mustafasrepo commented on code in PR #6734:
URL: https://github.com/apache/arrow-datafusion/pull/6734#discussion_r1238673844
##########
datafusion/core/tests/sqllogictests/test_files/groupby.slt:
##########
@@ -2076,18 +2076,18 @@ Projection: annotated_data_infinite2.a,
annotated_data_infinite2.b, FIRST_VALUE(
----TableScan: annotated_data_infinite2 projection=[a, b, c]
physical_plan
ProjectionExec: expr=[a@0 as a, b@1 as b,
FIRST_VALUE(annotated_data_infinite2.c) ORDER BY [annotated_data_infinite2.a
DESC NULLS FIRST]@2 as first_c]
---AggregateExec: mode=Single, gby=[a@0 as a, b@1 as b],
aggr=[FIRST_VALUE(annotated_data_infinite2.c)], ordering_mode=FullyOrdered
+--AggregateExec: mode=Single, gby=[a@0 as a, b@1 as b],
aggr=[LAST_VALUE(annotated_data_infinite2.c)], ordering_mode=FullyOrdered
----CsvExec: file_groups={1 group:
[[WORKSPACE_ROOT/datafusion/core/tests/data/window_2.csv]]}, projection=[a, b,
c], infinite_source=true, output_ordering=[a@0 ASC NULLS LAST, b@1 ASC NULLS
LAST, c@2 ASC NULLS LAST], has_header=true
query III
SELECT a, b, FIRST_VALUE(c ORDER BY a DESC) as first_c
FROM annotated_data_infinite2
GROUP BY a, b
----
-0 0 0
-0 1 25
-1 2 50
-1 3 75
+0 0 24
Review Comment:
The reason is that source is ordered by `a ASC`. Since aggregation requires
ordering by `a DESC` it swaps the aggregation to resolve requirement. However,
since requirement column is among the group by column (`group by a, b`), we are
sure that each group will involve unique `a` values. Hence in practice `a DESC`
and `a ASC` is already satisfied, since each group will involve same `a`
values. In short, previous plan and this plan are both valid.
--
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]
