Github user chenghao-intel commented on a diff in the pull request: https://github.com/apache/spark/pull/5717#discussion_r32802697 --- Diff: sql/core/src/main/scala/org/apache/spark/sql/execution/joins/SortMergeJoin.scala --- @@ -36,46 +36,91 @@ import org.apache.spark.util.collection.CompactBuffer case class SortMergeJoin( leftKeys: Seq[Expression], rightKeys: Seq[Expression], + joinType: JoinType, left: SparkPlan, - right: SparkPlan) extends BinaryNode { + right: SparkPlan, + condition: Option[Expression] = None) extends BinaryNode { - override def output: Seq[Attribute] = left.output ++ right.output + val (streamed, buffered, streamedKeys, bufferedKeys) = joinType match { + case RightOuter => (right, left, rightKeys, leftKeys) + case _ => (left, right, leftKeys, rightKeys) + } + + override def output: Seq[Attribute] = joinType match { + case Inner => + left.output ++ right.output + case LeftOuter => + left.output ++ right.output.map(_.withNullability(true)) + case RightOuter => + left.output.map(_.withNullability(true)) ++ right.output + case FullOuter => + left.output.map(_.withNullability(true)) ++ right.output.map(_.withNullability(true)) + case x => + throw new IllegalStateException(s"SortMergeJoin should not take $x as the JoinType") + } - override def outputPartitioning: Partitioning = left.outputPartitioning + override def outputPartitioning: Partitioning = joinType match { + case FullOuter => + // when doing Full Outer join, NULL rows from both sides are not so partitioned. + UnknownPartitioning(streamed.outputPartitioning.numPartitions) + case _ => streamed.outputPartitioning --- End diff -- I think @JoshRosen is right, but currently all of the existed equi-join has the same bug which described at https://issues.apache.org/jira/browse/SPARK-2205 , so for now, we'd better keep it the same as it was to be (just take the streaming output partitioning).
--- 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