zhengruifeng commented on a change in pull request #31480:
URL: https://github.com/apache/spark/pull/31480#discussion_r578958152
##########
File path: core/src/main/scala/org/apache/spark/rdd/OrderedRDDFunctions.scala
##########
@@ -73,7 +75,23 @@ class OrderedRDDFunctions[K : Ordering : ClassTag,
* because it can push the sorting down into the shuffle machinery.
*/
def repartitionAndSortWithinPartitions(partitioner: Partitioner): RDD[(K,
V)] = self.withScope {
- new ShuffledRDD[K, V, V](self, partitioner).setKeyOrdering(ordering)
+ if (self.partitioner == Some(partitioner)) {
+ self.mapPartitions(iter => {
+ val context = TaskContext.get
+ val sorter = new ExternalSorter[K, V, V](context, None, None,
Some(ordering))
+ sorter.insertAll(iter)
+ context.taskMetrics.incMemoryBytesSpilled(sorter.memoryBytesSpilled)
+ context.taskMetrics.incDiskBytesSpilled(sorter.diskBytesSpilled)
Review comment:
I review the related codes and it seems that `sorter.iterator` may spill
during traverse:
`isShuffleSort = false` in `def iterator` makes internals iterator
`destructiveIterator` a `SpillableIterator`.
We can add the update of `taskMetrics` to a task completion listener if
necessary, but maybe in a new ticket.
As to this PR, I perfer to keep the line with existing impl.
----------------------------------------------------------------
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:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
