DanielLeens commented on issue #12116: URL: https://github.com/apache/seatunnel/issues/12116#issuecomment-5584884264
Thanks for tracing this against the published contract. Current `dev` still has a static `LONG_ACCUMULATOR`, while the Assert documentation says `MIN_ROW` and `MAX_ROW` are evaluated once at close against the rows observed by that writer instance. The row-count part of the intended scope is therefore clear: the first fix should replace the JVM-wide accumulator with writer-owned state and preserve the existing close-time evaluation boundary. Please keep that PR limited to the row-count defect. Add a deterministic regression with two writers for the same table (and, where relevant, sequential writers in one JVM) proving that one writer's close result cannot be affected by another writer's rows. Do not fold the separate static `TABLE_NAMES` question into this change: aggregate-versus-writer-local table-name semantics need their own documented contract and lifecycle analysis before changing them. The fix must not silently redefine multi-writer assertions as job-global aggregation. -- 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]
