[
https://issues.apache.org/jira/browse/CALCITE-5043?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17513332#comment-17513332
]
Xurenhe commented on CALCITE-5043:
----------------------------------
It may be related to
[CALCITE-4512|https://issues.apache.org/jira/browse/CALCITE-4512]
But, I don't know the difference between
[CALCITE-4512|https://issues.apache.org/jira/browse/CALCITE-4512] and this.
I may miss some important things.
> GROUP BY alias or SELECT field should be determined by using sql dialect when
> alias and field are same
> ------------------------------------------------------------------------------------------------------
>
> Key: CALCITE-5043
> URL: https://issues.apache.org/jira/browse/CALCITE-5043
> Project: Calcite
> Issue Type: Bug
> Components: core
> Affects Versions: 1.29.0
> Environment: jdk8
> Reporter: yanjing.wang
> Assignee: yanjing.wang
> Priority: Major
> Fix For: 1.31.0
>
>
> {noformat}
> select length(ename) as ename from emp group by ename{noformat}
> PostgreSQL: group by empno will be expanded to group by emp.empno.
> BigQuery: group by empno will be expanded to group by length(ename).
> MySQL ONLY_FULL_GROUP_BY mode(default mode): as PostgreSQL.
> MySQL extension: as PostgreSQL.
> {noformat}
> select empno + deptno as empno from emp group by empno{noformat}
> PostgreSQL: throws column "emp.empno" must appear in the GROUP BY clause or
> be used in an aggregate function.
> BigQuery: group by empno will be expanded to group by empno + deptno.
> MySQL ONLY_FULL_GROUP_BY mode(default mode): throws as PostgreSQL.
> MySQL extension: group by empno will be expanded to group by emp.empno.
> {noformat}
> select a.deptno + b.deptno as deptno from emp a , dept b group by deptno
> {noformat}
> PostgreSQL: throws column reference "deptno" is ambiguous.
> BigQuery: group by deptno will be expanded to group by a.deptno + b.deptno.
> MySQL ONLY_FULL_GROUP_BY mode(default mode): throws as PostgreSQL.
> MySQL extension: as BigQuery.
> We need support expanding GROUP BY alias or SELECT field according to the sql
> dialect.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
