SEZ9 commented on issue #12818: URL: https://github.com/apache/seatunnel/issues/12818#issuecomment-6207648081
Thanks for filing this. A couple of things would help scope the work: - Could you point to the two places in the fixed-bucket routing path where the same row is fully converted, and whether this is observable on the current main branch or only on top of the parent #12366 change? If it depends on the parent landing, it may be worth noting that under "Blocked by". - Do you have any measurement (e.g. throughput or a profiler sample) showing how much the second conversion costs? That would help decide between reusing the already-converted row for routing versus extracting only the partition/primary-key fields. For coverage, a unit test asserting the optimized routing picks the same bucket as the existing logic, plus an e2e case with a multi-table fixed-bucket Paimon sink (bucket > 1, parallelism > 1) comparing results against current behavior, would satisfy the acceptance criteria nicely. <!-- streview-comment:1802 --> -- 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]
