Dandandan commented on a change in pull request #1732:
URL: https://github.com/apache/arrow-datafusion/pull/1732#discussion_r797892243
##########
File path: datafusion/src/physical_plan/projection.rs
##########
@@ -185,6 +185,10 @@ impl ExecutionPlan for ProjectionExec {
self.expr.iter().map(|(e, _)| Arc::clone(e)),
)
}
+
+ fn should_repartition_children(&self) -> bool {
+ false
Review comment:
A processing heavy projection (e.g. parsing during cast, complex
case/when statement, etc.) benefits from having multiple partitions, to run in
parallel.
##########
File path: datafusion/src/physical_plan/repartition.rs
##########
@@ -259,6 +259,10 @@ impl ExecutionPlan for RepartitionExec {
fn statistics(&self) -> Statistics {
self.input.statistics()
}
+
+ fn should_repartition_children(&self) -> bool {
+ false
Review comment:
Hash-repartition is quite expensive, so might benefit from being
repartitioned.
--
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]
