Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/5717#discussion_r32327932
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/joins/SortMergeJoin.scala
---
@@ -84,84 +129,164 @@ case class SortMergeJoin(
override final def next(): Row = {
if (hasNext) {
- // we are using the buffered right rows and run down left
iterator
- val joinedRow = joinRow(leftElement,
rightMatches(rightPosition))
- rightPosition += 1
- if (rightPosition >= rightMatches.size) {
- rightPosition = 0
- fetchLeft()
- if (leftElement == null || keyOrdering.compare(leftKey,
matchKey) != 0) {
- stop = false
- rightMatches = null
+ if (bufferedMatches == null || bufferedMatches.size == 0) {
+ // we just found a row with no join match and we are here to
produce a row
+ // with this row with a standard null row from the other
side.
+ if (continueStreamed) {
+ val joinedRow = smartJoinRow(streamedElement,
bufferedNullRow.copy())
--- End diff --
Why do we have to copy the standard null rows? Are we worried about
downstream operators mutating them?
---
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]
