anshulsingh-py commented on PR #19853: URL: https://github.com/apache/hudi/pull/19853#issuecomment-5584669926
Addressed the last review comment: `lookupFunction` skips the analyzer's implicit-cast pass, so a structurally valid call against the wrong column type (e.g. `concat(id, 'x')` on an int column) was resolving and passing validation, then throwing at eval — same failure mode this whole PR has been closing. Added `resolved && checkInputDataTypes().isSuccess` to the guard, with test coverage. Also applied the two nits: renamed `tableResolved` → `hardcodedResolved`, and tightened the `resolveViaFunctionRegistry` comments into one upfront explanation instead of scattered narration. Verified locally: 27/27 in `TestHoodieProcedureFilterUtils`, 256/260 in the full `procedure` package (the 4 failures are a pre-existing local build artifact issue in an unrelated `hudi-cli`/bootstrap module, not reproducible from a clean build — confirmed the stack trace points to a stale `SparkKeyGenUtils` classpath resolution, unrelated to this change). -- 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]
