Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/8354#discussion_r42435971
--- Diff: sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveShim.scala
---
@@ -196,14 +194,8 @@ private[hive] object HiveShim {
if (instance != null) {
instance.asInstanceOf[UDFType]
} else {
- val func = Utils.getContextOrSparkClassLoader
+ Utils.getContextOrSparkClassLoader
.loadClass(functionClassName).newInstance.asInstanceOf[UDFType]
- if (!func.isInstanceOf[UDF]) {
--- End diff --
I guess the problem was that caching of the function instance was causing
things to break? If so, is the right fix to disable caching for _all_ things
which aren't UDFs? Does this risk performance regressions or correctness issues?
I'm not familiar enough with this code to know for sure, so it would be
great if you could comment here to help clear this up.
Also, if it does turn out that removing this caching is the correct
approach, then I think that we may be able to simply remove `instance`, since
it will always be `null` now as far as I can tell.
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
