rs-0.53.0

## 🏆 Highlights

- Add Extension types (#25322)

## 🚀 Performance improvements

- Don't always rechunk on gather of nested types (#26478)
- Enable zero-copy object\_store `put` upload for IPC sink (#26288)
- Resolve file schema's and metadata concurrently (#26325)
- Run elementwise CSEE for the streaming engine (#26278)
- Disable morsel splitting for fast-count on streaming engine (#26245)
- Implement streaming decompression for scan\_ndjson and scan\_lines (#26200)
- Improve string slicing performance (#26206)
- Refactor `scan_delta` to use python dataset interface (#26190)
- Add dedicated kernel for group-by `arg_max/arg_min` (#26093)
- Add streaming merge-join (#25964)
- Generalize Bitmap::new\_zeroed opt for Buffer::zeroed (#26142)
- Reduce fs stat calls in path expansion (#26173)
- Lower streaming group\_by n\_unique to unique().len() (#26109)
- Speed up `SQL` interface "UNION" clauses (#26039)
- Speed up `SQL` interface "ORDER BY" clauses (#26037)
- Add fast kernel for is\_nan and use it for numpy NaN->null conversion (#26034)
- Optimize ArrayFromIter implementations for ObjectArray (#25712)
- New streaming NDJSON sink pipeline (#25948)
- New streaming CSV sink pipeline (#25900)
- Dispatch partitioned usage of `sink_*` functions to new-streaming by default (#25910)
- Replace ryu with faster zmij (#25885)
- Reduce memory usage for .item() count in grouped first/last (#25787)
- Skip schema inference if schema provided for `scan_csv/ndjson` (#25757)
- Add width-aware chunking to prevent degradation with wide data (#25764)
- Use new sink pipeline for write/sink\_ipc (#25746)
- Reduce memory usage when scanning multiple parquet files in streaming (#25747)
- Don't call cluster\_with\_columns optimization if not needed (#25724)
- Tune partitioned sink\_parquet cloud performance (#25687)
- New single file IO sink pipeline enabled for sink\_parquet (#25670)
- New partitioned IO sink pipeline enabled for sink\_parquet (#25629)
- Correct overly eager local predicate insertion for unpivot (#25644)
- Reduce HuggingFace API calls (#25521)
- Use strong hash instead of traversal for CSPE equality (#25537)
- Fix panic in is\_between support in streaming Parquet predicate push down (#25476)
- Faster kernels for rle\_lengths (#25448)
- Allow detecting plan sortedness in more cases (#25408)
- Enable predicate expressions on unsigned integers (#25416)
- Mark output of more non-order-maintaining ops as unordered (#25419)
- Fast find start window in `group_by_dynamic` with large `offset` (#25376)
- Add streaming native `LazyFrame.group_by_dynamic` (#25342)
- Add streaming sorted Group-By (#25013)
- Add parquet prefiltering for string regexes (#25381)
- Use fast path for `agg_min`/`agg_max` when nulls present (#25374)
- Fuse positive `slice` into streaming `LazyFrame.rolling` (#25338)
- Mark `Expr.reshape((-1,))` as row separable (#25326)
- Use bitmap instead of Vec\<bool> in first/last w. skip\_nulls (#25318)
- Return references from `aexpr_to_leaf_names_iter` (#25319)

