1 Release published by 1 person

• py-1.32.0-beta.1 Python Polars 1.32.0-beta.1

  published Jul 26, 2025

55 Pull requests merged by 11 people

• fix: Make meta.pop operate on Expr only

  #23808 merged Jul 30, 2025

• perf: Lower pl.repeat to streaming engine

  #23804 merged Jul 29, 2025

• feat: Implement repeat_by for Array and Null

  #23794 merged Jul 29, 2025

• chore: Fix typos path

  #23803 merged Jul 29, 2025

• chore: Remove deserialize_with_unknown_fields

  #23802 merged Jul 29, 2025

• feat: Add DeprecationWarning on passing physical ordering to Categorical

  #23779 merged Jul 29, 2025

• test(python): Add pyiceberg to dev dependencies

  #23791 merged Jul 29, 2025

• refactor(rust): Remove incorrect DeletionFilesList::slice

  #23796 merged Jul 29, 2025

• feat: Pre-filtered decode and row group skipping with Iceberg / Delta / scans with cast options

  #23792 merged Jul 29, 2025

• fix: Stack overflow in DslPlan serde

  #23801 merged Jul 29, 2025

• chore: Remove old schema file

  #23798 merged Jul 29, 2025

• refactor(rust): Remove Default for StreamingExecutionState

  #23729 merged Jul 28, 2025

• fix: Clear credentials cached in Python when rebuilding object store

  #23756 merged Jul 28, 2025

• perf: Remove redundant clone

  #23771 merged Jul 28, 2025

• feat(python): Update BytecodeParser opcode awareness for upcoming Python 3.14

  #23782 merged Jul 28, 2025

• refactor(rust): Explicit match to smaller dtypes before cast to Int32 in asof join

  #23776 merged Jul 28, 2025

• fix: Datetime selectors with mixed timezone info

  #23774 merged Jul 28, 2025

• chore: Mark more tests as ready for cloud

  #23743 merged Jul 28, 2025

• docs: Correct the name.* methods on their removal of aliases

  #23773 merged Jul 28, 2025

• fix: Support i128 in asof join

  #23770 merged Jul 28, 2025

• fix: Remove sleep for credential refresh

  #23768 merged Jul 27, 2025

• chore: Reduce required deps for pyo3-polars

  #23761 merged Jul 27, 2025

• refactor(rust): Expose PlPathRef via polars::prelude

  #23754 merged Jul 27, 2025

• refactor(python): Remove unused functions from the rust side

  #23757 merged Jul 27, 2025

• ci: Add hashes json

  #23758 merged Jul 26, 2025

• Python Polars 1.32 pre-release

  #23750 merged Jul 26, 2025

• fix(python): Load _expiry_time from botocore Credentials in CredentialProviderAWS

  #23753 merged Jul 26, 2025

• feat(python): Add Python-side caching for credentials and provider auto-initialization

  #23736 merged Jul 25, 2025

• fix: Fix credential refresh logic

  #23730 merged Jul 25, 2025

• chore: Add AExpr::is_expr_equal_to

  #23740 merged Jul 25, 2025

• fix: Fix to_datetime() fallible identification

  #23735 merged Jul 25, 2025

• fix: Correct output datatype for dt.with_time_unit

  #23734 merged Jul 25, 2025

• fix: Fix incorrect native Iceberg scan from tables with renamed/dropped columns/fields

