SEZ9 commented on PR #12619: URL: https://github.com/apache/seatunnel/pull/12619#issuecomment-6817042960
@goutamadwant Thanks for the detailed breakdown on F4, and sorry for the slow reply. (a) The `projectsTransformFromPhysicalDagInfo` test you describe (running `lineage_transform.conf` through `DAGUtils.getJobDAGInfo(..., true, ...)` and asserting the TRANSFORM node, `NOT_APPLICABLE` and empty table paths) sounds like it targets exactly the branch I was concerned about. I have not yet verified it against the code at 441ee3e5f8331c0038f7d1fb2a024087f4ce20ac, so I will do that before marking this part of F4 resolved. Given your note that `JobMaster`, `CoordinatorService` and `BaseService` all pass `true`, prioritising the physical flavour in this PR makes sense to me. (b) Deferring the logical flavour (`getJobDAGInfo(..., false, ...)`) is fine. The `ActionStateKey` mismatch you describe in `DAGUtils` is pre-existing and out of scope for the projector, and handling it together with chained-transform and union-split mapping in its own PR is reasonable. Two small asks so the gap stays visible: 1. Please add a short note (a one-line comment near the engine-backed tests in `JobLineageProjectorTest`, or an entry in the PR description checklist) stating that only the physical `JobDAGInfo` shape is exercised and why. 2. Once the follow-up issue/PR exists, link both from the checklist as you proposed; I will treat F4 as fully closed once that link is in place and I have confirmed the test in the diff. Nothing else outstanding from me on F4. <!-- streview-comment:2283 --> -- 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]
