-encoding (#15559)

✨ Enhancements

• add Expr.dt.add_business_days and Series.dt.add_business_days (#15595)
• Add str.head and str.tail (#14425)
• Add union/or operator for pl.Enum (#14965)
• Extended BytecodeParser to handle additional math functions, and imports from the global namespace (#15627)
• Push down is_between expressions to Arrow (#15180)
• add holidays argument to business_day_count (#15580)
• change default to write parquet statistics (#15597)
• Expressify to_integer (#15604)
• Optimizer; remove double SORT and redundant projections (#15573)
• Add null_on_oob parameter to expr.array.get (#15426)
• support weekend argument in business_day_count (#15544)
• Enable is_first/last_distinct for not nested non-numeric list (#15552)
• Turn off cse if cache node found (#15554)
• Tag concat list as elementwise (#15545)

🐞 Bug fixes

• Return appropriate data type for time mean and median (#14471)
• Fix issue in write_excel that could lead to incorrect spanning range determination (#15631)
• Output correct dtype for mean_horizontal on a single column (#15118)
• Recompute RowIndex schema after projection pd (#15625)
• Mean of boolean in streaming group_by incorrectly always gave NULL (#15616)
• Include cloud creds in cache key (#15609)
• Fix elementwise-apply if any input is AggregatedScalar (#15606)
• Explode list should take validity into account (#15572)
• use larger recursive stack in debug mode (#15593)
• SQL interface "off-by-one' indexing error with GROUP BY clauses that use position ordinals (#15584)
• Enable missing features in polars-time (#15558)
• Handle quoted identifiers when registering CTEs in the SQL engine (#15564)
• Decompress moved out of schema initialization (#15550)
• Turn off cse if cache node found (#15554)

📖 Documentation

• Add legacy CPU install instructions in user guide (#13676)
• Examples for errors (#13724)
• Add docstring examples for reading json (#14481)
• Add security warning in LazyFrame.deserialize() docstring (#15282)
• Various minor updates to User Guide's SQL intro section (#15557)

🛠️ Other improvements

• Replace most deprecated calls with bounded version (#15632)
• use bound api (#15630)
• Initial PyO3 0.21 support (#15622)
• Don't run streaming group-by in partitionable gb (#15611)
• pref(rust!, python): Unify sort with SortOptions and SortMultipleOptions (#15590)
• Set up CodSpeed (#15537)

Thank you to all our contributors for making this release possible!
@CanglongCl, @ChayimFriedman2, @Fokko, @JamesCE2001, @MarcoGorelli, @NedJWestern, @TrevorWinstral, @alexander-beedie, @deanm0000, @douglas-raillard-arm, @eitsupi, @filabrazilska, @i-aki-y, @itamarst, @leoforney, @mcrumiller, @nameexhaustion, @orlp, @ozgrakkurt, @reswqa, @ritchie46 and @stinodego

🚀 Performance improvements

• Replace std::thread spawn with tokio block_in_place (#15517)
• speed up offset_by when a single offset is passed (#15493)
• Avoid allocation in the hot path for struct JSON serialization (#15449)

✨ Enhancements

• Support list group-by of non numeric lists (#15540)
• add business_day_count function (#15512)
• Add SQL support for MEDIAN aggfunc (#15519)
• Implement string, boolean and binary dtype in top_k (#15488)
• Add SQL support for TRUNCATE TABLE command (#15513)
• Add SQL support for GREATEST and LEAST (#15511)
• Allow specifying Hive schema in read/scan_parquet (#15434)
• Implements agg_list for NullChunked (#15439)

🐞 Bug fixes

• dot product of two integer series is cast to float (#15502)
• Resolve function names and prune all aliases. (#15522)
• Pass skip_rows_after_header to pyarrow csv reader (#15533)
• No longer error when schema_overrides contains nonexistent columns (#15528)
• list.get should take validity into account (#15516)
• block decimal in streaming (#15520)
• group_by partitioned with literal Series panic (#15487)
• Initialize validity for GroupsProxy::Slice windows (#15509)
• Fix struct name resolving (#15507)
• pow return type evaluation (#15506)
• Address issue with read_database draining iter_batches early (#15504)
• Allow selectors inside frame-level .filter() (#15445)
• Don't prune alias in AnonymousFunction subtree (#15453)
• Raise if pass a negative n into clear (#15432)
• Fix deadlock in async parquet scan (#15440)