  #23713 merged Jul 25, 2025

• chore: Fix rank test to respect maintain order

  #23723 merged Jul 24, 2025

• refactor: IR inputs and exprs iterators

  #23722 merged Jul 24, 2025

• fix: Allow DataType expressions with selectors

  #23720 merged Jul 24, 2025

• feat: Expand on DataTypeExpr

  #23249 merged Jul 24, 2025

• refactor: Store more granular schema hashes to reduce merge conflicts

  #23709 merged Jul 24, 2025

• refactor(rust): Add assertions for unique ID

  #23711 merged Jul 24, 2025

• fix: Match output type to engine for interpolate on Decimal

  #23706 merged Jul 24, 2025

• fix: Remaining bugs in with_exprs_and_input and pruning

  #23710 merged Jul 24, 2025

• refactor(rust): Use RelaxedCell in multiscan

  #23712 merged Jul 24, 2025

• refactor(rust): Debug assert ColumnTransform cast is non-strict

  #23717 merged Jul 24, 2025

• refactor: Use UUID for UniqueID

  #23704 merged Jul 23, 2025

• fix: Match output dtype to engine for cum_sum_horizontal

  #23686 merged Jul 23, 2025

• fix: Field names for pl.struct in group-by

  #23703 merged Jul 23, 2025

• docs(python): Fix str.replace_many examples trigger deprecation warning

  #23695 merged Jul 23, 2025

• fix: Fix output for str.extract_groups with empty string pattern

  #23698 merged Jul 23, 2025

• fix: Match output type to engine for rolling_map

  #23702 merged Jul 23, 2025

• refactor: Remove scan id

  #23697 merged Jul 23, 2025

• refactor: Propagate Iceberg physical ID schema to IR

  #23671 merged Jul 23, 2025

• chore: Remove unused and confusing match arm

  #23691 merged Jul 23, 2025

• fix(python): Moved passing DeltaTable._storage_options

  #23673 merged Jul 23, 2025

• chore: Remove unused ALLOW_GROUP_AWARE flag

  #23690 merged Jul 23, 2025

• fix: Fix incorrect join on single Int128 column for in-memory engine

  #23694 merged Jul 23, 2025

24 Pull requests opened by 10 people

• refactor: Rewrite `evaluate_on_groups` for `.gather` / `.get`

  #23700 opened Jul 23, 2025

• docs: Update user guide aggregation

  #23718 opened Jul 24, 2025

• refactor(rust): Add subphase and query task spawning to StreamingExecState

  #23725 opened Jul 24, 2025

• docs: Improve examples aggregation

  #23726 opened Jul 24, 2025

• feat(python): Making `Expr.rolling_*_by` methods available to `pl.Series`

  #23742 opened Jul 25, 2025

• fix(rust): Fix schema consistency for `Categorical` in when-then expressions

  #23762 opened Jul 27, 2025

• refactor(rust): Replace `Either` with `SchemaProvider` for PythonScan DSL

  #23775 opened Jul 28, 2025

• docs(python): Update resampling.md

  #23778 opened Jul 28, 2025

• feat: Implement temporal dtypes for `hist`

  #23786 opened Jul 28, 2025

• docs: Document the pureness requirements of udfs

  #23787 opened Jul 29, 2025

• feat(python): Add caching to user credential providers

  #23789 opened Jul 29, 2025

• fix: Set `shrink_dtype` output type to `Unknown`

  #23793 opened Jul 29, 2025

• refactor(rust): Make `GetBatchState` polling functions unsafe

  #23795 opened Jul 29, 2025

• fix: Match output type to engine for `Struct` arithmetic

  #23805 opened Jul 29, 2025

• perf(python): Optimise `BytecodeParser` usage from `warn_on_inefficient_map`

  #23809 opened Jul 30, 2025

• build(python): Workaround for pyiceberg `make requirements` on Python 3.13

  #23810 opened Jul 30, 2025

• feat: Default to native reader for `scan_iceberg`

  #23811 opened Jul 30, 2025

• feat: Pass `endpoint_url` loaded from `CredentialProviderAWS` to `scan/write_delta`

  #23812 opened Jul 30, 2025

• refactor(python): Improve internal `LRUCache` implementation and move into "_utils" module

  #23813 opened Jul 30, 2025

• chore: Deprecate `Expr.shrink_dtype` and remove from lazy execution

  #23814 opened Jul 30, 2025

• fix!: Make `Expr.list.to_struct` respect the Polars data-model

  #23815 opened Jul 30, 2025

• ci: Automatically tag patches mentioning "SQL" with the appropriate label

  #23816 opened Jul 30, 2025

• feat(python): Ensure `CachingCredentialProvider` returns copied credentials dict

