github-actions[bot] commented on code in PR #65005:
URL: https://github.com/apache/doris/pull/65005#discussion_r3503000429


##########
regression-test/suites/query_p0/eager_agg/bilateral_eager_agg.groovy:
##########
@@ -894,7 +895,59 @@ suite("bilateral_eager_agg") {
          GROUP BY t1.k;
      """
 
-    // Reset session variables to defaults
-    sql "SET eager_aggregation_mode = 0;"
-    sql "SET force_eager_agg_hint = '';"
+     sql """
+         SET force_eager_agg_hint = '';
+         set eager_aggregation_mode=1;
+         DROP TABLE IF EXISTS bilateral_left;
+         DROP TABLE IF EXISTS bilateral_right;
+         
+         CREATE TABLE bilateral_left (
+             filter_date DATE NULL,
+             value_date  DATE NULL,
+             flag        BOOLEAN NULL
+         )
+         DUPLICATE KEY(filter_date)
+         DISTRIBUTED BY HASH(filter_date) BUCKETS 1
+         PROPERTIES (
+             "replication_num" = "1"
+         );
+         
+         CREATE TABLE bilateral_right (
+             lower_bound BIGINT NULL,
+             upper_bound BIGINT NULL
+         )
+         DUPLICATE KEY(lower_bound)
+         DISTRIBUTED BY HASH(lower_bound) BUCKETS 1
+         PROPERTIES (
+             "replication_num" = "1"
+         );
+         
+         INSERT INTO bilateral_left VALUES
+             ('2018-01-08', '2019-01-01', FALSE),
+             ('2018-01-08', '2020-01-01', FALSE),
+             ('2018-01-08', '2021-01-01', TRUE),
+             ('2018-01-09', '2022-01-01', TRUE);
+         
+         INSERT INTO bilateral_right VALUES
+             (1, 2),
+             (2, 2),
+             (3, 2),
+             (NULL, 2);
+     """
+
+     order_qt_bilateral_max """
+         SELECT

Review Comment:
   This regression does not appear to exercise the branch this PR changes. The 
new condition in `EagerAggRewriter` only matters when both join children are 
pushed (`leftChildContext` and `rightChildContext` are present) and no count 
slot is produced, but this query's only aggregate is `MAX(CASE ... l.flag ... 
l.value_date ...)`, so the aggregate inputs all come from `bilateral_left`. 
`decideJoinPushSide` will only choose that side, and `MAX` does not request an 
output-count slot, so the new `(leftChildContext && rightChildContext)` branch 
is not covered. Please make the regression use `MAX`/`MIN` aggregates from both 
join children, or otherwise assert a plan shape that proves both sides are 
pushed, so it would fail on the pre-fix code.



##########
regression-test/suites/query_p0/eager_agg/bilateral_eager_agg.groovy:
##########
@@ -894,7 +895,59 @@ suite("bilateral_eager_agg") {
          GROUP BY t1.k;
      """
 
-    // Reset session variables to defaults
-    sql "SET eager_aggregation_mode = 0;"
-    sql "SET force_eager_agg_hint = '';"
+     sql """
+         SET force_eager_agg_hint = '';
+         set eager_aggregation_mode=1;
+         DROP TABLE IF EXISTS bilateral_left;
+         DROP TABLE IF EXISTS bilateral_right;

Review Comment:
   `git diff --check` fails on this new block because the blank separator lines 
at 903, 914, 924, and 930 contain spaces. Please remove the trailing whitespace 
so the PR passes the repository's standard diff hygiene check.



-- 
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]

Reply via email to