📖 Documentation

• Update leftover references of by parameter to group_by in DataFrame/LazyFrame.upsample/group_by_dynamic/rolling (#15527)
• Add make docs command, DataType docs/layout tweak, minor README updates (#15386)
• Add example for Series.list.median. (#15451)

🛠️ Other improvements

• Remove unused code paths in read_parquet (#15532)
• Organize utils for I/O functionality (#15529)
• Remove private DataFrame._read classmethods (#15521)
• Move dedicated inference code out of io.database executor module (#15526)
• Add unstable warning to hive_schema functionality (#15508)

Thank you to all our contributors for making this release possible!
@CanglongCl, @ChayimFriedman2, @MarcoGorelli, @alexander-beedie, @cmdlineluser, @dependabot, @dependabot[bot], @henryharbeck, @mbuhidar, @nameexhaustion, @reswqa, @ritchie46, @rob-sil and @stinodego

🚀 Performance improvements

• CSV reading memory usage tests and fixes (#15422)
• avoid double-allocation in rolling_apply_agg_window (#15423)
• Make LogicalPlan immutable (#15416)
• Add non-order preserving variable row-encoding (#15414)
• Use row-encoding for multiple key group by (#15392)

✨ Enhancements

• Supports explode_by_offsets for decimal (#15417)
• Add read_clipboard and DataFrame.write_clipboard (#15272)
• Add null_on_oob parameter to expr.list.get (#15395)
• make Series.__bool__ error message Rusttier (#15407)
• CSV-writer escape carriage return (#15399)

🐞 Bug fixes

• datetime operations (e.g. .dt.year) were raising when null values were backed by out-of-range integers (#15420)
• Ensure Binary -> Binview cast doesn't overflow the buffer size (#15408)
• Don't prune alias in function subtree (#15406)
• Return 0 for n_unique() in group-by context when group is empty (#15289)
• Unset UpdateGroups after group-sensitive expression (#15400)
• to_any_value should supports all LiteralValue type (#15387)
• Hash failure combining hash of two numeric columns containing equal values (#15397)
• Add FixedSizeBinary to arrow field conversion (#15389)
• Conversion of expr_ir in partition fast path (#15388)
• fix panic when doing a scan_parquet with hive partioning (#15381)
• sort for series with unsupported dtype should raise instead of panic (#15385)

📖 Documentation

• Added example for explode mapping strategy in pl.Expr.over (#15402)
• Add outer_coalesce join strategy in the user guide (#15405)
• Change the example to series for series/array.py (#15383)
• Add "See Also" for arg_sort and arg_sort_by (#15348)

🛠️ Other improvements

• make dsl immutable and cheap to clone (#15394)
• use recursive crate, add missing recursive tag (#15393)
• Update CODEOWNERS (polars-sql) (#15384)

Thank you to all our contributors for making this release possible!
@CanglongCl, @JamesCE2001, @MarcoGorelli, @Sol-Hee, @alexander-beedie, @dependabot, @dependabot[bot], @itamarst, @kszlim, @mcrumiller, @nameexhaustion, @orlp, @reswqa, @ritchie46, @rob-sil and @thomaslin2020

🏆 Highlights

• Full plan CSE (#15264)

⚠️ Deprecations

• Rename parameter by to group_by in DataFrame.upsample/group_by_dynamic/rolling (#14840)
• Rename from_repr parameter from tbl to data (#15156)

🚀 Performance improvements

• load bits one word at a time for BitmapIter (#15333)
• Ipc exec multiple paths (#15040)
• add SIMD support for if-then-else kernels (#15131)

