fresh-borzoni opened a new pull request, #29697: URL: https://github.com/apache/flink/pull/29697
## What is the purpose of the change With multi-join enabled, the upsert keys of a multi-join could include a key that is not unique in its output, so the sink dropped the SinkUpsertMaterializer and a delete could remove a row that still exists. FlinkRelMdUpsertKeys used input 0's join columns as the left join key of every step instead of the columns that step actually joins on. This PR uses each step's own left join columns. ## Brief change log - Add StreamPhysicalMultiJoin#getLeftJoinKeyIndices, the columns an input's equi-join conditions reference in the inputs before it. - Use it as the left join key of each step in FlinkRelMdUpsertKeys#getUpsertKeys for StreamPhysicalMultiJoin. ## Verifying this change - MultiJoinTest - JoinITCase ## Does this pull request potentially affect one of the following parts: - Dependencies (does it add or upgrade a dependency): no - The public API, i.e., is any changed class annotated with `@Public(Evolving)`: no - The serializers: no - The runtime per-record code paths (performance sensitive): no - Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: no - The S3 file system connector: no -- 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]