  #23817 opened Jul 30, 2025

• chore: Update `typos` package

  #23818 opened Jul 30, 2025

34 Issues closed by 10 people

• Plotters with Polars

  #23807 closed Jul 30, 2025

• Regression in 1.32.0-beta.1, `meta.pop` fails for `map_batches`

  #23760 closed Jul 30, 2025

• Reading parquet files with metadata footer larger than 2GB aborts

  #23162 closed Jul 30, 2025

• `repeat_by` breaks on array entries

  #23788 closed Jul 29, 2025

• `pl.Categorical(ordering='physical')` has no warnings or errors but isn't supported

  #23766 closed Jul 29, 2025

• Schema mismatch panic in `scan_parquet` with `n_rows` and `filter`

  #23790 closed Jul 29, 2025

• Scan parquet fails when shrinking `dtype` as part of the lazy operations

  #15728 closed Jul 29, 2025

• Regression (0.20.2 -> 0.20.3): failure to initialize null column with nested struct dtypes (`ComputeError`)

  #13602 closed Jul 28, 2025

• `scan_csv` does not parse the time column correctly - while `read_csv` does

  #14038 closed Jul 28, 2025

• Supertype of datetime/durations is the lower-precision value

  #23783 closed Jul 28, 2025

• Multi-output, multi-sink lazy polars

  #18235 closed Jul 28, 2025

• sink_parquet_all support

  #18364 closed Jul 28, 2025

• `scan_parquet` + `sink_parquet` with same filename PanicException and file truncation

  #12843 closed Jul 28, 2025

• scan_iceberg: can't pass cast_options

  #23781 closed Jul 28, 2025

• Add a `print` or similar method which accepts pl.Config arguments

  #23705 closed Jul 28, 2025

• `pl.selectors.datetime(time_zone=[None, 'UTC'])` no longer selects timezone-naive columns

  #23767 closed Jul 28, 2025

• [pre-release issue] `.name.suffix` no longer undoes `alias`

  #23765 closed Jul 28, 2025

• Incorrect results with `join_asof` with `by` parameter

  #23751 closed Jul 28, 2025

• Add ability to free unused memory

  #23755 closed Jul 27, 2025

• sink_parquet() with float sign() column fails

  #23759 closed Jul 27, 2025

• Floating point arithmetic errors up to 1ULP with respect to IEEE 754

  #23741 closed Jul 25, 2025

• Cannot select `List` or `Array` types by dtype using `pl.List` or `pl.Array` without specifying dtype details

  #23749 closed Jul 25, 2025

• Slow `CredentialProviderAWS` due to `botocore` calling `ProcessProvider`

  #23732 closed Jul 25, 2025

• `__exit__` fails in `tempfile` when reading with `pl.read_csv`

  #23460 closed Jul 25, 2025

• Incorrect results on native Iceberg scans when columns have been renamed

  #23428 closed Jul 25, 2025

• Gotcha: `pl.read_excel`, using default calamine engine, has a low default `schema_sample_rows` and `dtype_coercion` is non-strict by default

  #23728 closed Jul 24, 2025

• `cum_sum_horizontal` InvalidOperationError on main

  #23719 closed Jul 24, 2025

• Unexpected error on `unnest` of horizontal expression for LazyFrame

  #22076 closed Jul 24, 2025

• `pl.cum_sum_horizontal(...).struct.unnest()` errors with `StructFieldNotFoundError: literal`

  #19765 closed Jul 24, 2025

• bug: pl.cum_sum_horizontal(pl.all()).struct.unnest() raises

  #20277 closed Jul 24, 2025

• Usage of `UniqueId::default()` is not sound when there are `Plain` variants in memory

  #23674 closed Jul 23, 2025

• Expression name of nested struct expression is silently dropped during collection

  #23701 closed Jul 23, 2025

• `str.replace_many` example triggers deprecation warning

  #23517 closed Jul 23, 2025

• Incorrect results when joining between u64 and i64 (i.e. joining on i128)

