davidzollo commented on PR #12142: URL: https://github.com/apache/seatunnel/pull/12142#issuecomment-6886005014
### Implementation Notes and Trade-offs Head: `04798d7ffec1b657d6a9efebea89cfd98e41d2e9` I addressed the existing Doris E2E feedback by using the same source relation for the bounded sink wait and the row-by-row comparison. The unique-key cases now wait for the rows that satisfy `F_ID > 50`; the duplicate-key case still uses all source rows. This fixes the mismatch between the old unfiltered count (91) and the 45 qualifying rows actually written to the sink. The wait still requires a nonempty source and an exact matching sink count, then runs the original schema and full content assertions. It retries SQL query errors only while waiting for the sink. Sharing the predicate avoids a second copy that could drift; increasing the timeout alone would still wait for the wrong count. The original startup budget is unchanged, and the separately merged `testCustomSql` behavior comes from `dev`. Scoped Spotless apply/check and the diff whitespace check passed. GitHub confirms that the PR has no merge conflict and still changes only the two Doris test files. The updated [Build](https://github.com/apache/seatunnel/actions/runs/43653995528) was queued at readback; both Doris matrix legs still need to finish. -- 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]
