py-1.33.1
📦 polarsView on GitHub →
✨ 7 features🐛 13 fixes🔧 8 symbols
Summary
This release focuses heavily on performance improvements, particularly around Parquet decoding and memory allocation. It also introduces several enhancements, including S3 URI support and a new security policy, alongside numerous bug fixes.
✨ New Features
- Support S3 virtual-hosted–style URI
- Remove explicit file create for local async writes
- Add PyCapsule `__arrow_c_schema__` interface to `pl.Schema`
- Support Partitioning sinks in cloud
- User-friendly error message on empty path expansion
- Add unstable `pre_execution_query` parameter to `read_database_uri`
- Add Polars security policy
🐛 Bug Fixes
- Correct `sink_ipc` overload for compression
- Enable all integer dtypes for `by` parameter in `join_asof`
- Fix Group-By + filter aggregation performs subsequent operations on all data instead of only filtered data
- Wrap deprecated top-level imports in TYPE_CHECKING
- Fix incorrect output ordering for row-separable exprs
- Fix `Series.__arrow_c_stream__` for Decimal and other logical types
- Match output type to engine for `Struct` arithmetic
- Make mmap use MAP_PRIVATE rather than MAP_SHARED
- Fix cloud iceberg scan DATASET_PROVIDER_VTABLE error
- Don't throw away type information for NumPy numeric values when using lit()
- Incorrect logic in negative streaming slice
- Ensure `read_database_uri` with ADBC works as expected with DuckDB URIs
- Do not error on non-list `Sequence` for `columns` parameter in `read_excel`
🔧 Affected Symbols
pl.Schemasink_ipcjoin_asofSeries.__arrow_c_stream__lit()read_database_uriread_excelregister_io_source