  #23688 closed Jul 23, 2025

22 Issues opened by 16 people

• RFC: `schema_provider` + `schema_sampler` option for `scan_parquet` to allow for programmatic/custom schema management

  #23806 opened Jul 29, 2025

• `int_range` as a `join_where` predicate incorrect result

  #23800 opened Jul 29, 2025

• Arithmetic on struct does not cast inner to supertype

  #23797 opened Jul 29, 2025

• `pl.Series` should inherit from `Sequence`

  #23785 opened Jul 28, 2025

• Regression in 1.32.0-beta.1, function passed to `map_batches` called an additional time

  #23784 opened Jul 28, 2025

• LazyFrame from scan_delta - leveraging file level statistics for predicate pushdown

  #23780 opened Jul 28, 2025

• User Guide still mentions group_by_rolling / example rolling missing

  #23777 opened Jul 28, 2025

• Decimal zeroes are stripped when reading float values as string in read_json

  #23772 opened Jul 28, 2025

• SEGFAULT When adding as a dependency to a PyO3 project

  #23769 opened Jul 27, 2025

• Improve readability of show_graph / explain output

  #23764 opened Jul 27, 2025

• User Guide datasets are scattered and/or missing

  #23763 opened Jul 27, 2025

• Provide some sort of callback for `LazyFrame` to handle errors when collecting?

  #23752 opened Jul 25, 2025

• Output type mismatch in `when-then`

  #23733 opened Jul 25, 2025

• Column-wise type cast failed when using pl.when

  #23731 opened Jul 25, 2025

• Polars IO plugin full LazyFrame push down

  #23727 opened Jul 24, 2025

• Memory-efficient streaming: group_by max-by aggregation

  #23721 opened Jul 24, 2025

• Add `sink_iceberg`

  #23716 opened Jul 24, 2025

• Iceberg: Outer NULLs are written as empty lists on columns with multiple nesting levels

  #23715 opened Jul 24, 2025

• Allow constructing `Schema` from a pyrrow schema using `pl.Schema(pyarrow_schema)`

  #23714 opened Jul 24, 2025

• `pl.Expr.meta.has_multiple_outputs()` is broken on main

  #23708 opened Jul 23, 2025

• Add global setting for infer_schema_length

  #23707 opened Jul 23, 2025

• LazyFrame join -> over(mapping_strategy="join") -> collect(engine="streaming") does not preserve list order.

  #23699 opened Jul 23, 2025

46 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• docs: Update multiple pages in the Polars Cloud user guide

  #23661 commented on Jul 30, 2025 • 10 new comments

• feat: Add `having` to `group_by` context

  #23550 commented on Jul 24, 2025 • 1 new comment

• Add lance as a supported data format to read/write

  #14452 commented on Jul 28, 2025 • 0 new comments

• Group by and selectors not working with head()

  #14383 commented on Jul 28, 2025 • 0 new comments

• Inconsistent behavior with schema_overrides in read_csv

  #20903 commented on Jul 28, 2025 • 0 new comments

• pl.concat_list(pl.col("struct_of_structs")) out of memory and crashes

  #19805 commented on Jul 28, 2025 • 0 new comments

• Autodetect hive partitioning when reading from S3

  #23591 commented on Jul 28, 2025 • 0 new comments

• `list.eval(pl.element().fill_null())` doesn't work on empty DataFrames

  #23294 commented on Jul 28, 2025 • 0 new comments

• `cannot get ref Float64 from Float32` when dividing `Float32` by an `Int32`

  #23677 commented on Jul 29, 2025 • 0 new comments

• Polars scan_iceberg inefficiently loads Iceberg table data into memory leading to out-of-memory errors

