Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/23176#discussion_r237397442
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/predicates.scala
---
@@ -367,11 +367,29 @@ case class InSet(child: Expression, hset: Set[Any])
extends UnaryExpression with
}
@transient lazy val set: Set[Any] = child.dataType match {
- case _: AtomicType => hset
+ case t: AtomicType if !t.isInstanceOf[BinaryType] => hset
case _: NullType => hset
case _ =>
+ val ord = TypeUtils.getInterpretedOrdering(child.dataType)
+ val ordering = if (hasNull) {
+ new Ordering[Any] {
+ override def compare(x: Any, y: Any): Int = {
--- End diff --
InSet overrides nullSafeEval, and for codegen we look into `set` only if
`!ev.isNull`, so I think we only need to consider the case one side is null.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
