LuciferYang commented on a change in pull request #29434:
URL: https://github.com/apache/spark/pull/29434#discussion_r471898154
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/util/ArrayBasedMapData.scala
##########
@@ -136,7 +136,8 @@ object ArrayBasedMapData {
keys.zip(values).toMap
}
- def toScalaMap(keys: Seq[Any], values: Seq[Any]): Map[Any, Any] = {
+ def toScalaMap(keys: scala.collection.Seq[Any],
Review comment:
Yep, I think now it must be `s.c.Seq` unless we accept `case
ArrayType(et, nullable)` branch of `RowEncoder.deserializerFor` method has a
2.13-only branch to use `immutable.ArraySeq.unsafeMake` create a
`immutable.Seq`. We can't use uniformly` immutable.ArraySeq ` because they are
in different packages with different Scala version, I haven't found a perfect
replacement of `WrappedArray.make` method yet.
Now `WrappedArray.make` create a `mutable.ArraySeq` instance in Scala 2.13 ,
so the `toScalaMap` method input args are `mutable.ArraySeq` not
`immutable.ArraySeq` and `Seq` alias representative `immutable.Seq` in Scala
2.13, this not match. On the other hand, `toScalaMap` method is called in
reflection mode, inconvenient to add 'toSeq'.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
