HyukjinKwon commented on a change in pull request #23746: [SPARK-26761][SQL][R]
Vectorized R gapply() implementation
URL: https://github.com/apache/spark/pull/23746#discussion_r255854287
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/object.scala
##########
@@ -437,19 +438,30 @@ object FlatMapGroupsInR {
groupingAttributes: Seq[Attribute],
dataAttributes: Seq[Attribute],
child: LogicalPlan): LogicalPlan = {
- val mapped = FlatMapGroupsInR(
- func,
- packageNames,
- broadcastVars,
- inputSchema,
- schema,
- UnresolvedDeserializer(keyDeserializer, groupingAttributes),
- UnresolvedDeserializer(valueDeserializer, dataAttributes),
- groupingAttributes,
- dataAttributes,
- CatalystSerde.generateObjAttr(RowEncoder(schema)),
- child)
- CatalystSerde.serialize(mapped)(RowEncoder(schema))
+ if (SQLConf.get.arrowEnabled) {
+ FlatMapGroupsInRWithArrow(
+ func,
+ packageNames,
+ broadcastVars,
+ inputSchema,
+ schema.toAttributes,
+ UnresolvedDeserializer(keyDeserializer, groupingAttributes),
+ groupingAttributes,
+ child)
Review comment:
Yea, that's actually related with `gapplyCollect`. I have a neat idea to
avoid unnecessary de/encode. I will make a fix after this PR, when I fix
`gapplyCollect` against https://issues.apache.org/jira/browse/SPARK-26858.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
