adiranjan25 commented on PR #18918: URL: https://github.com/apache/iceberg/pull/18918#issuecomment-6830988773
Hi maintainers, I've submitted this PR as initial regression coverage for the subset-key `MERGE INTO` partition-grouping behavior discussed in #18597. The tests reproduce the difference in partition grouping with partially clustered distribution enabled versus disabled, while verifying MERGE correctness. I'd appreciate technical feedback on three points: 1. **Regression coverage:** Does this test adequately capture the partition-concentration behavior described in #18597? Are there additional scenarios or assertions that would make the coverage more valuable? 2. **Technical significance:** How important is this behavior for Iceberg users running large-scale MERGE workloads, and what additional evidence would help establish its practical impact? 3. **Next implementation:** Would an Iceberg-side planning improvement or node-local shuffle mechanism be the appropriate next step, or is there a preferred alternative? I'm happy to strengthen the tests, run additional reproducible benchmarks, and contribute a follow-on implementation based on maintainer guidance. Thank you for reviewing this contribution. -- 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]
