narwhals 2.16.0

Narwhals v2.16.0

on Python PyPI

Changes

• ci: Unpin polars==1.34.0 in --group typing (#3434)
• ci: Bump duckdb==1.4.4 in --group typing (#3433)
• test: DuckDB XPass (#3426)
• refactor: Simplify pd.ArrowDtype -> nw.DType (#3413)
• ci: Temporarily pin sqlglot, ignore pyspark warning (#3412)
• Remove interchange from non v1 (#3403)

✨ Enhancements

• feat: add separator argument in read_csv/scan_csv (#2989)
• feat: Allow nested structures in lit (#3424)
• enh: Introduce narwhals.sql (#3254)
• enh: Introduce (optional) order_by in first / last (#3372)
• feat: support window functions in filter (#3401)
• feat: Improve support for Decimal DType (#3377)
• feat: Support concat(..., how="diagonal") for ibis (#3404)
• feat: Enablelist.{sort, sum} for sqlframe (#3400)
• feat: Add str.pad_{start,end} (#3395)
• feat: Add {Expr,Series}.cos (#3392)
• feat: Add testing.assert_frame_equal (#3220)

🐞 Bug fixes

• fix(test): Pin correct polars version in lit_test (#3438)
• refactor: Avoid subprocess to test TPCH queries, and fix q8 (#3419)
• ci(fix): Temporary pin numba & llvmlite for darts downstream test (#3406)
• fix(test): Change error message for polars, un-xfail sqlframe list.mean (#3397)

🛠️ Other improvements

• ci: Pin sqlglot<28.6.0 in --group typing (#3432)
• chore: pin sqlglot to get ci green (#3428)
• chore(typing): Improve tpch typing (#3420)
• ci: pin pandas in some downstream jobs (#3417)
• [pre-commit.ci] pre-commit autoupdate (#3275)

Thank you to all our contributors for making this release possible!
@FBruzzesi, @MarcoGorelli, @camriddell, @dangotbanned, @liamholmes31, and @raisadz