Github user hvanhovell commented on a diff in the pull request: https://github.com/apache/spark/pull/9566#discussion_r44514711 --- Diff: sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/AggregationQuerySuite.scala --- @@ -545,19 +576,21 @@ abstract class AggregationQuerySuite extends QueryTest with SQLTestUtils with Te | count(distinct value2), | sum(distinct value2), | count(distinct value1, value2), + | longProductSum(distinct value1, value2), --- End diff -- Yes and No. The input we care about only consists of these tuples: ```[value1=null, value2=null], [value1=null, value2=1], [value1=1, value2=null], and [value1=1, value2=1]``` However in the current implementation a distinct aggregate will see more input than those. It will also see records from other groups. However, the values in these records are nulled out. The assumption here is that an AggregateFunction is not changed by an all NULL update. The only case I can think of that would be problematic is a ```FIRST(DISTINCT ...)```; which shouldn't be used like that anyway. We could solve this by wrapping AggregateFunctions with an operator which will only update if the group id is correct.
--- 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