## ✨ Enhancements

- Add primitive filter -> agg lowering in streaming GroupBy (#26459)
- Support for the SQL `FETCH` clause (#26449)
- Add get() to retrieve a byte from binary data (#26454)
- Remove with\_context in SQL lowering (#26416)
- Avoid OOM for scan\_ndjson and scan\_lines if input is compressed and negative slice (#26396)
- Add JoinBuildSide (#26403)
- Support annoymous agg in-mem (#26376)
- Add unstable `arrow_schema` parameter to `sink_parquet` (#26323)
- Improve error message formatting for structs (#26349)
- Remove parquet field overwrites (#26236)
- Enable zero-copy object\_store `put` upload for IPC sink (#26288)
- Improved disambiguation for qualified wildcard columns in SQL projections (#26301)
- Expose `upload_concurrency` through env var (#26263)
- Allow quantile to compute multiple quantiles at once (#25516)
- Allow empty LazyFrame in `LazyFrame.group_by(...).map_groups` (#26275)
- Use delta file statistics for batch predicate pushdown (#26242)
- Add streaming UnorderedUnion (#26240)
- Implement compression support for sink\_ndjson (#26212)
- Add unstable record batch statistics flags to `{sink/scan}_ipc` (#26254)
- Cloud retry/backoff configuration via `storage_options` (#26204)
- Use same sort order for expanded paths across local / cloud / directory / glob (#26191)
- Expose physical plan `NodeStyle` (#26184)
- Add streaming merge-join (#25964)
- Serialize optimization flags for cloud plan (#26168)
- Add compression support to write\_csv and sink\_csv (#26111)
- Add `scan_lines` (#26112)
- Support regex in `str.split` (#26060)
- Add unstable IPC Statistics read/write to `scan_ipc`/`sink_ipc` (#26079)
- Add nulls support for all rolling\_by operations (#26081)
- ArrowStreamExportable and sink\_delta (#25994)
- Release musl builds (#25894)
- Implement streaming decompression for CSV `COUNT(*)` fast path (#25988)
- Add nulls support for rolling\_mean\_by (#25917)
- Add lazy `collect_all` (#25991)
- Add streaming decompression for NDJSON schema inference (#25992)
- Improved handling of unqualified SQL `JOIN` columns that are ambiguous (#25761)
- Expose record batch size in `{sink,write}_ipc` (#25958)
- Add `null_on_oob` parameter to `expr.get` (#25957)
- Suggest correct timezone if timezone validation fails (#25937)
- Support streaming IPC scan from S3 object store (#25868)
- Implement streaming CSV schema inference (#25911)
- Support hashing of meta expressions (#25916)
- Improve `SQLContext` recognition of possible table objects in the Python globals (#25749)
- Add pl.Expr.(min|max)\_by (#25905)
- Improve MemSlice Debug impl (#25913)
- Implement or fix json encode/decode for (U)Int128, Categorical, Enum, Decimal (#25896)
- Expand scatter to more dtypes (#25874)
- Implement streaming CSV decompression (#25842)
- Add Series `sql` method for API consistency (#25792)
- Mark Polars as safe for free-threading (#25677)
- Support Binary and Decimal in arg\_(min|max) (#25839)
- Allow Decimal parsing in str.json\_decode (#25797)
- Add `shift` support for Object data type (#25769)
- Add node status to NodeMetrics (#25760)
- Allow scientific notation when parsing Decimals (#25711)
- Allow creation of `Object` literal (#25690)
- Don't collect schema in SQL union processing (#25675)
- Add `bin.slice()`, `bin.head()`, and `bin.tail()` methods (#25647)
- Add SQL support for the `QUALIFY` clause (#25652)
- New partitioned IO sink pipeline enabled for sink\_parquet (#25629)
- Add SQL syntax support for `CROSS JOIN UNNEST(col)` (#25623)
- Add separate env var to log tracked metrics (#25586)
- Expose fields for generating physical plan visualization data (#25562)
- Allow pl.Object in pivot value (#25533)
- Extend SQL `UNNEST` support to handle multiple array expressions (#25418)
- Minor improvement for `as_struct` repr (#25529)
- Temporal `quantile` in rolling context (#25479)
- Add support for `Float16` dtype (#25185)
- Add strict parameter to pl.concat(how='horizontal') (#25452)
- Add leftmost option to `str.replace_many / str.find_many / str.extract_many` (#25398)
- Add `quantile` for missing temporals (#25464)
- Expose and document pl.Categories (#25443)
- Support decimals in search\_sorted (#25450)
- Use reference to Graph pipes when flushing metrics (#25442)
- Add SQL support for named `WINDOW` references (#25400)
- Add Extension types (#25322)
- Add `having` to `group_by` context (#23550)
- Allow elementwise `Expr.over` in aggregation context (#25402)
- Add SQL support for `ROW_NUMBER`, `RANK`, and `DENSE_RANK` functions (#25409)
- Automatically Parquet dictionary encode floats (#25387)
- Add `empty_as_null` and `keep_nulls` to `{Lazy,Data}Frame.explode` (#25369)
- Allow `hash` for all `List` dtypes (#25372)
- Support `unique_counts` for all datatypes (#25379)
- Add `maintain_order` to `Expr.mode` (#25377)
- Display function of streaming physical plan `map` node (#25368)
- Allow `slice` on scalar in aggregation context (#25358)
- Allow `implode` and aggregation in aggregation context (#25357)
- Add `empty_as_null` and `keep_nulls` flags to `Expr.explode` (#25289)
- Add `ignore_nulls` to `first` / `last` (#25105)
- Move GraphMetrics into StreamingQuery (#25310)
- Allow `Expr.unique` on `List`/`Array` with non-numeric types (#25285)
- Allow `Expr.rolling` in aggregation contexts (#25258)
- Support additional forms of SQL `CREATE TABLE` statements (#25191)
- Add `LazyFrame.pivot` (#25016)
- Support column-positional SQL `UNION` operations (#25183)
- Allow arbitrary expressions as the `Expr.rolling` `index_column` (#25117)
- Allow arbitrary Expressions in "subset" parameter of `unique` frame method (#25099)
- Support arbitrary expressions in SQL `JOIN` constraints (#25132)

