pan3793 commented on code in PR #55927:
URL: https://github.com/apache/spark/pull/55927#discussion_r3393079883
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala:
##########
@@ -967,6 +967,20 @@ object SQLConf {
.checkValue(_ > 0, "The value of spark.sql.shuffle.partitions must be
positive")
.createWithDefault(200)
+ val SHUFFLE_SPREAD_NULL_JOIN_KEYS_ENABLED =
+ buildConf("spark.sql.shuffle.spreadNullJoinKeys.enabled")
+ .doc("When true, Spark may spread rows with NULL equi-join keys across
shuffle partitions " +
+ "for shuffled LEFT, RIGHT, and FULL OUTER equi-joins on nullable keys
to reduce " +
+ "shuffle skew. Null-aware join output partitioning does not satisfy a
strict " +
+ "ClusteredDistribution, so downstream grouping, windowing, or
equi-joins may require " +
+ "an extra shuffle. If one input is already hash partitioned, only the
other input may " +
+ "be reshuffled into the null-aware layout, so the pre-shuffled input
can keep its NULL " +
+ "skew.")
+ .version("4.1.0")
Review Comment:
should it be 4.3.0?
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
