回复： BUG #18568: BUG: Result wrong when do group by on partition table!

PartCollMatchesExprColl()  is not strict enough.
If partcoll == InvalidOid and groupcoll != InvalidOid, PartCollMatchesExprColl() return true always.
Just determine whether groupcoll is equal to partcoll, like this "partcoll == groupcoll".




We cannot delete "if (IsA(groupexpr, RelableTyple) " branch, 
becasuse if groupexpr is RelabelType and  partcoll equal to groupcoll,  the "equal(groupexpr, partexpr) && PartKeyCollMatchesExprColl(partcoll, groupexpr_coll)" condition return false!!!
This is not what we expect.






We could rename "groupexpr_coll" to groupcoll, it looks more elegant.






狂奔的蜗牛
1105066510@qq.com



 




------------------ 原始邮件 ------------------
发件人:                                                                                                                        "Tender Wang"                                                                                    <tndrwang@gmail.com&gt;;
发送时间:&nbsp;2024年8月7日(星期三) 下午2:57
收件人:&nbsp;"狂奔的蜗牛"<1105066510@qq.com&gt;;
抄送:&nbsp;"pgsql-bugs"<pgsql-bugs@lists.postgresql.org&gt;;"aleksander"<aleksander@timescale.com&gt;;
主题:&nbsp;Re: BUG #18568: BUG: Result wrong when do group by on partition table!






狂奔的蜗牛 <1105066510@qq.com&gt; 于2024年8月7日周三 13:35写道：

this case will enter `if (IsA(groupexpr, RelabelType))` branch.&nbsp;


We set "C" as groupkey's collation, and it's not equal to column's.
In transformGroupClause(), syntax `COLLATE xxx` will create CollateExpr node, and CollateExpr's arg is Var node.
And then, planner will call&nbsp;eval_const_expressions_mutator() to transform CollateExpr to RelableType if CollateExpr-&gt;collOid not equal to Collate-&gt;arg's collation.


Yech. Thanks for the explanation.
Because exprCollation() return the resultcollid if the nodetag is Relabletype, the expr() &amp;&amp; PartKeyCollMatchesExprColl() can handle this situation.
So we don't need "if (IsA(groupexpr, RelableTyple) " this branch.
&nbsp;




About V3 patch, PartCollMatchesExprColl() may be not&nbsp;suitable, because collation of partkey must be equal to groupkey's, even though they are all InvalidOid.


I'm not sure about what you said.&nbsp; I uesd it because partition prune do this way. I keep it temporarily.&nbsp;


The test case in v3 patch only has EXPLAIN statement. I add SELECT statement and make the result more stable according to Aleksander advices.

-- 
Tender Wang