| From: | shihao zhong <zhong950419(at)gmail(dot)com> |
|---|---|
| To: | Corey Huinker <corey(dot)huinker(at)gmail(dot)com> |
| Cc: | Michael Paquier <michael(at)paquier(dot)xyz>, pgsql-hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>, tomas(at)vondra(dot)me |
| Subject: | Re: [PG19] pg_dump --statistics restores extended statistics by attribute number |
| Date: | 2026-10-16 03:05:06 |
| Message-ID: | CAGRkXqQGYqpFPBQ6TJCXe8fD-6w4de=OFgNLacscXL7m+3ykLA@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi Corey,
v2 does not remap anything. It only rejects the stats when the column
names do not match, so the table then shows up in vacuumdb
--missing-stats-only, which is the path you describe. It is 79 lines
plus tests.
The silent case has no dropped column. It is an inheritance child with
its own column order.
create table p (a int, b int, z int);
create table ch (z int, b int, a int);
alter table ch inherit p;
create statistics ch_s on z, b, a from ch;
After a restore the stats of z land on a, with no warning. A list of
tables with dropped columns does not find it.
Michael, is v2 the rejection you meant, or is the warning we get today
enough?
Thanks,
Shihao
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Chao Li | 2026-10-16 03:10:13 | Re: [PATCH] Doc fix for autovacuum_max_parallel_workers |
| Previous Message | 杨伯宇 (长堂) | 2026-10-16 03:01:59 | Re:logical decoding: skip unnecessary snapshot distribution. |