Github user srowen commented on the issue: https://github.com/apache/spark/pull/19565 Regarding caching: I think that can be ignored for purposes of this change. All this does is add a filter, and it doesn't cause an RDD to computed more than it was before. The only question is whether the filtering is worth it; does it filter out enough that it makes later work faster?
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org