Github user mgaido91 commented on the issue: https://github.com/apache/spark/pull/22029 @cloud-fan if we consider only the expression IN with literals, yes, the behavior is very similar, with the following difference: Presto throws exception when null is present on both sides. But what I'd really like to highlight is that there is a huge difference: both Hive and Presto (as well as the RDBMs are coherent in their behavior between IN with literals and IN with subquery). Actually, Hive doesn't support subqueries with more than one output field; Presto instead behaves in the same way in the case with literals and subqueries, which means it behaves like Spark before this PR with literals, but different from Spark when subqueries are involved.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org