[SPARK-46158][PYTHON][PS] Support axis columns in DataFrame.xs

[manhha2502]

What changes were proposed in this pull request?

This PR adds support for DataFrame.xs with axis=1 / axis="columns" in pandas API on Spark.

It supports selecting cross-sections from columns, including single-level columns, MultiIndex columns, and level-based selection by level number or level name.

Why are the changes needed?

DataFrame.xs currently supports index-axis selection, but column-axis selection raises NotImplementedError.

This improves compatibility with pandas and addresses SPARK-46158.

Does this PR introduce any user-facing change?

Yes.

Previously, DataFrame.xs(..., axis=1) raised NotImplementedError.

After this change, users can select column cross-sections, for example:

psdf.xs("metrics", axis=1)
psdf.xs("metrics", axis="columns")
psdf.xs("num_legs", axis=1, level="feature")

How was this patch tested?

Added positive and negative test cases in FrameIndexingMixin.test_xs covering axis=1, axis="columns", MultiIndex columns, level-based selection, missing column keys, and invalid column levels.

Also ran:

python -m py_compile python/pyspark/pandas/frame.py python/pyspark/pandas/tests/indexes/test_indexing.py
git diff --check

I attempted to run the targeted unittest locally, but the source checkout has not built Spark jars yet.

Was this patch authored or co-authored using generative AI tooling?

Generated-by: OpenAI Codex (GPT-5)

[HyukjinKwon]

cc @devin-petersohn fyi