Github user wzhfy commented on the issue: https://github.com/apache/spark/pull/19321 @felixcheung For percentiles, I think the type of results should be the same as input data type. In these tests, the type of data is int, so actually `30` is more accurate than `30.0`. The previous answer is `30.0` because `ApproximatePercentile` only accepts double input type and outputs double results before this PR.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org