  #18912 commented on Jul 29, 2025 • 0 new comments

• Tracking for Iceberg cloud enablement

  #22450 commented on Jul 29, 2025 • 0 new comments

• Expression output field must match engine output

  #23663 commented on Jul 30, 2025 • 0 new comments

• Tracking Issue for dealing with Unknown datatype

  #23625 commented on Jul 30, 2025 • 0 new comments

• feat: Change SQL-Explode/UNNEST to Dataframe.explode method

  #22546 commented on Jul 25, 2025 • 0 new comments

• feat: Reinterpret binary data to fixed size numerical array

  #22840 commented on Jul 25, 2025 • 0 new comments

• feat: Experimental support for sinking to Iceberg

  #23011 commented on Jul 24, 2025 • 0 new comments

• feat: Implement `dt.days_in_month` function

  #23119 commented on Jul 28, 2025 • 0 new comments

• fix: Improve identification of "non group-key" aggregates in SQL `GROUP BY` queries

  #23191 commented on Jul 29, 2025 • 0 new comments

• feat: Add `categories` to `to_dummies`

  #23270 commented on Jul 28, 2025 • 0 new comments

• feat: Add `list.zip`

  #23316 commented on Jul 28, 2025 • 0 new comments

• feat: Add `num_samples` parameter to `date_range` and `datetime_range`

  #23425 commented on Jul 23, 2025 • 0 new comments

• feat: Support `pre_execution_query` parameter from `connectorx`

  #23634 commented on Jul 27, 2025 • 0 new comments

• fix: Create file for streaming sink even if unspawned

  #23672 commented on Jul 23, 2025 • 0 new comments

• Free RAM not released to OS after heavy dataframe operations

  #23128 commented on Jul 23, 2025 • 0 new comments

• `jemalloc` future/replacement?

  #23222 commented on Jul 23, 2025 • 0 new comments

• Group-By + filter aggregation performs subsequent operations on all data instead of only filtered data

  #23682 commented on Jul 23, 2025 • 0 new comments

• Tracking issue for the new streaming engine

  #20947 commented on Jul 24, 2025 • 0 new comments

• `scan_iceberg` does not push predicate filters into the table scan.

  #23286 commented on Jul 24, 2025 • 0 new comments

• Add `df.write_iceberg(table, mode="upsert")` feature (or maybe `df.upsert_iceberg()`)

  #23197 commented on Jul 24, 2025 • 0 new comments

• Lazy/streaming writes to iceberg tables

  #22336 commented on Jul 24, 2025 • 0 new comments

• Bug: engine="streaming" parameter not implemented when reading Iceberg tables from S3

  #22473 commented on Jul 24, 2025 • 0 new comments

• Significant performance difference depending on how I use the "filter" method

  #18531 commented on Jul 24, 2025 • 0 new comments

• Element-wise casting before SUM with OVER clause results in PanicException

  #23283 commented on Jul 24, 2025 • 0 new comments

• Partition filtering on Iceberg no longer working

  #22978 commented on Jul 24, 2025 • 0 new comments

• Typing error with SchemaDict

  #14468 commented on Jul 24, 2025 • 0 new comments

• Add `pl.case_when()` functionality

  #22939 commented on Jul 24, 2025 • 0 new comments

• Duration `std` aggregation is not implemented for new streaming engine.

  #23608 commented on Jul 25, 2025 • 0 new comments

• `Expr.rolling_*_by` exist but these method are not on `Series`

  #22066 commented on Jul 25, 2025 • 0 new comments

• Tracking issue for missing methods in `arr` namespace

  #21302 commented on Jul 25, 2025 • 0 new comments

• Support datetime ceiling and flooring

  #22204 commented on Jul 25, 2025 • 0 new comments

• Fixed-width text file reader

  #3151 commented on Jul 25, 2025 • 0 new comments

• Add SAS database read support (.sas7bdat) as pl.read_sas(filepath:str, **kwargs)

  #14587 commented on Jul 25, 2025 • 0 new comments

• Have convert_dtypes() function for dataframes

  #23180 commented on Jul 26, 2025 • 0 new comments

• Tracking issue for Polars Cloud

  #21487 commented on Jul 26, 2025 • 0 new comments

• `polars.DataFrame.write_json` panics for `pl.List(pl.Enum)` nested column

  #23459 commented on Jul 28, 2025 • 0 new comments

• Add support to read and write Amazon ION files

  #12069 commented on Jul 28, 2025 • 0 new comments