## 🐞 Bug fixes

- Do not overwrite used names in cluster\_with\_columns pushdown (#26467)
- Do not mark output of concat\_str on multiple inputs as sorted (#26468)
- Fix CSV schema inference content line duplication bug (#26452)
- Fix InvalidOperationError using `scan_delta` with filter (#26448)
- Alias giving missing column after streaming GroupBy CSE (#26447)
- Ensure `by_name` selector selects only names (#26437)
- Restore compatibility of strings written to parquet with pyarrow filter (#26436)
- Update schema in cluster\_with\_columns optimization (#26430)
- Fix negative slice in groups slicing (#26442)
- Don't run CPU check on aarch64 musl (#26439)
- Remove the `POLARS_IDEAL_MORSEL_SIZE` monkeypatching in the parametric merge-join test (#26418)
- Correct off-by-one in RLE row counting for nullable dictionary-encoded columns (#26411)
- Support very large integers in env var limits (#26399)
- Fix PlPath panic from incorrect slicing of UTF8 boundaries (#26389)
- Fix Float dtype for spearman correlation (#26392)
- Fix optimizer panic in right joins with type coercion (#26365)
- Don't serialize retry config from local environment vars (#26289)
- Fix `PartitionBy` with scalar key expressions and `diff()` (#26370)
- Add {Float16, Float32} -> Float32 lossless upcast (#26373)
- Fix panic using `with_columns` and `collect_all` (#26366)
- Add multi-page support for writing dictionary-encoded Parquet columns (#26360)
- Ensure slice advancement when skipping non-inlinable values in `is_in` with inlinable needles (#26361)
- Pin `xlsx2csv` version temporarily (#26352)
- Bugs in ViewArray total\_bytes\_len (#26328)
- Overflow in i128::abs in Decimal fits check (#26341)
- Make Expr.hash on Categorical mapping-independent (#26340)
- Clone shared GroupBy node before mutation in physical plan creation (#26327)
- Fix lazy evaluation of replace\_strict by making it fallible (#26267)
- Consider the "current location" of an item when computing `rolling_rank_by` (#26287)
- Reset `is_count_star` flag between queries in collect\_all (#26256)
- Fix incorrect is\_between filter on scan\_parquet (#26284)
- Lower AnonymousStreamingAgg in group-by as aggregate (#26258)
- Avoid overflow in `pl.duration` scalar arguments case (#26213)
- Broadcast arr.get on single array with multiple indices (#26219)
- Fix panic on CSPE with sorts (#26231)
- Fix UB in `DataFrame::transpose_from_dtype` (#26203)
- Eager `DataFrame.slice` with negative offset and `length=None` (#26215)
- Use correct schema side for streaming merge join lowering (#26218)
- Implement expression keys for merge-join (#26202)
- Overflow panic in `scan_csv` with multiple files and `skip_rows + n_rows` larger than total row count (#26128)
- Respect `allow_object` flag after cache (#26196)
- Raise error on non-elementwise PartitionBy keys (#26194)
- Allow ordered categorical dictionary in scan\_parquet (#26180)
- Allow excess bytes on IPC bitmap compressed length (#26176)
- Address buggy quadratic scaling fix in scan\_csv (#26175)
- Address a macOS-specific compile issue (#26172)
- Fix deadlock on `hash_rows()` of 0-width DataFrame (#26154)
- Fix NameError filtering pyarrow dataset (#26166)
- Fix concat\_arr panic when using categoricals/enums (#26146)
- Fix NDJSON/scan\_lines negative slice splitting with extremely long lines (#26132)
- Incorrect group\_by min/max fast path (#26139)
- Remove a source of non-determinism from lowering (#26137)
- Error when `with_row_index` or `unpivot` create duplicate columns on a `LazyFrame` (#26107)
- Panics on shift with head (#26099)
- Optimize slicing support on compressed IPC (#26071)
- CPU check for musl builds (#26076)
- Fix slicing on compressed IPC (#26066)
- Release GIL on collect\_batches (#26033)
- Missing buffer update in String is\_in Parquet pushdown (#26019)
- Make `struct.with_fields` data model coherent (#25610)
- Incorrect output order for order sensitive operations after join\_asof (#25990)
- Use SeriesExport for pyo3-polars FFI (#26000)
- Don't write Parquet min/max statistics for i128 (#25986)
- Ensure chunk consistency in in-memory join (#25979)
- Fix varying block metadata length in IPC reader (#25975)
- Implement collect\_batches properly in Rust (#25918)
- Fix panic on arithmetic with bools in list (#25898)
- Convert to index type with strict cast in some places (#25912)
- Empty dataframe in streaming non-strict hconcat (#25903)
- Infer large u64 in json as i128 (#25904)
- Set http client timeouts to 10 minutes (#25902)
- Prevent panic when comparing `Date` with `Duration` types (#25856)
- Correct lexicographic ordering for Parquet BYTE\_ARRAY statistics (#25886)
- Raise error on duplicate `group_by` names in `upsample()` (#25811)
- Correctly export view buffer sizes nested in Extension types (#25853)
- Fix `DataFrame.estimated_size` not handling overlapping chunks correctly (#25775)
- Ensure Kahan sum does not introduce NaN from infinities (#25850)
- Trim excess bytes in parquet decode (#25829)
- Reshape checks size to match exactly (#25571)
- Fix panic/deadlock sinking parquet with rows larger than 64MB estimated size (#25836)
- Fix quantile `midpoint` interpolation (#25824)
- Don't use cast when converting from physical in list.get (#25831)
- Invalid null count on int -> categorical cast (#25816)
- Update groups in `list.eval` (#25826)
- Use downcast before FFI conversion in PythonScan (#25815)
- Double-counting of row metrics (#25810)
- Cast nulls to expected type in streaming union node (#25802)
- Incorrect slice pushdown into map\_groups (#25809)
- Fix panic writing parquet with single bool column (#25807)
- Fix upsample with `group_by` incorrectly introduced NULLs on group key columns (#25794)
- Panic in top\_k pruning (#25798)
- Fix documentation for new() (#25791)
- Fix incorrect `collect_schema` for unpivot followed by join (#25782)
- Fix documentation for `tail()` (#25784)
- Verify `arr` namespace is called from array column (#25650)
- Ensure `LazyFrame.serialize()` unchanged after `collect_schema()` (#25780)
