narwhals 2.2.0

Narwhals v2.2.0

on Python PyPI

Changes

✨ Enhancements

• feat: Add {Expr,Series}.fill_nan (#3030)
• feat: Allow mode(..., keep="any") (#3019)
• feat: add DataFrame.top_k and LazyFrame.top_k (#2977)
• feat(typing): Add IntoBackend alias (#2971)
• feat: Add returns_scalar argument in map_batches (#2998)
• feat: Add Series.from_iterable (#2933)

🐞 Bug fixes

• test(typing): Avoid masking fixed overloads (#3033)
• fix(typing): Avoid overlapping DataFrame, LazyFrame (#2944)
• fix: map_batches parsing result type (#3008)
• fix: Add missing v2 constructor overrides (#3006)
• ci: Remove duckdb nightly from cov (#3015)
• fix: expr & series bool ambiguity (#2985)
• fix: GroupBy with mixed length exprs as keys (#3004)
• fix: Workaround for clip with modin[pyarrow] and cudf backends (#2986)

📖 Documentation

• docs: fix typo, remove outdated "you can only execute this once" (#3035)
• feat(typing): Add IntoBackend alias (#2971)
• docs: Fix indentation of coalesce example in doc string (#3011)
• docs: Include DType methods, properties in API Reference (#2999)
• fix(tests): ensure doctests work on windows (#2994)

🛠️ Other improvements

• chore: xfail top k multiple in old polars (#3034)
• chore: introduce Expr._with_unary / _expr_with_n_ary_op (#3021)
• chore: update ruff config (#3009)
• chore: replace zip with zip_strict (#3003)
• test: reuse eager_backend fixture (#3005)
• fix(tests): ensure doctests work on windows (#2994)
• test: Simplify read_scan_test, spark session (#3024)

Thank you to all our contributors for making this release possible!
@FBruzzesi, @MarcoGorelli, @camriddell, @dangotbanned, @kklein and @raisadz