mingmwang commented on code in PR #4043:
URL: https://github.com/apache/arrow-datafusion/pull/4043#discussion_r1012379795
##########
datafusion/core/src/physical_plan/aggregates/mod.rs:
##########
@@ -255,25 +276,58 @@ impl ExecutionPlan for AggregateExec {
/// Get the output partitioning of this plan
fn output_partitioning(&self) -> Partitioning {
- self.input.output_partitioning()
+ match &self.mode {
+ AggregateMode::Partial => {
+ // Partial Aggregation will not change the output partitioning
but need to respect the Alias
+ let input_partition = self.input.output_partitioning();
+ match input_partition {
+ Partitioning::Hash(exprs, part) => {
+ let normalized_exprs = exprs
+ .into_iter()
+ .map(|expr| {
+ normalize_out_expr_with_alias_schema(
+ expr,
+ &self.alias_map,
+ &self.schema,
+ )
+ })
+ .collect::<Vec<_>>();
+ Partitioning::Hash(normalized_exprs, part)
+ }
+ _ => input_partition,
+ }
+ }
+ // Final Aggregation's output partitioning is the same as its real
input
+ _ => self.input.output_partitioning(),
+ }
}
+ // TODO check the output ordering of AggregateExec
Review Comment:
Yes, I will remove it. The AggregateExec can not keep the ordering.
--
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]
