[GitHub] [arrow-datafusion] ozankabak commented on a diff in pull request #4885: Do not repartition inputs whose sort order is required

[GitBox <gi...@apache.org>]

ozankabak commented on code in PR #4885:
URL: https://github.com/apache/arrow-datafusion/pull/4885#discussion_r1070081678


##########
datafusion/core/src/physical_optimizer/repartition.rs:
##########
@@ -163,10 +178,34 @@ fn optimize_partitions(
             .children()
             .iter()
             .map(|child| {
+                // does plan itelf (not parent) require its input to
+                // be sorted in some way?
+                let required_input_ordering =
+                    plan_has_required_input_ordering(plan.as_ref());
+
+                let can_reorder_child = if can_reorder {
+                    // parent of `plan` will not use any particular order
+
+                    // if `plan` itself doesn't need order OR
+                    !required_input_ordering ||
+                    // child has no order to preserve
+                        child.output_ordering().is_none()
+                } else {
+                    // parent would like to use the `plan`'s output
+                    // order.
+
+                    // if `plan` doesn't maintain the input order and
+                    // doesn't need the child's output order itself
+                    (!plan.maintains_input_order() &&  !required_input_ordering) ||
+                    // child has no ordering to preserve
+                        child.output_ordering().is_none()
+                };

Review Comment:
   Sure; will take a look, discuss with the team and get back to you.



-- 
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: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org