viirya commented on PR #59848: URL: https://github.com/apache/spark/pull/59848#issuecomment-6476396083
> Should the throwable-order guarantee also cover Python and Pandas UDF projections? Confirmed by running it. The Python UDF query raises `USER_RAISED_EXCEPTION` for both regular and pandas UDFs, and with `avoidDoubleFilterEval` off as well. It comes from the generic `UnaryNode` case (`pushDownPredicate`), which pushes `assert_true(id > 5) IS NULL` alone below `ArrowEvalPython`. That predates both SPARK-47672 and SPARK-55014, and it applies to every operator in `canPushThrough` (Generate, Expand, Sort, ...). So I filed SPARK-60633 for it, and narrowed this PR's description to the `Filter` over `Project` case. -- 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]
