cloud-fan commented on a change in pull request #23750: [SPARK-19712][SQL] 
Pushing Left Semi and Left Anti joins through Project, Aggregate, Window, Union 
etc.
URL: https://github.com/apache/spark/pull/23750#discussion_r270134939
 
 

 ##########
 File path: 
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/PushDownLeftSemiAntiJoin.scala
 ##########
 @@ -0,0 +1,197 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.sql.catalyst.optimizer
+
+import org.apache.spark.sql.catalyst.expressions._
+import org.apache.spark.sql.catalyst.plans._
+import org.apache.spark.sql.catalyst.plans.logical._
+import org.apache.spark.sql.catalyst.rules.Rule
+
+/**
+ * This rule is a variant of [[PushDownPredicate]] which can handle
+ * pushing down Left semi and Left Anti joins below the following operators.
+ *  1) Project
+ *  2) Window
+ *  3) Union
+ *  4) Aggregate
+ *  5) Other permissible unary operators. please see 
[[PushDownPredicate.canPushThrough]].
+ */
+object PushDownLeftSemiAntiJoin extends Rule[LogicalPlan] with PredicateHelper 
{
+  def apply(plan: LogicalPlan): LogicalPlan = plan transform {
+    // LeftSemi/LeftAnti over Project
+    case Join(p @ Project(pList, gChild), rightOp, LeftSemiOrAnti(joinType), 
joinCond, hint)
+      if pList.forall(_.deterministic) &&
+        !pList.exists(ScalarSubquery.hasCorrelatedScalarSubquery) &&
+        canPushThroughCondition(Seq(gChild), joinCond, rightOp) =>
+      if (joinCond.isEmpty) {
+        // No join condition, just push down the Join below Project
+        p.copy(child = Join(gChild, rightOp, joinType, joinCond, hint))
+      } else {
+        val aliasMap = PushDownPredicate.getAliasMap(p)
+        val newJoinCond = if (aliasMap.nonEmpty) {
+          Option(replaceAlias(joinCond.get, aliasMap))
+        } else {
+          joinCond
+        }
+        p.copy(child = Join(gChild, rightOp, joinType, newJoinCond, hint))
+      }
+
+    // LeftSemi/LeftAnti over Aggregate
+    case join @ Join(agg: Aggregate, rightOp, LeftSemiOrAnti(joinType), 
joinCond, hint)
+      if agg.aggregateExpressions.forall(_.deterministic) && 
agg.groupingExpressions.nonEmpty &&
+        
!agg.aggregateExpressions.exists(ScalarSubquery.hasCorrelatedScalarSubquery) =>
+      if (joinCond.isEmpty) {
+        // No join condition, just push down Join below Aggregate
+        agg.copy(child = Join(agg.child, rightOp, joinType, joinCond, hint))
+      } else {
+        val aliasMap = PushDownPredicate.getAliasMap(agg)
+
+        // For each join condition, expand the alias and check if the 
condition can be evaluated
+        // using attributes produced by the aggregate operator's child 
operator.
+        val (pushDown, stayUp) = 
splitConjunctivePredicates(joinCond.get).partition { cond =>
+          val replaced = replaceAlias(cond, aliasMap)
+          cond.references.nonEmpty &&
+            replaced.references.subsetOf(agg.child.outputSet ++ 
rightOp.outputSet)
+        }
+
+        // Check if the remaining predicates do not contain columns from the 
right
+        // hand side of the join. Since the remaining predicates will be kept
+        // as a filter over aggregate, this check is necessary after the left 
semi
+        // or left anti join is moved below aggregate. The reason is, for this 
kind
+        // of join, we only output from the left leg of the join.
+        val rightOpColumns = 
AttributeSet(stayUp.toSet).intersect(rightOp.outputSet)
+
+        if (pushDown.nonEmpty && rightOpColumns.isEmpty) {
+          val pushDownPredicate = pushDown.reduce(And)
+          val replaced = replaceAlias(pushDownPredicate, aliasMap)
+          val newAgg = agg.copy(child = Join(agg.child, rightOp, joinType, 
Option(replaced), hint))
+          // If there is no more filter to stay up, just return the Aggregate 
over Join.
+          // Otherwise, create "Filter(stayUp) <- Aggregate <- 
Join(pushDownPredicate)".
 
 Review comment:
   Left-anti join outputs records that do NOT satisfy the join condition. Let's 
say the join condition is `a && b`, and this rule turns `Join(Aggregate, ..., a 
&& b)` to `Filter(b, Aggregate(Join(..., ..., a)))`.
   
   This seems problematic. Previously we get result satisfying `Not(a && b)`, 
now we get `Not(a) && b`. @hvanhovell is this your concern?

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to