> On Jul 19, 2025, at 13:15, Zhang Mingli <zmlpostg...@gmail.com> wrote:
>
> Hi,
>
> On Jul 18, 2025 at 22:52 +0800, feichanghong <feichangh...@qq.com>, wrote:
>
> explain select * from (select * from t where a = 1 and b > 1 order by a, b)
> order by a, b limit 1;
> QUERY PLAN
> -----------------------------------------------------------------------------------------
> Limit (cost=366.26..366.27 rows=1 width=8)
> -> Sort (cost=366.26..391.26 rows=9999 width=8)
> Sort Key: t.a, t.b
> -> Index Only Scan using t_a_b_idx on t (cost=0.29..316.27
> rows=9999 width=8)
> Index Cond: ((a = 1) AND (b > 1))
> (5 rows)
> ```
>
> Should we retain the complete `pathkeys` in `Path->pathkeys` for use by the
> upper layers of the subquery, rather than just keeping the portion trimmed by
> `PlannerInfo->query_pathkeys`? I'm not sure if my understanding is correct.
>
> The subquery has a qualifier a = 1 that forms an Equivalence Class (EC) whose
> ec_member contains a constant.
> As a result, subroot->sort_pathkeys doesn't need to include column a.
> However, in the outer query, there are no such qualifiers to form a similar
> EC, and the ORDER BY a, b clause means root->sort_pathkeys requires both
> columns a and b.
> When convert_subquery_pathkeys is called, the subpath lacks the pathkeys for
> column a.
> Furthermore, is there a more efficient way to write this, to avoid the
> `Sort` node mentioned above?
Yes, your understanding is basically consistent with mine.
> A simple solution is to add an EC using a qual:
>
>
>
> EXPLAIN SELECT * FROM (SELECT * FROM t WHERE a = 1 AND b > 1 ORDER BY a, b)
> WHERE a = 1 ORDER BY a, b LIMIT 1;
> QUERY PLAN
> -----------------------------------------------------------------------------------
> Limit (cost=0.29..0.32 rows=1 width=8)
> -> Index Only Scan using t_a_b_idx on t (cost=0.29..316.27 rows=9999 width=8)
> Index Cond: ((a = 1) AND (b > 1))
> (3 rows)
Thank you for your suggestion, this method can address simple subquery
scenarios. However, my situation involves a union all, so it's not possible to
add the corresponding equality qualifier at the top level. The SQL is as
follows:
```sql
explain select a, b from (
(select a, b from t t1 where a > 19000 order by a, b)
union all
(select a, b from t t2 where a = 1 and b > 1 order by a, b)
) t order by a, b limit 1;
```
Currently, I have not found a better way to rewrite this, except by optimizing
this scenario from the pg kernel side.
Best Regards,
Fei Changhong
