voonhous commented on PR #19791: URL: https://github.com/apache/hudi/pull/19791#issuecomment-5461872342
The existing Spark check is a drive-by from HUDI-8866 (`a93c6eec6788`): once each NB-CC bucket got one fixed file id, overwrite had no valid choice (the fixed id collides with the replaced group, a random id splits the bucket), so it was rejected, untested. The invariant is broader than insert overwrite: `delete_partition` then re-insert (`SparkDeletePartitionCommitActionExecutor.java:66`) hides data the same way. Could the description state "a fixed-id file group must never be listed in a replacecommit", and is bumping the `-0` generation in `newBucketFileIdForNBCC` on replace the root fix worth tracking instead of per-operation guards? -- 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]