✨ Enhancements

• Remove 'FileCacher' optimization (#15357)
• check input type in entropy (#15351)
• Implements arr.n_unique (#15296)
• CSE don't scan share if predicate pushdown predicates don't match (#15328)
• Add read_database support for SurrealDB ("ws" and "http") (#15269)
• Only allow inputs of type Sequence in from_records (#15329)
• In hypothesis testing strategies, enable Decimal strategy by default (#15321)
• Remove cached nodes when finished (#15310)
• Full plan CSE (#15264)
• More robust handling of async database calls (#15202)
• Add name parameter to GroupBy.len method (#15235)
• Add IR for expressions. (#15168)
• Improve read_database when reading from Kùzu graph database (#15218)
• Warn if map_elements is called without return_dtype specified (#15188)
• Add support for async SQLAlchemy connections to read_database (#15162)
• Infer time_unit in pl.duration when nanoseconds is specified (#14987)
• Add strict parameter to from_dict/from_records (#15158)

🐞 Bug fixes

• Return correct dtype for s.clear() when dtype is Object (#15315)
• ensure first datapoint is always included in group_by_dynamic (#15312)
• Non-exhaustive patterns: arrow-schema::DataType in polars-arrow (#15250)
• use dynamic stacks for problematic recursive functions (#15355)
• Adding default ddof for Series.list.std and Series.list.var (#15267)
• Raise properly for slices not supported by LazyFrame (#15331)
• Propagate strictness in from_dicts (#15344)
• Raise error when schema_overrides contains nonexistent column name (#15290)
• Enforce integer dtype input for int_range and int_ranges (#15339)
• Preserve Decimal precision when constructing empty Series (#15320)
• Fix cache dot visualization (#15311)
• Handle special case correctly when slicing a LazyFrame (#15297)
• Properly propagate strict flag when constructing a Struct Series from any values (#15302)
• Consistent expansion of nested struct data during DataFrame init from dict (#15217)
• Raise when join projects name with suffix that doesn't exist (#15256)
• Ensure ooc_start is set (#15255)
• Fix bug where rolling operations were ignoring check_sorted in some cases (#15227)
• Fix lazy schema for rle expression (#15248)
• incorrect negative offset in multi-byte string slicing (#15140)
• do not clamp negative offsets to start of array prematurely (#15242)
• allow null index in list.get and array.get (#15239)
• Avoid loading all columns in read_parquet when columns parameter is specified (#15229)
• properly support nulls_last + descending (#15212)
• fix nested runtime panic (#15216)
• Block rounding/truncating to negative durations (#15175)
• Ensure the cs.temporal() selector uses wildcard time zone matching for Datetime (#13683)
• Consistently raise TypeError on constructor failure (#15178)
• Properly propagate strictness in some constructor cases (#15166)
• Fix constructing a Series from a list of Series with given dtype (#15144)

📖 Documentation

• Fix time unit in timestamp example (#15281)
• Fix link to renamed method (.list.lengths -> .list.len) (#15228)
• Update Excel and database pages in user guide (#14721)
• Add examples for Series.search_sorted (#14737)
• Correct the definition of an expression in the user guide (#14750)
• Add a note about the behaviour of lower/upper bounds for is_between, and add an example (#15197)

📦 Build system

• Update Cargo lock (#15370)

🛠️ Other improvements

• Memory usage test infrastructure, plus a test for #15098 (#15285)
• Update CODEOWNERS (#15352)
• remove try_apply_values (#15336)
• always use non-legacy float_sum for mean (#15343)
• remove legacy bitmap module (#15335)
• Fix test not writing to temporary directory (#15318)
• Reorganize tests for clear operation (#15304)
• Rename Cache[count] to Cache[cache_hits] (#15300)
• Cleanup file_caching optimization call (#15299)
• Minor refactor of PyDataFrame.from_dicts (#15274)
• remove 'fast-projection' node (#15253)
• cleanup column names in optimizer (#15252)
• remove left_most_input_name from expr ir (#15251)
• add AlignedBitmapSlice (#15171)
• Run non-benchmark tests in benchmark workflow (#15207)
• Add wrapping_abs to arithmetic kernel (#15210)
• remove raw buffers from BinViewArray (#15206)
• Enable RUST_BACKTRACE=1 in the CI test suite (#15204)
• Split read_database functionality into cleaner module structure (#15201)
• Clean up some of the AnyValue conversion logic (#15190)
• remove parts of legacy bit_util (#15169)
• remove legacy arrow compute (#15164)
• Split up dataframe module in PyO3 bindings (#15165)
• Remove unused private constructors (#15160)

