SEZ9 commented on PR #13092: URL: https://github.com/apache/seatunnel/pull/13092#issuecomment-6905463219
@SEPURI-SAI-KRISHNA thanks for the point-by-point reply, and sorry for the lag — my notes were indeed written against the earlier version. I have not re-read `6a59aa1a32` yet, so I can't mark anything resolved from the description alone, but here is where I expect things to land once I do: - F3, F4, F5 and F8 were all about the NULL probe / extra NULL split / new split class. If the current head has none of those, as you describe, they should no longer apply; I'll confirm against the diff. - F2 was about how the extra split interacts with the explicit bounds; with no extra split there should be nothing left to document there. - F6: the concern was the "same splits as before" wording. From your description, tables with a nullable primary key or unique key column are now split on the next NOT NULL key column or read as one split, so splits do change for those tables. Could you confirm the rewritten `incompatible-changes.md` entry no longer claims unchanged splits for that case (or quote the current sentence)? - F1: if `findSplitKey` returns no key for a `query` without `partition_column` and the table is read as one split, that removes the silent-skip concern. Is there a `DynamicChunkSplitterTest` case that pins the "`query` without `partition_column` → one split" behaviour, or is it covered only in docs/Javadoc at the moment? - F7: with `isNullableSplitKey` gone, the remaining question is the fallback when the table metadata does not report nullability for a candidate key column — is it treated as NOT NULL and used, or skipped with the WARN? If the `isAutomaticSplitColumn` Javadoc already says this, point me at it and I'll drop the item. I'll re-read `6a59aa1a32` with the above in mind and withdraw whatever no longer applies. @nzw921rx the re-approval request is yours once you are happy with the wording. <!-- streview-comment:2376 --> -- 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]
