JingsongLi commented on PR #10158: URL: https://github.com/apache/paimon/pull/10158#issuecomment-5832151751
Reviewed the current head as a cross-engine data-location fix. The new option is passed to Python's `FileStorePathFactory` for writes and to all three split generators when resolving manifest data files for reads. The directory is inserted between the table root and partition/bucket path, matching Java `FileStorePathFactory`'s `dataFilePathDirectory` layout. Native write, plan, read, and commit dispatch all fall back to Python when this option is set, so they cannot silently use the bucket-root layout. The unset option keeps the existing path. Local verification: 124 directory, data-evolution, native-plan, and native-read tests passed, plus 54 subtests; `git diff --check` passed. The real-table test wrote and read relocated files, and the native-requested tests exercised each fallback. The current CI is green for Native CI and all scheduled Python versions. The earlier native-path and test-double failures are addressed at this head. I found no blocking regression. A small Java-write/Python-read or Python-write/Java-read fixture with `data-file.path-directory` would be valuable as a durable cross-engine compatibility check, since that is the user-visible failure this fixes. -- 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]
