File python-narwhals.changes of Package python-narwhals
-------------------------------------------------------------------
Mon Mar 9 10:44:31 UTC 2026 - Dirk Müller <dmueller@suse.com>
- update to 2.17.0:
* feat: Add support for indices and values `Series` in
`Series.scatter`
* fix: Allow `Decimal` dtypes in `{Series, Expr}.is_close`
* fix: `offset_by` was raising unnecessarily for sql-like
* fix: add test for PySpark min supported version in CI and fix
broken methods
* fix: Address DuckDB 1.5 `fetch_arrow_table` deprecation
- update to 2.16.0:
* feat: add `separator` argument in `read_csv`/`scan_csv`
* feat: Allow nested structures in `lit`
* enh: Introduce `narwhals.sql`
* enh: Introduce (optional) `order_by` in `first` / `last`
* feat: support window functions in filter
* feat: Improve support for `Decimal` DType
* feat: Support `concat(..., how="diagonal")` for `ibis`
* feat: Enable`list.{sort, sum}` for sqlframe
* feat: Add `str.pad_{start,end}`
* feat: Add `{Expr,Series}.cos`
* feat: Add `testing.assert_frame_equal`
-------------------------------------------------------------------
Mon Jan 26 20:24:35 UTC 2026 - Dirk Stoecker <opensuse@dstoecker.de>
- update to 2.15.0
## Changes
* test: unxfail sqlframe tests for list functions (#3383)
* ci: Test fairlearn using pytest marker (#3234)
## Enhancements
* test: unxfail sqlframe for list.median (#3387)
* feat: Add {Expr,Series}.sin (#3365)
* feat: add list.sort (#3356)
## Bug fixes
* test: Various GPU fixes (#3390)
* test: separate numpy array for tests (#3385)
* fix(docs): Keep table filter only in api-completeness page (#3367)
* ci: Fix hierarchicalforecast installation (#3362)
## Documentation
* fix(docs): Keep table filter only in api-completeness page (#3367)
## Other improvements
* chore: Refactor pandas-like pyarrow branching (#3361)
-------------------------------------------------------------------
Sat Dec 20 17:51:40 UTC 2025 - Ben Greiner <code@bnavigator.de>
- Update to 2.14.0
## Enhancements
* feat: add list aggregate methods (#3332)
* feat: Add {Expr,Series}.any_value (#3315)
* enh: Add support for slice(None) in series getitem (#3342)
## Bug fixes
* test: fix median tests for list.agg methods (#3354)
* fix: ArrowSeries.fill_null(strategy=..., limit=...) and None at
the "edge" (#3335)
- Lots of other changes since 1.46, see Releases page on GitHub
https://github.com/narwhals-dev/narwhals/releases
-------------------------------------------------------------------
Mon Jul 14 12:28:34 UTC 2025 - Dirk Müller <dmueller@suse.com>
- update to 1.46.0:
* refactor: Cache, reuse `Implementation._backend_version`
* enh!: simplify exceptions, consistently raise
InvalidOperationError instead of
ShapeError/OrderDependentExprError/LengthChangingExprError
* feat: Add `Expr.dt.offset_by()`
* enh: allow for `over(order_by=...)` for `rank`
* fix: Respect session time zone when casting to `Datetime` for
PySpark
* enh: Support casting to Duration, and Datetime with different
units, for DuckDB
* feat: Add support for `quantile` and `ewm_mean` in over
context
* feat: Support `DataFrame().lazy("ibis")`
* chore: xfail cudf offset_by
* ci: cudf tests fixup
* ci: unxfail truncate tests for pyspark by checking local
results
* fix: type annotation default for `named: Literal[False]` was
missing in `DataFrame.iter_rows`
* refactor: Cache, reuse `Implementation._backend_version`
-------------------------------------------------------------------
Sat Jul 5 10:25:21 UTC 2025 - Dirk Müller <dmueller@suse.com>
- update to 1.45.0:
* feat: Add `Expr|Series.str.to_date`
* feat: add `ignore_nulls` to `any_horizontal` and
`all_horizontal`
* feat: Add `order_by` option in `with_row_index` method
* patch: Preserve pandas "pyarrow" dtype_backend in
`str.to_datetime`
* fix: zfill for old polars and pyarrow versions
* fix: ensure `fill_null` doesn't silently downcast for pandas
per-3.0
* test: Suppress `NotImplementedError` for `cuDF`
* test: xfail floordiv polars on GPU engine
* ci: Fail polars 0.20.6 in floordiv by zero
* ci: use try_divide in PySpark for division
* test: div by zero returns nan
* chore: share binary comparisons between duckdb/ibis/spark
* typing: Return `Self` whenever possible
* ci: pin sqlglot, use different cache_suffixes for different
jobs
* chore: Add pandas-like `import_array_module` util function
* chore: Add `LazyExprNamespace`
* ci: fixup old versions for floordiv test
- update to 1.44.0:
* fix, perf: avoid np.repeat for broadcasting
* feat: Add support for `Expr.str.zfill`
* fix: Set Ibis time_unit and time_zone for Datetime dtype
conversion
* fix: raise NotImplementedError instead of returning it
* fix, perf: avoid np.repeat for broadcasting
* fix(typing): Resolve intermittent `*GroupBy` variance issue
* fix: ensure `generate_temporary_column_name` result always
starts with character
-------------------------------------------------------------------
Thu Jun 19 11:57:59 UTC 2025 - Daniel Garcia <daniel.garcia@suse.com>
- Initial version 1.43.1