- Function map\_(rows|elements) with return\_dtype = pl.Object (#25753)
- Avoid visiting nodes multiple times in PhysicalPlanVisualizationDataGenerator (#25737)
- Fix incorrect cargo sub-feature (#25738)
- Fix deadlock on empty scan IR (#25716)
- Don't invalidate node in cluster-with-columns (#25714)
- Move `boto3` extra from s3fs in dev requirements (#25667)
- Binary slice methods missing from Series and docs (#25683)
- Mix-up of variable\_name/value\_name in unpivot (#25685)
- Invalid usage of `drop_first` in `to_dummies` when nulls present (#25435)
- Rechunk on nested dtypes in `take_unchecked_impl` parallel path (#25662)
- New single file IO sink pipeline enabled for sink\_parquet (#25670)
- Fix streaming `SchemaMismatch` panic on `list.drop_nulls` (#25661)
- Correct overly eager local predicate insertion for unpivot (#25644)
- Fix "dtype is unknown" panic in cross joins with literals (#25658)
- Fix panic on Boolean `rolling_sum` calculation for list or array eval (#25660)
- Preserve List inner dtype during chunked take operations (#25634)
- Fix `panic` edge-case when scanning hive partitioned data (#25656)
- Fix lifetime for `AmortSeries` lazy group iterator (#25620)
- Improve SQL `GROUP BY` and `ORDER BY` expression resolution, handling aliasing edge-cases (#25637)
- Fix empty format handling (#25638)
- Prevent false positives in is\_in for large integers (#25608)
- Optimize projection pushdown through HConcat (#25371)
- Differentiate between empty list an no list for unpivot (#25597)
- Properly resolve `HAVING` clause during SQL `GROUP BY` operations (#25615)
- Fix spearman panicking on nulls (#25619)
- Increase precision when constructing `float` Series (#25323)
- Make sum on strings error in group\_by context (#25456)
- Hang in multi-chunk DataFrame .rows() (#25582)
- Bug in boolean unique\_counts (#25587)
- Set `Float16` parquet schema type to `Float16` (#25578)
- Correct arr\_to\_any\_value for object arrays (#25581)
- Have `PySeries::new_f16` receive `pf16`s instead of `f32`s (#25579)
- Fix occurence of exact matches of `.join_asof(strategy="nearest", allow_exact_matches=False, ...)` (#25506)
- Raise error on out-of-range dates in temporal operations (#25471)
- Fix incorrect `.list.eval` after slicing operations (#25540)
- Reduce HuggingFace API calls (#25521)
- Strict conversion AnyValue to Struct (#25536)
- Fix panic in is\_between support in streaming Parquet predicate push down (#25476)
- Always respect return\_dtype in map\_elements and map\_rows (#25504)
- Rolling `mean`/`median` for temporals (#25512)
- Add `.rolling_rank()` support for temporal types and `pl.Boolean` (#25509)
- Fix dictionary replacement error in `write_ipc()` (#25497)
- Fix group lengths check in `sort_by` with `AggregatedScalar` (#25503)
- Fix expr slice pushdown causing shape error on literals (#25485)
- Allow empty list in `sort_by` in `list.eval` context (#25481)
- Prevent panic when joining sorted LazyFrame with itself (#25453)
- Apply CSV dict overrides by name only (#25436)
- Incorrect result in aggregated `first`/`last` with `ignore_nulls` (#25414)
- Fix off-by-one bug in `ColumnPredicates` generation for inequalities operating on integer columns (#25412)
- Fix `arr.{eval,agg}` in aggregation context (#25390)
- Support `AggregatedList` in `list.{eval,agg}` context (#25385)
- Improve SQL `UNNEST` behaviour (#22546)
- Remove `ClosableFile` (#25330)
- Use Cargo.template.toml to prevent git dependencies from using template (#25392)
- Resolve edge-case with SQL aggregates that have the same name as one of the `GROUP BY` keys (#25362)
- Revert `pl.format` behavior with nulls (#25370)
- Remove `Expr` casts in `pl.lit` invocations (#25373)
- Nested dtypes in streaming `first_non_null`/`last_non_null` (#25375)
- Correct `eq_missing` for struct with nulls (#25363)
- Unique on literal in aggregation context (#25359)
- Allow `implode` and aggregation in aggregation context (#25357)
- Aggregation with `drop_nulls` on literal (#25356)
- Address multiple issues with SQL `OVER` clause behaviour for window functions (#25249)
- Schema mismatch with `list.agg`, `unique` and scalar (#25348)
- Correct `drop_items` for scalar input (#25351)
- SQL `NATURAL` joins should coalesce the key columns (#25353)
- Mark `{forward,backward}_fill` as `length_preserving` (#25352)
- Nested dtypes in streaming `first`/`last` (#25298)
- AnyValue::to\_physical for categoricals (#25341)
- Fix link errors reported by `markdown-link-check` (#25314)
- Parquet `is_in` for mixed validity pages (#25313)
- Fix length preserving check for `eval` expressions in streaming engine (#25294)
- Fix building polars-plan with features lazy,concat\_str (but no strings) (#25306)
- Fix building polars-mem-engine with the async feature (#25300)
- Don't quietly allow unsupported SQL `SELECT` clauses (#25282)
- Fix small bug with `PyExpr` to `PyObject` conversion (#25265)
- Reverse on chunked `struct` (#25281)
- Panic exception when calling `Expr.rolling` in `.over` (#25283)
- Correct `{first,last}_non_null` if there are empty chunks (#25279)
- Incorrect results for aggregated `{n_,}unique` on bools (#25275)
- Fix building polars-expr without timezones feature (#25254)
- Ensure out-of-range integers and other edge case values don't give wrong results for index\_of() (#24369)
- Correctly prune projected columns in hints (#25250)
- Allow `Null` dtype values in `scatter` (#25245)
- Correct handle requested stops in streaming shift (#25239)
- Make `str.json_decode` output deterministic with lists (#25240)
- Wide-table join performance regression (#25222)
- Fix single-column CSV header duplication with leading empty lines (#25186)
- Enhanced column resolution/tracking through multi-way SQL joins (#25181)
- Fix serialization of lazyframes containing huge tables (#25190)
- Use (i64, u64) for VisualizationData (offset, length) slices (#25203)
- Fix assertion panic on `group_by` (#25179)
- Fix `format_str` in case of multiple chunks (#25162)
- Fix incorrect `drop_nans()` result when used in `group_by()` / `over()` (#25146)

