ForVic commented on PR #59486: URL: https://github.com/apache/spark/pull/59486#issuecomment-6357413782
Thanks @cloud-fan and @peter-toth for the reviews. I've addressed them in 07d5f8d. @cloud-fan, on whether the reader should accept expensive key-only conjuncts such as UDFs and regexes: I went with Peter's suggestion. They're now behind their own conf, `spark.sql.parquet.storageFilterPushdown.keyConditions.enabled`, off by default. Its doc names the reordering and the fatal-error exposure. The bloom alone stays under the existing conf. I didn't reuse `InjectRuntimeFilter.isSimpleExpression`. It would keep out the expensive cases, but not every expression that can fail fatally, and it would also keep out reasonable conditions. Happy to add it on top of the conf if you'd prefer both. -- 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]