Thank you to all our contributors for making this release possible!
@MarcoGorelli, @alexander-beedie, @braaannigan, @c-peters, @cojmeister, @deanm0000, @dependabot, @dependabot[bot], @itamarst, @kszlim, @mbuhidar, @mcrumiller, @mickvangelderen, @orlp, @petrosbar, @reswqa, @ritchie46, @rob-sil, @sportfloh, @stinodego and @yutannihilation

🚀 Performance improvements

• add new when-then-otherwise kernels (#15089)
• Coerce sorted flag of unit arrays during concat (#15104)
• Use sorted flag for (first|last)_non_null (#15050)
• OOC sort improvements (#14994)

✨ Enhancements

• raise if both closed and by are passed to rolling_* aggregations (#15108)
• raise informative error for rolling_* aggs with by of invalid dtype (#15088)
• add non_existent arg to replace_time_zone (#15062)
• Support single nested row encodings (#15105)
• make ooc sort configurable (#15084)
• Make register_plugin a standalone function and include shared lib discovery (#14804)
• Async parquet: Decode parquet on a blocking thread pool (#15083)
• let "ambiguous" take "null" value (#14961)
• Raise informative error message when join would introduce duplicate column name (#15042)
• Allow cast of decimal to boolean (#15015)
• Return error when no supertype can be determined in AnyValue constructor when strict=false (#15025)
• Implement IpcReaderAsync (#14984)
• Support Array statistics in parquet (#15031)
• Support decimal groupby (#15000)
• Add thread names to rayon thread pool (#15024)
• Support decimal uniq (#15001)
• expose timings in verbose state of OOC sort (#14979)

🐞 Bug fixes

• Fix Series construction from nested list with mixed data types (#15046)
• Support BinaryView in row decoder to prevent a panic in streaming group by (#15117)
• Binview chunked gather; don't modify inlined view (#15124)
• Fix chunked_id gather for binview buffers (#15123)
• Don't cache HTTP object stores as they maintain URL state (#15121)
• use wrapping_add in csv line snooping (#15109)
• Output u32 when sum_horizontal provided with single boolean column (#15114)
• Ensure eprintln! is only called within debug/verbose context (#15100)
• Propagate error instead of panicking when calling product on an invalid type (#15093)
• Raise error when casting Array to different width (#14995)
• Fix file scan bugs for ipc, csv and parquet that occur with combinations of glob paths, row indices and predicates (#15065)
• Incorrectly preserved sorted flag when concatenating sorted series containing nulls (#15082)
• Return largest non-NaN value for max() on sorted float arrays if it exists instead of NaN (#15060)
• return NaN for all-NaN min/max (#15066)
• Prevent "index out of range for slice" error in parquet reader (#15021)
• Respect nulls_last in streaming sort (#15061)
• Fix Series construction from nested list with mixed data types (#15046)
• Don't count nulls in streaming count agg (#15051)
• agg_list on decimal lost scale (#15054)
• Block predicate pushdown on equality that are use in join (#15055)
• Enum equality based on categories (#15053)
• Strict cast in when/then/otherwise operation (#15052)
• Don't panic in string_addition_to_linear_concat (#15006)
• CSV do utf8-validation after escaping fields (#15004)
• Use primitive constructors to create a Series of lists when dtype is provided (#15002)
• replace_time_zone with single-null-element "ambiguous" was panicking (#14971)

📖 Documentation

• Fix typo in comment (#14997)

🛠️ Other improvements

• Extend and speed up scan tests (#15127)
• always assert on ChunkedArray::get (#15120)
• Use ObjectStore instead of AsyncRead in parquet get metadata (#15069)
• Minor refactor of Rust any value constructors (#15077)
• Simplify streaming execution (#15039)
• Ensure we hit the spilled source path in ooc sort test (#15010)
• Refactor constructor code (#15009)
• Apply clippy:assigning_clones lint (#14999)
• fix features (#14977)

