[
https://issues.apache.org/jira/browse/SPARK-43098?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Josh Rosen updated SPARK-43098:
-------------------------------
Labels: correctness (was: )
> Should not handle the COUNT bug when the GROUP BY clause of a correlated
> scalar subquery is non-empty
> -----------------------------------------------------------------------------------------------------
>
> Key: SPARK-43098
> URL: https://issues.apache.org/jira/browse/SPARK-43098
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Affects Versions: 3.2.0
> Reporter: Jack Chen
> Assignee: Jack Chen
> Priority: Major
> Labels: correctness
> Fix For: 3.4.1, 3.5.0
>
>
> From [~allisonwang-db] :
> There is no COUNT bug when the correlated equality predicates are also in the
> group by clause. However, the current logic to handle the COUNT bug still
> adds default aggregate function value and returns incorrect results.
>
> {code:java}
> create view t1(c1, c2) as values (0, 1), (1, 2);
> create view t2(c1, c2) as values (0, 2), (0, 3);
> select c1, c2, (select count(*) from t2 where t1.c1 = t2.c1 group by c1) from
> t1;
> -- Correct answer: [(0, 1, 2), (1, 2, null)]
> +---+---+------------------+
> |c1 |c2 |scalarsubquery(c1)|
> +---+---+------------------+
> |0 |1 |2 |
> |1 |2 |0 |
> +---+---+------------------+
> {code}
>
> This bug affects scalar subqueries in RewriteCorrelatedScalarSubquery, but
> lateral subqueries handle it correctly in DecorrelateInnerQuery. Related:
> https://issues.apache.org/jira/browse/SPARK-36113
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