## 📖 Documentation

- Fix typo in max\_by docstring (#26404)
- Remove deprecated `cublet_id` (#26260)
- Update for new release (#26255)
- Update MCP server section with new URL (#26241)
- Fix unmatched paren and punctuation in pandas migration guide (#26251)
- Add observatory database\_path to docs (#26201)
- Note plugins in Python user-defined functions (#26138)
- Clarify min\_by/max\_by behavior on ties (#26077)
- Add `QUALIFY` clause and `SUBSTRING` function to the SQL docs (#25779)
- Update mixed-offset datetime parsing example in user guide (#25915)
- Update bare-metal docs for mounted anonymous results (#25801)
- Fix credential parameter name in cloud-storage.py (#25788)
- Configuration options update (#25756)
- Fix typos in Excel and Pandas migration guides (#25709)
- Add "right" to `how` options in `join()` docstrings (#25678)
- Document schema parameter in meta methods (#25543)
- Correct link to `datetime_range` instead of `date_range` in resampling page (#25532)
- Explain aggregation \& sorting of lists (#25260)
- Update `LazyFrame.collect_schema()` docstring (#25508)
- Remove lzo from parquet write options (#25522)
- Update on-premise documentation (#25489)
- Fix incorrect 'bitwise' in `any_horizontal`/`all_horizontal` docstring (#25469)
- Add Extension and BaseExtension to doc index (#25444)
- Add polars-on-premise documentation (#25431)
- Fix link errors reported by `markdown-link-check` (#25314)
- Fix LanceDB URL (#25198)

## 📦 Build system

- Address remaining Python 3.14 issues with `make requirements-all` (#26195)
- Address a macOS-specific compile issue (#26172)
- Fix `make fmt` and `make lint` commands (#25200)

