Github user thunterdb commented on a diff in the pull request: https://github.com/apache/spark/pull/16971#discussion_r102592174 --- Diff: sql/core/src/main/scala/org/apache/spark/sql/execution/stat/StatFunctions.scala --- @@ -78,7 +80,12 @@ object StatFunctions extends Logging { def apply(summaries: Array[QuantileSummaries], row: Row): Array[QuantileSummaries] = { var i = 0 while (i < summaries.length) { - summaries(i) = summaries(i).insert(row.getDouble(i)) + if (!row.isNullAt(i)) { --- End diff -- Thank you for fixing this issue, it was an oversight in my original implementation. The current exception being thrown depends on an implementation detail (calling `sampled.head`). Can you modify the function `def query` below to explicitly throw an exception if `sampled` is empty, and document this behavior in that function? This way, we will not forget it if decide to change the semantics of that class. As @MLnick was mentioning above, it would be preferrable to either return an Option, null or NaN eventually, but this can wait for more consensus.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org