Github user nongli commented on a diff in the pull request:
https://github.com/apache/spark/pull/11372#discussion_r54990168
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/Optimizer.scala
---
@@ -586,6 +587,52 @@ object NullPropagation extends Rule[LogicalPlan] {
}
/**
+ * Attempts to eliminate reading (unnecessary) NULL values if they are not
required for correctness
+ * by inserting isNotNull filters is the query plan. These filters are
currently inserted beneath
+ * existing Filters and Join operators and are inferred based on their
data constraints.
+ *
+ * Note: While this optimization is applicable to all types of join, it
primarily benefits Inner and
+ * LeftSemi joins.
+ */
+object NullFiltering extends Rule[LogicalPlan] with PredicateHelper {
+ def apply(plan: LogicalPlan): LogicalPlan = plan transform {
+ case filter @ Filter(condition, child: LogicalPlan) =>
+ // We generate a list of additional isNotNull filters from the
operator's existing constraints
+ // but remove those that are either already part of the filter
condition or are part of the
+ // operator's child constraints.
+ val newIsNotNullConstraints =
filter.constraints.filter(_.isInstanceOf[IsNotNull]) --
+ (child.constraints ++ splitConjunctivePredicates(condition))
+ val newCondition = if (newIsNotNullConstraints.nonEmpty) {
+ And(newIsNotNullConstraints.reduce(And), condition)
+ } else {
+ condition
+ }
+ Filter(newCondition, child)
+
+ case join @ Join(left: LogicalPlan, right: LogicalPlan, joinType:
JoinType,
+ condition: Option[Expression]) =>
+ val leftIsNotNullConstraints = join.constraints
--- End diff --
indenting
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
