Skip to content

[SPARK-23349][SQL]ShuffleExchangeExec: Duplicate and redundant type determination for ShuffleManager Object #20530

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wujianping10043419 wants to merge 1 commit into from
Closed

[SPARK-23349][SQL]ShuffleExchangeExec: Duplicate and redundant type determination for ShuffleManager Object #20530

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 3 additions & 6 deletions sql/core/src/main/scala/org/apache/spark/sql/execution/exchange/ShuffleExchangeExec.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -163,13 +163,10 @@ object ShuffleExchangeExec {
// passed instead of directly passing the number of partitions in order to guard against
// corner-cases where a partitioner constructed with `numPartitions` partitions may output
// fewer partitions (like RangePartitioner, for example).
val conf = SparkEnv.get.conf
val shuffleManager = SparkEnv.get.shuffleManager
val sortBasedShuffleOn = shuffleManager.isInstanceOf[SortShuffleManager]
val bypassMergeThreshold = conf.getInt("spark.shuffle.sort.bypassMergeThreshold", 200)
val sortBasedShuffleOn = SparkEnv.get.shuffleManager.isInstanceOf[SortShuffleManager]
if (sortBasedShuffleOn) {
val bypassIsSupported = SparkEnv.get.shuffleManager.isInstanceOf[SortShuffleManager]
if (bypassIsSupported && partitioner.numPartitions <= bypassMergeThreshold) {
val bypassMergeThreshold = SparkEnv.get.conf.getInt("spark.shuffle.sort.bypassMergeThreshold", 200)
if (partitioner.numPartitions <= bypassMergeThreshold) {
// If we're using the original SortShuffleManager and the number of output partitions is
// sufficiently small, then Spark will fall back to the hash-based shuffle write path, which
// doesn't buffer deserialized records.
Expand Down