pedrumj2 commented on code in PR #13016:
URL: https://github.com/apache/gluten/pull/13016#discussion_r4020498439
##########
backends-velox/src/main/scala/org/apache/spark/sql/expression/UDFResolver.scala:
##########
@@ -338,6 +341,35 @@ object UDFResolver extends Logging {
.toBoolean
}
+ /**
+ * One Spark function per loaded UDF whose name contains no dot. A dotted
name is a Hive UDF class
+ * name, which VeloxHiveUDFTransformer already resolves, so it is skipped
here.
+ *
+ * A name is also skipped when it collides with a Spark built-in: the names
are unqualified, so
+ * injecting one would redirect that built-in to a native implementation
with possibly different
+ * semantics for every query on the session.
+ */
+ def getFunctionDescriptions: Seq[FunctionDescription] = {
+ val (shadowing, injectable) = UDFNames.toSeq
+ .filterNot(_.contains("."))
+ .sorted
+ .partition(name =>
FunctionRegistry.builtin.functionExists(FunctionIdentifier(name)))
Review Comment:
Updated
--
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]
