YannByron opened a new pull request, #9411: URL: https://github.com/apache/paimon/pull/9411
### Purpose Add a public strict append-only HDF5 ingestion API to MultimodalTable. - Discover individual files, path lists, and directories recursively in deterministic order. - Stream Arrow tables or record batches from transforms while enforcing the exact target schema. - Use one writer and one commit for all files in a call, returning structured file, batch, row, and snapshot results. - Abort and close resources on every pre-commit failure while preserving unknown commit-result semantics. - Read seekable local, HDFS, OSS, S3, and GCS sources through FileIO with isolated source options. - Document that repeated calls append duplicate rows and are not retry-safe. ### Tests - Focused HDF5, FileIO, and legacy OSS unit tests: 58 passed. - Deprecation warnings-as-errors path tests: 3 passed. - Ruff, Flake8, git diff --check, Python license headers, and Apache RAT passed. - Isolated wheel build and content inspection passed, including the hdf5 extra and h5py dependency. - Earlier PyPaimon full suite before the final path-only cleanup: 4351 passed, 133 skipped; the final squashed tree was rechecked with the focused tests above. -- 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]