Thank you to all our contributors for making this release possible!
@JackRolfe, @MKisilyov, @MarcoGorelli, @alexander-beedie, @c-peters, @flisky, @jqnatividad, @mcrumiller, @mickvangelderen, @nameexhaustion, @orlp, @petrosbar, @ritchie46, @stinodego and @trueb2

🚀 Performance improvements

• add new when-then-otherwise kernels (#15089)
• Coerce sorted flag of unit arrays during concat (#15104)
• Use sorted flag for (first|last)_non_null (#15050)
• OOC sort improvements (#14994)

✨ Enhancements

• improved dtype inference/refinement for read_database results (#15126)
• raise if both closed and by are passed to rolling_* aggregations (#15108)
• raise informative error for rolling_* aggs with by of invalid dtype (#15088)
• add non_existent arg to replace_time_zone (#15062)
• Support single nested row encodings (#15105)
• make ooc sort configurable (#15084)
• Make register_plugin a standalone function and include shared lib discovery (#14804)
• Expose infer_schema_length parameter on read_database (#15076)
• Async parquet: Decode parquet on a blocking thread pool (#15083)
• let "ambiguous" take "null" value (#14961)
• Raise informative error message when join would introduce duplicate column name (#15042)
• Allow cast of decimal to boolean (#15015)
• Add strict parameter to DataFrame constructor to allow non-strict construction (#15034)
• Support Array statistics in parquet (#15031)
• Support decimal groupby (#15000)
• Add thread names to rayon thread pool (#15024)
• Support decimal uniq (#15001)
• expose timings in verbose state of OOC sort (#14979)

🐞 Bug fixes

• Support BinaryView in row decoder to prevent a panic in streaming group by (#15117)
• Binview chunked gather; don't modify inlined view (#15124)
• Fix chunked_id gather for binview buffers (#15123)
• Don't cache HTTP object stores as they maintain URL state (#15121)
• Output u32 when sum_horizontal provided with single boolean column (#15114)
• Propagate error instead of panicking when calling product on an invalid type (#15093)
• Raise error when casting Array to different width (#14995)
• Fix file scan bugs for ipc, csv and parquet that occur with combinations of glob paths, row indices and predicates (#15065)
• Incorrectly preserved sorted flag when concatenating sorted series containing nulls (#15082)
• Return largest non-NaN value for max() on sorted float arrays if it exists instead of NaN (#15060)
• return NaN for all-NaN min/max (#15066)
• Prevent "index out of range for slice" error in parquet reader (#15021)
• Respect nulls_last in streaming sort (#15061)
• Fix Series construction from nested list with mixed data types (#15046)
• Don't count nulls in streaming count agg (#15051)
• agg_list on decimal lost scale (#15054)
• Block predicate pushdown on equality that are use in join (#15055)
• Enum equality based on categories (#15053)
• Don't panic in string_addition_to_linear_concat (#15006)
• CSV do utf8-validation after escaping fields (#15004)
• Use primitive constructors to create a Series of lists when dtype is provided (#15002)
• replace_time_zone with single-null-element "ambiguous" was panicking (#14971)

📖 Documentation

• Update write_database code blocks in user guide (#15106)
• Add missing docstring examples in the Struct namespace (#15071)
• Improve API reference landing page (#14888)
• improve join_asof example (#14993)
• Fix inadvertent swap of new and old parameters in replace description (#15019)

🛠️ Other improvements

• Extend and speed up scan tests (#15127)
• Add parameterized-scan-tests (#15057)
• Simplify streaming execution (#15039)
• Ensure we hit the spilled source path in ooc sort test (#15010)
• Refactor constructor code (#15009)
• fix features (#14977)
• Revert pinning PyPI publish action (#14975)

Thank you to all our contributors for making this release possible!
@JackRolfe, @MKisilyov, @MarcoGorelli, @alexander-beedie, @c-peters, @flisky, @jqnatividad, @mcrumiller, @mickvangelderen, @nameexhaustion, @orlp, @petrosbar, @ritchie46, @stinodego and @trueb2