## 🛠️ Other improvements

- Move IO source metrics instrumentation to `PolarsObjectStore` (#26414)
- More SQL to IR conversion `execute_isolated` (#26455)
- Cleanup unused attributes in optimizer (#26464)
- Use `Expr::Display` as catch all for IR - DSL asymmetry (#26471)
- Remove the `POLARS_IDEAL_MORSEL_SIZE` monkeypatching in the parametric merge-join test (#26418)
- Move IO metrics struct to `polars-io` and use new timer (#26397)
- Reduce blocking on computational executor threads in multiscan init (#26407)
- Cleanup the parametric merge-join test (#26413)
- Ensure local doctests skip `from_torch` if module not installed (#26405)
- Implement various deprecations (#26314)
- Refactor `MinBy` and `MaxBy` as `IRFunctions` (#26307)
- Rename `Operator::Divide` to `RustDivide` (#26339)
- Properly disable the Pyodide tests (#26382)
- Add LiveTimer (#26384)
- Use derived serialization on `PlRefPath` (#26167)
- Add metadata to `ArrowSchema` struct (#26318)
- Remove unused field (#26367)
- Fix runtime nesting (#26359)
- Remove `xlsx2csv` dependency pin (#26355)
- Allow unchecked IPC reads (#26354)
- Use outer runtime if exists in to\_alp (#26353)
- Make CategoricalMapping::new pub(crate) to avoid misuse (#26308)
- Clarify IPC buffer read limit/length paramter (#26334)
- Improve accuracy of active IO time metric (#26315)
- Mark `VarState` as `repr(C)` (#26309)
- IO metrics for streaming Parquet / IPC sources (#26300)
- Replace panicking index access with error handling in `dictionaries_to_encode` (#26059)
- Remove unnecessary match and move early return in testing (#26297)
- Add dtype test coverage for delta predicate filter (#26291)
- Add property-based tests for `Scalar::cast_with_options` (#25744)
- Add AI policy (#26286)
- Remove MemSlice (#26259)
- Remove recursion from `upsample_impl` (#26250)
- Remove all non CSV fast-count paths (#26233)
- Replace MemReader with Cursor (#26216)
- Add `serde(default)` to new CSV compression fields (#26210)
- Add a couple of `SAFETY` comments in merge-join node (#26197)
- Expose physical plan `NodeStyle` (#26184)
- Ensure optimization flag modification happens local (#26185)
- Use NullChunked as default for Series (#26181)
- In merge-sorted node, when buffering, request a stop on \*every\* unbuffered morsel (#26178)
- Rename `io_sinks2 -> io_sinks` (#26159)
- Lint leftover fixme (#26122)
- Move Buffer and SharedStorage to polars-buffer crate (#26113)
- Remove old sink IR (#26130)
- Use derived serialization on `PlRefPath` (#26062)
- Improve backtrace for `POLARS_PANIC_ON_ERR` (#26125)
- Fix Python docs build (#26117)
- Remove old streaming sink implementation (#26102)
- Disable `unused-ignore` mypy lint (#26110)
- Remove unused equality impls for IR / FunctionIR (#26106)
- Ignore mypy warning (#26105)
- Preserve order for string concatenation (#26101)
- Raise error on `file://hostname/path` (#26061)
- Disable debug info for docs workflow (#26086)
- Remove IR / physical plan visualization data generators (#26090)
- Update docs for next polars cloud release (#26091)
- Support Python 3.14 in dev environment (#26073)
- Mark top slow normal tests as slow (#26080)
- Simplify `PlPath` (#26053)
- Update breaking deps (#26055)
- Fix for upstream url bug and update deps (#26052)
- Properly pin chrono (#26051)
- Don't run rust doctests (#26046)
- Update deps (#26042)
- Ignore very slow test (#26041)
- Add Send bound for SharedStorage owner (#26040)
- Update rust compiler (#26017)
- Improve csv test coverage (#25980)
- Use `from_any_values_and_dtype` in `Series::extend_constant` (#26006)
- Pass `sync_on_close` and `num_pipelines` via `start_file_writer` for IO sinks (#25950)
- Add `broadcast_nulls` field to `RowEncodingVariant` and `_get_rows_encoded_{ca,arr}` (#26001)
- Ramp up CSV read size (#25997)
- Rename `FileType` to `FileWriteFormat` (#25951)
- Don't unwrap first sink morsel send (#25981)
- Update `ruff` action and simplify version handling (#25940)
- Cleanup Rust DataFrame interface (#25976)
- Export PhysNode related struct (#25987)
- Restructure Sort variant in logical and physical plans visualization data (#25978)
- Run python lint target as part of pre-commit (#25982)
- Allow multiple inputs to streaming GroupBy node (#25961)
- Disable HTTP timeout for receiving response body (#25970)
- Add AI contribution policy (#25956)
- Remove unused sink code (#25949)
- Add detailed Sink info to IRNodeProperties (#25954)
- Wrap `FileScanIR::Csv` enum variant in `Arc` (#25952)
- Use PlSmallStr for CSV format strings (#25901)
- Add unsafe bound to MemSlice::from\_arc (#25920)
- Improve MemSlice Debug impl (#25913)
- Remove manual cmp impls for `&[u8]` (#25890)
- Remove and deprecate batched csv reader (#25884)
- Remove unused AnonymousScan functions (#25872)
- Use Buffer\<T> instead of Arc\<[T]> to store stringview buffers (#25870)
- Add `TakeableRowsProvider` for IO sinks (#25858)
- Filter DeprecationWarning from pyparsing indirectly through pyiceberg (#25854)
- Various small improvements (#25835)
