diff options
Diffstat (limited to 'sql/catalyst')
| -rw-r--r-- | sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala | 16 |
1 files changed, 15 insertions, 1 deletions
diff --git a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala index b6ad5db74e..6ba8b33b3f 100644 --- a/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala +++ b/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala @@ -689,7 +689,7 @@ object PushDownPredicate extends Rule[LogicalPlan] with PredicateHelper { // state and all the input rows processed before. In another word, the order of input rows // matters for non-deterministic expressions, while pushing down predicates changes the order. case filter @ Filter(condition, project @ Project(fields, grandChild)) - if fields.forall(_.deterministic) => + if fields.forall(_.deterministic) && canPushThroughCondition(grandChild, condition) => // Create a map of Aliases to their values from the child projection. // e.g., 'SELECT a + b AS c, d ...' produces Map(c -> a + b). @@ -830,6 +830,20 @@ object PushDownPredicate extends Rule[LogicalPlan] with PredicateHelper { filter } } + + /** + * Check if we can safely push a filter through a projection, by making sure that predicate + * subqueries in the condition do not contain the same attributes as the plan they are moved + * into. This can happen when the plan and predicate subquery have the same source. + */ + private def canPushThroughCondition(plan: LogicalPlan, condition: Expression): Boolean = { + val attributes = plan.outputSet + val matched = condition.find { + case PredicateSubquery(p, _, _, _) => p.outputSet.intersect(attributes).nonEmpty + case _ => false + } + matched.isEmpty + } } /** |