SEZ9 commented on PR #11577: URL: https://github.com/apache/seatunnel/pull/11577#issuecomment-5390959277
@DanielLeens — thanks for picking this up so quickly, and answering your direct question first: the 2026-08-01 possible-API-change concern on `AbstractBigQuerySinkWriter` is **still open**, not resolved. Your own trace of the current head (`0a25d27e5051`) noted that `AbstractBigQuerySinkWriter.applySchemaChange` was unchanged by `cedc9b4f439a`, which matches what I saw — nothing landed that addresses that concern, so I simply didn't repeat it this round to keep the new findings focused. Please carry it forward alongside Issues 1, 4, and 5 in your re-review. That said, I agree with your prioritization: Issue 1 is the blocking item. An unvalidated CDC-supplied column name concatenated into the `ALTER TABLE ... ADD COLUMN` statement reaching a BigQuery multi-statement query job needs to be fixed before merge regardless of the other findings. Concrete asks for the next commit: 1. **Issue 1 (blocking):** validate or safely quote/escape identifiers before they enter the DDL string in `BigQuerySchemaChangeManager`, rather than raw concatenation. 2. **Issue 4:** address the writer-schema-refresh propagation race. 3. **Issue 5:** ensure unsupported DDL doesn't put the task into a crash loop — this and Issue 4 both touch restore/consistency behavior, so they should land together. 4. Any movement on the `AbstractBigQuerySinkWriter` API concern, or a note explaining why the current shape is intentional, so we can close it out explicitly. Once a new commit lands with those, I'll do another full pass on the schema-evolution path in parallel with your re-review. Sound good? <!-- streview-comment:509 --> -- 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]
