Github user gatorsmile commented on the issue: https://github.com/apache/spark/pull/21320 The feature has already been developed for almost two years. I am feeling sorry to merge it in Spark 2.4 release. Personally, I think we should not block merging this PR to Spark 2.4 release, even if the solution might still have a few issues we need to address. This PR only includes a rule and it does not touch any common code. Thus, the Spark users will face zero risk if we just turn off the conf `spark.sql.nestedSchemaPruning.enabled` by default. More importantly, if we merge it now, we can collect the feedbacks from Spark users who are waiting for this feature and fix the holes if existed in the next releases. @mallman Could you remove the changes made in ParquetRowConverter.scala and also turn off `spark.sql.nestedSchemaPruning.enabled` by default in this PR?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org