Github user tdas commented on a diff in the pull request: https://github.com/apache/spark/pull/19452#discussion_r144437093 --- Diff: sql/core/src/main/scala/org/apache/spark/sql/execution/streaming/StreamingSymmetricHashJoinExec.scala --- @@ -161,6 +164,10 @@ case class StreamingSymmetricHashJoinExec( new SerializableConfiguration(SessionState.newHadoopConf( sparkContext.hadoopConfiguration, sqlContext.conf))) + + val nullLeft = new GenericInternalRow(left.output.map(_.withNullability(true)).length) + val nullRight = new GenericInternalRow(right.output.map(_.withNullability(true)).length) --- End diff -- This can also be moved into the Joiner. Use `Joiner.inputAttributes`, instead of `left.output`/`right.output` Basically if you have to define variables outside the Joine with the name "left", then it highly likely that you can move into the Joiner, as it is meant capture all the context needed to process one side.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org