cloud-fan commented on code in PR #36295:
URL: https://github.com/apache/spark/pull/36295#discussion_r903765555
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/V2ScanRelationPushDown.scala:
##########
@@ -407,7 +407,66 @@ object V2ScanRelationPushDown extends Rule[LogicalPlan]
with PredicateHelper wit
case other => (other, false)
}
- def pushDownLimits(plan: LogicalPlan): LogicalPlan = plan.transform {
+ private def pushDownOffset(
+ plan: LogicalPlan,
+ offset: Int): Boolean = plan match {
+ case sHolder: ScanBuilderHolder =>
+ val isPushed = PushDownUtils.pushOffset(sHolder.builder, offset)
+ if (isPushed) {
+ sHolder.pushedOffset = Some(offset)
+ }
+ isPushed
+ case p: Project =>
+ pushDownOffset(p.child, offset)
+ case _ => false
+ }
+
+ def pushDownLimitAndOffset(plan: LogicalPlan): LogicalPlan = plan.transform {
+ case offset @ LimitAndOffset(limit, offsetValue, child) =>
+ val (newChild, canRemoveLimit) = pushDownLimit(child, limit)
+ if (canRemoveLimit) {
+ // If we can remove limit, it indicates data source only have one
partition.
Review Comment:
I don't think it applies to all data sources. Let's be more general here:
```
// Try to push down OFFSET only if the LIMIT operator has been pushed and
can be removed.
```
--
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]
