tomvanbussel commented on a change in pull request #29787:
URL: https://github.com/apache/spark/pull/29787#discussion_r490774220
##########
File path:
core/src/main/java/org/apache/spark/util/collection/unsafe/sort/UnsafeExternalSorter.java
##########
@@ -527,19 +527,23 @@ public long getCurrentPageNumber() {
public long spill() throws IOException {
synchronized (this) {
- if (inMemSorter == null || numRecords <= 0) {
+ if (inMemSorter == null) {
return 0L;
}
long currentPageNumber = upstream.getCurrentPageNumber();
ShuffleWriteMetrics writeMetrics = new ShuffleWriteMetrics();
- // Iterate over the records that have not been returned and spill them.
- final UnsafeSorterSpillWriter spillWriter =
- new UnsafeSorterSpillWriter(blockManager, fileBufferSizeBytes,
writeMetrics, numRecords);
- spillIterator(upstream, spillWriter);
- spillWriters.add(spillWriter);
- upstream = spillWriter.getReader(serializerManager);
+ if (numRecords > 0) {
+ // Iterate over the records that have not been returned and spill
them.
+ final UnsafeSorterSpillWriter spillWriter = new
UnsafeSorterSpillWriter(
+ blockManager, fileBufferSizeBytes, writeMetrics, numRecords);
+ spillIterator(upstream, spillWriter);
+ spillWriters.add(spillWriter);
+ upstream = spillWriter.getReader(serializerManager);
+ } else {
+ upstream = null;
Review comment:
There is nothing to spill to disk here, as all records have already been
read, but we still need to free the memory. Before this PR we would keep the
memory allocated.
Added a comment to the code to clarify this.
----------------------------------------------------------------
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
