cfmcgrady commented on code in PR #37439:
URL: https://github.com/apache/spark/pull/37439#discussion_r952211694


##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/UnwrapCastInBinaryComparison.scala:
##########
@@ -155,7 +155,16 @@ object UnwrapCastInBinaryComparison extends 
Rule[LogicalPlan] {
       list.foreach {
         case lit @ Literal(null, _) => nullList += lit
         case lit @ NonNullLiteral(_, _) =>
-          unwrapCast(EqualTo(in.value, lit)) match {
+          val originalEqualTo = EqualTo(in.value, lit)
+          unwrapCast(originalEqualTo) match {
+            // the function `unwrapCast` may returns original expression when 
the literal can not
+            // cast to fromType, for instance: (the boundreference is of type 
DECIMAL(5,2))
+            //     CAST(boundreference() AS DECIMAL(10,4)) = 123456.1234BD
+            // Due to `cast(lit, fromExp.dataType) == null` we simply return
+            // `falseIfNotNull(fromExp)`.
+            case equalTo: EqualTo if equalTo == originalEqualTo &&
+              Cast(lit, fromExp.dataType).eval() == null =>

Review Comment:
   if `Cast(lit, fromExp.dataType).eval()` is not equal to `null`, then the 
return expression of `unwrapCast` should be either `EqualTo(fromExp, lit)` or 
`falseIfNotNull(fromExp)` which was handled by L168-L169.
   
   
https://github.com/apache/spark/blob/5a71a7f7b5c1762677ddbfe39a7c35d645c25e94/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/UnwrapCastInBinaryComparison.scala#L305-L346
   
   > we may have a return value from L322, L331 or L340.



-- 
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]

Reply via email to