| From: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> | 
|---|---|
| To: | Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp> | 
| Cc: | pgsql-hackers(at)postgresql(dot)org, Teodor Sigaev <teodor(at)sigaev(dot)ru>, Alexander Korotkov <aekorotkov(at)gmail(dot)com> | 
| Subject: | Re: Fix error message when trying to alter statistics on included column | 
| Date: | 2018-07-17 00:04:34 | 
| Message-ID: | 20180717000433.ljup6n5mgxlgmeev@alvherre.pgsql | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-hackers | 
On 2018-Jun-28, Yugo Nagata wrote:
> According to the error message, it is not allowed to alter statistics on
> included column because this is "non-expression column".
> 
>  postgres=# create table test (i int, d int);
>  CREATE TABLE
>  postgres=# create index idx on test(i) include (d);
>  CREATE INDEX
>  postgres=# alter index idx alter column 2 set statistics 10;
>  ERROR:  cannot alter statistics on non-expression column "d" of index "idx"
>  HINT:  Alter statistics on table column instead.
> 
> However, I think this should be forbidded in that this is not a key column 
> but a included column. Even if we decide to support expressions in included
> columns in future, it is meaningless to do this because any statistics on 
> included column is never used by the planner.
I agree with this reasoning, so I pushed this patch.  Thanks!  I added a
couple of lines in the regress file for this feature also.
Teodor, Alexander, now would be the time to express dissent :-)
-- 
Álvaro Herrera                https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | David Rowley | 2018-07-17 00:21:00 | Re: Internal error XX000 with enable_partition_pruning=on, pg 11 beta1 on Debian | 
| Previous Message | Michael Paquier | 2018-07-17 00:01:29 | Re: [HACKERS] WAL logging problem in 9.4.3? |