- Clear venv with appropriate version of Python (#25851)
- Move CSV write logic to `CsvSerializer` (#25828)
- Ensure Polars Object extension type is registered (#25813)
- Harden Python object process ID (#25812)
- Skip schema inference if schema provided for `scan_csv/ndjson` (#25757)
- Ensure proper async connection cleanup on DB test exit (#25766)
- Flip `has_residual_predicate` -> `no_residual_predicate` (#25755)
- Track original length before file filtering in scan IR (#25717)
- Ensure we uninstall other Polars runtimes in CI (#25739)
- Make 'make requirements' more robust (#25693)
- Remove duplicate compression level types (#25723)
- Replace async blocks with named components in new parquet write pipeline (#25695)
- Move Object `lit` fix earlier in the function (#25713)
- Remove unused decimal file (#25701)
- Move `boto3` extra from s3fs in dev requirements (#25667)
- Upgrade to latest version of `sqlparser-rs` (#25673)
- Update slab to version without RUSTSEC (#25686)
- Fix typo (#25684)
- Avoid rechunk requirement for Series.iter() (#25603)
- Use dtype for group\_aware evaluation on `ApplyExpr` (#25639)
- Make polars-plan constants more consistent (#25645)
- Add "panic" and "streaming" tagging to `issue-labeler` workflow (#25657)
- Add support for multi-column reductions (#25640)
- Fix rolling kernel dispatch with `monotonic` group attribute (#25494)
- Simplify \_write\_any\_value (#25622)
- Ensure we hash all attributes and visit all children in `traverse_and_hash_aexpr` (#25627)
- Ensure literal-only `SELECT` broadcast conforms to SQL semantics (#25633)
- Add parquet file write pipeline for new IO sinks (#25618)
- Rename polars-on-premise to polars-on-premises (#25617)
- Constrain new `issue-labeler` workflow to the Issue title (#25614)
- Add streaming IO sink components (#25594)
- Help categorise Issues by automatically applying labels (using the same patterns  used for labelling PRs) (#25599)
- Show on streaming engine (#25589)
- Add `arg_sort()` and `Writeable::as_buffered()` (#25583)
- Take task priority argument in `parallelize_first_to_local` (#25563)
- Skip existing files in pypi upload (#25576)
- Fix template path in release-python workflow (#25565)
- Skip rust integration tests for coverage in CI (#25558)
- Add asserts and tests for `list.eval` on multiple chunks with slicing (#25559)
- Rename `URL_ENCODE_CHARSET` to `HIVE_ENCODE_CHARSET` (#25554)
- Add `assert_sql_matches` coverage for SQL `DISTINCT` and `DISTINCT ON` syntax (#25440)
- Use strong hash instead of traversal for CSPE equality (#25537)
- Update partitioned sink IR (#25524)
- Print expected DSL schema hashes if mismatched (#25526)
- Remove verbose prints on file opens (#25523)
- Add `proptest` `AnyValue` strategies (#25510)
- Fix --uv argument for benchmark-remote (#25513)
- Add `proptest` `DataFrame` strategy (#25446)
- Run `maturin` with `--uv` option (#25490)
- Remove some dead argminmax impl code (#25501)
- Fix feature gating TZ\_AWARE\_RE again (#25493)
- Take `sync` parameter in `Writeable::close()` (#25475)
- Fix unsoundness in ChunkedArray::{first, last} (#25449)
- Add some cleanup (#25445)
- Test for `group_by(...).having(...)` (#25430)
- Accept multiple files in `pipe_with_schema` (#25388)
- Remove aggregation context `Context` (#25424)
- Take `&dyn Any` instead of `Box<dyn Any>` in python object converters (#25421)
- Refactor sink IR (#25308)
- Remove `ClosableFile` (#25330)
- Remove debug file write from test suite (#25393)
- Add `ElementExpr` for `_eval` expressions (#25199)
- Dispatch `Series.set` to `zip_with_same_dtype` (#25327)
- Better coverage for `group_by` aggregations (#25290)
- Add oneshot channel to polars-stream (#25378)
- Enable more streaming tests (#25364)
- Remove `Column::Partitioned` (#25324)
- Remove incorrect cast in reduce code (#25321)
- Add toolchain file to runtimes for sdist (#25311)
- Remove `PyPartitioning` (#25303)
- Directly take `CloudScheme` in `parse_cloud_options()` (#25304)
- Refactor `dt_range` functions (#25225)
- Fix typo in CI release workflow (#25309)
- Use dedicated runtime packages from template (#25284)
- Add `proptest` strategies for Series nested types (#25220)
- Simplify sink parameter passing from Python (#25302)
- Add test for unique with column subset (#25241)
- Fix Decimal precision annotation (#25227)
- Add `LazyFrame.pivot` (#25016)
- Clean up CSPE callsite (#25215)
- Avoid relabelling changes-dsl on every commit (#25216)
- Move ewm variance code to polars-compute (#25188)
- Upgrade to schemars 0.9.0 (#25158)
- Update markdown link checker (#25201)
- Automatically label pull requests that change the DSL (#25177)
- Add reliable test for `pl.format` on multiple chunks (#25164)
- Move supertype determination and casting to IR for `date_range` and related functions (#24084)
- Make python docs build again (#25165)
- Make `pipe_with_schema` work on Arced schema (#25155)
- Add functions for `scan_lines` (#25136)
- Remove lower\_ir conversion from Scan to InMemorySource (#25150)
- Update versions (#25141)

