HyukjinKwon commented on PR #57899: URL: https://github.com/apache/spark/pull/57899#issuecomment-5248672428
Thanks for the review, @dongjoon-hyun! Addressed all the nits in d8444567c4a9: 1. **Arrow return-type guard** — the non-iterator element-wise Arrow branch now calls `verify_scalar_result`, so a non-array-like result raises the friendly `UDF_RETURN_TYPE` error (matching the base `SQL_SCALAR_ARROW_UDF` path) instead of a bare `TypeError` from `len()`. 2. **Overstated test comment** — corrected in `test_scalar_pandas_udf_in_lambda`; the second assertion only exercises arithmetic composition, so the comment no longer claims `filter`/`array_sort` coverage. 3. **ArrowEvalPythonExec Scaladoc** — extended the supported-eval-types list to include the element-wise types (102 and the new 103-106). Also fully-qualified the dangling `[[PythonEvalType.SQL_ARROW_ELEMENTWISE_UDF]]` link in `ExtractPythonUDFFromLambda`. 4. **Added coverage** — `test_chained_vectorized_udfs_in_lambda` (`f(g(x))` for both non-iterator and iterator flavors, asserted against the native equivalent) and `test_scalar_iter_udf_struct_element_return_type` (iterator struct/DataFrame return). I left the duplicated re-nest logic (`renest_spec` in the 102 path vs `_elementwise_renest`) as a follow-up, as suggested. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
