Github user WeichenXu123 commented on the issue: https://github.com/apache/spark/pull/21120 I doubt that this will slow down the summarizer performance because you add sum statistics internally (and this sum value will possible to overflow). We can directly use `count * mean` to get the sum if we want to use it.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org