kiyeonjeon21 commented on issue #15369: URL: https://github.com/apache/iceberg/issues/15369#issuecomment-4187340224
I'd like to take this on. I looked through the code and the main changes I see are: 1. Adding a `toBranch` method to the `RemoveDanglingDeleteFiles` API (with a default impl to avoid breaking changes) 2. Updating `RemoveDanglingDeletesSparkAction` to scope metadata reads and commits to the target branch 3. Passing the branch through from `RewriteDataFilesSparkAction` when it invokes the action 4. Revisiting the early return for unpartitioned tables One thing I want to confirm: for the unpartitioned table case, the current comment says "ManifestFilterManager already performs this table-wide delete on each commit." Is the intent to keep skipping but log a warning, or should we actually process unpartitioned tables? The issue author's use case (equality deletes from Flink upsert on a branch) seems like it could hit this path. Happy to discuss before I start coding. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
