agrawaldevesh commented on a change in pull request #29304:
URL: https://github.com/apache/spark/pull/29304#discussion_r464165213
##########
File path:
sql/core/src/main/scala/org/apache/spark/sql/execution/joins/HashedRelation.scala
##########
@@ -327,11 +327,27 @@ private[joins] object UnsafeHashedRelation {
// Create a mapping of buildKeys -> rows
val keyGenerator = UnsafeProjection.create(key)
var numFields = 0
+ val nullPaddingCombinations: Seq[UnsafeProjection] = if (isNullAware) {
+ // C(numKeys, 0), C(numKeys, 1) ... C(numKeys, numKeys - 1)
+ // In total 2^numKeys - 1 records will be appended.
+ key.indices.flatMap { n =>
+ key.indices.combinations(n).map { combination =>
+ // combination is Seq[Int] indicates which key should be replaced to
null padding
+ val exprs = key.indices.map { index =>
+ if (combination.contains(index)) {
+ Literal.create(null, key(index).dataType)
Review comment:
Makes sense and great observation. Worth explaining this in the comments
and adding a test for this.
I am wondering if you still have a test suite that compares the old-school
BNLJ results vs with this optimization ? This would be a great test case.
----------------------------------------------------------------
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]
