Github user rxin commented on a diff in the pull request: https://github.com/apache/spark/pull/13906#discussion_r68495048 --- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala --- @@ -1053,6 +1055,34 @@ object PruneFilters extends Rule[LogicalPlan] with PredicateHelper { } /** + * Collapse plans consisting all empty local relations generated by [[PruneFilters]]. + * Note that the ObjectProducer/Consumer and direct aggregations are the exceptions. + * {{{ + * SELECT a, b FROM t WHERE 1=0 GROUP BY a, b ORDER BY a, b ==> empty result + * SELECT SUM(a) FROM t WHERE 1=0 GROUP BY a HAVING COUNT(*)>1 ORDER BY a (Not optimized) + * }}} + */ +object CollapseEmptyPlan extends Rule[LogicalPlan] with PredicateHelper { + private def isEmptyLocalRelation(plan: LogicalPlan): Boolean = + plan.isInstanceOf[LocalRelation] && plan.asInstanceOf[LocalRelation].data.isEmpty + + def apply(plan: LogicalPlan): LogicalPlan = plan transformUp { + case x if x.isInstanceOf[ObjectProducer] || x.isInstanceOf[ObjectConsumer] => x + + // Case 1: If groupingExpressions contains all aggregation expressions, the result is empty. + case a @ Aggregate(ge, ae, child) if isEmptyLocalRelation(child) && ae.forall(ge.contains(_)) => --- End diff -- this kind of blacklisting approach is too risky -- if we were to introduce a new logical node in the future, most likely we will forget to update this rule.
--- 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