SEPURI-SAI-KRISHNA commented on issue #12968: URL: https://github.com/apache/seatunnel/issues/12968#issuecomment-6410239627
Confirmed on all three, @SEZ9, and nothing in the PR has moved since. - The three `ArrowToSeatunnelRowReaderNullCellTest` cases stay as described: a null MAP cell reads as `null`, a null ROW cell reads as `null` without aborting the read, and `testEmptyMapCellStaysAnEmptyMap` keeps empty distinct from null. - The EN and ZH `incompatible-changes.md` entries for the MAP behaviour change stay in, including the note telling a downstream consumer that assumed a non-null map to handle `null`. - The older `// todo check map` in `ArrowToSeatunnelRowReaderTest` is untouched, so the PR stays scoped to this one problem. Current state of #12969: rebased onto `dev`, `MERGEABLE` with no conflict, at `a66150a38`. @DanielLeens re-reviewed that head and closed every prior numbered finding with no blocking findings. The one open item is CI, and it is not this change. The last run failed on `AmazonDocumentDBIT.testProjection` under Spark 3.3, which asserts `MIN_ROW` 3 and saw `row num: 0`, so the source read nothing. That connector has zero references to arrow in its main sources; the job only ran at all because `connector-common` is an upstream module of every connector, so a change there expands the updated-modules set. I have rerun it. So from my side the PR needs only a green Build and a write-access review. -- 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]