Thank you to all our contributors for making this release possible!
@AndreaBozzo, @Atarust, @DannyStoll1, @EndPositive, @JakubValtar, @Jesse-Bakker, @Kevin-Patyk, @LeeviLindgren, @MarcoGorelli, @Matt711, @MrAttoAttoAtto, @TNieuwdorp, @Voultapher, @WaffleLapkin, @agossard, @alex-gregory-ds, @alexander-beedie, @anosrepenilno, @arlyon, @azimafroozeh, @bayoumi17m, @borchero, @c-peters, @cBournhonesque, @camriddell, @carnarez, @cmdlineluser, @coastalwhite, @cr7pt0gr4ph7, @davanstrien, @davidia, @dependabot[bot], @dsprenkels, @edizeqiri, @eitanf, @etiennebacher, @feliblo, @gab23r, @guilhem-dvr, @hallmason17, @hamdanal, @henryharbeck, @hutch3232, @ion-elgreco, @itamarst, @jamesfricker, @jannickj, @jetuk, @jqnatividad, @kdn36, @lun3x, @m1guelperez, @marinegor, @mcrumiller, @nameexhaustion, @orlp, @pomo-mondreganto, @qxzcode, @r-brink, @ritchie46, @sachinn854, @stijnherfst, @sweb, @tlauli, @vyasr, @wtn, @yonikremer and [dependabot[bot]](https://github.com/apps/dependabot)