Github user cloud-fan commented on the issue: https://github.com/apache/spark/pull/22561 Please describe the idea clearly in the PR description. After reading the code, I think the idea is, for any predicate `p`, we can rewrite it to `p'` that only contains partition columns. The requirement is, partitions returned by the partition predicate `p`, must also be returned by `p'`. We need to be careful here, to not introduce correctness issues. Can you prove that your rewrite is safe?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org