SEPURI-SAI-KRISHNA commented on issue #11923: URL: https://github.com/apache/seatunnel/issues/11923#issuecomment-5371697748
PR is up: #11925. @DanielLeens thanks for the direction, I followed option 1, and the PR description lists the source paths mirrored so the doc/code mapping can be checked quickly. Two notes on how I read your item 4, since I did not want to either skip `zh` or pad it: `docs/zh/architecture/features/multi-table.md` has no Java listing at all, it is prose plus one formula, so most of the drift you and I listed simply is not present there. What *was* wrong is narrower but real: it described rows as routed directly to the underlying writer (they are queued and drained by `MultiTableWriterRunnable` workers), it omitted the unknown-table / `MultiTableFailurePolicy` branch, and it omitted the schema-change barrier. Those three are fixed, and the §5.3 formula becomes `(hash(pk) & Integer.MAX_VALUE) mod replicaNum` with both the `Math.abs` and `System.nanoTime()` traps called out. So `zh` is updated in the same PR, just with a smaller diff than `en` (+24/-11 vs +116/-84) because there was less that was wrong. I also extended slightly beyond the writer and replica-selection sections, to **§6.2 Schema Evolution Routing**. It carried the identical defect, `handleSchemaChange(...)` (the real method is `applySchemaChange`), the non-existent `writers` field, and a direct loop over replicas where the real path enqueues a `SchemaChangeBarrier` onto every queue. Leaving a `writers.get(identifier)` call in the same file would have contradicted the PR's own claim to have aligned the page. §4.1 I checked and deliberately left alone: `MultiTableSink` really does have `sinks`, `replicaNum`, and `catalogTables`, so it was already correct. One coordination point for whoever reviews: **#11925 and #11721 touch the same regions of this file**, so whichever merges second needs a trivial rebase, which I am happy to do either way. The routing line in #11925 is written in the post-#11721 masked form, because documenting the current `Math.abs` line as the reference pattern would mean documenting a known bug. If you would rather #11925 wait until #11721 lands, say so and I will hold it. @priyanshuvishwakarma273403 apologies for the overlap, I had already started on this one when you asked, which is why I claimed it. Thanks for checking first rather than duplicating work. -- 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]
