Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/11868#discussion_r59145545
--- Diff: sql/catalyst/src/main/scala/org/apache/spark/sql/Row.scala ---
@@ -402,11 +402,13 @@ trait Row extends Serializable {
return false
}
case f1: Float if java.lang.Float.isNaN(f1) =>
- if (!o2.isInstanceOf[Float] || !
java.lang.Float.isNaN(o2.asInstanceOf[Float])) {
+ if (!(o2.isInstanceOf[Float] &&
java.lang.Float.isNaN(o2.asInstanceOf[Float]) ||
--- End diff --
Hm, in general NaN never equals NaN. There might be some reason to treat it
differently here. On the one hand I tend to agree with this change anyway, on
the grounds that it implements something like automatic promotion in
Scala/Java. But clearly we're already not implementing the language's
semantics, and trying to achieve something more like bitwise-equal semantics.
In that case this wouldn't quite be right. Ask the author of the original
change why it was made?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
