TheR1sing3un opened a new pull request, #10060: URL: https://github.com/apache/paimon/pull/10060
### Purpose `map_with_blobs()` rejects ARRAY and MAP BLOB columns that can already be read locally. Include these columns in source metadata and use the shared coalesced payload reader to provide row-aligned lists and key-value pairs to Ray callbacks, preserving null and empty values. Pass nested column information to workers so the table method also works after row-level Ray filters convert MAP columns to Python object extension arrays. The standalone function accepts optional `map_blob_columns` and `array_blob_columns`; native Arrow nested types remain usable without those hints. Exclude all source BLOB columns from the scalar batch and extend foreign-descriptor checks to nested and Python object columns. ### Tests - Five Ray BLOB regressions pass with Ray 2.54, covering scalar BLOB compatibility, direct scan metadata, table mapping after a row filter, projections, and empty results. - Additional regressions cover Python object columns, foreign descriptors after inline values, sliced arrays, empty chunks, and invalid nested column metadata. The two focused batch tests also pass on PyArrow 16. - Flake8, changed-file license headers, Python 3.6 grammar checks, and `git diff --check` passed. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
