Re: COMMENT ON INDEX silently fails

Michael Herold <quabla(at)hemio(dot)de> writes:
> The statement
> COMMENT ON INDEX object_name IS '...'
> does not fail if object_name is a table constraint. However, it does
> not seem to have any effect.

Sure it does.

regression=# create table foo (f1 int primary key);
CREATE TABLE
regression=# \d foo
Table "public.foo"
Column | Type | Modifiers
--------+---------+-----------
f1 | integer | not null
Indexes:
"foo_pkey" PRIMARY KEY, btree (f1)

regression=# select * from pg_description where objoid = 'foo_pkey'::regclass;
objoid | classoid | objsubid | description
--------+----------+----------+-------------
(0 rows)

regression=# comment on index foo_pkey is 'test comment';
COMMENT
regression=# select * from pg_description where objoid = 'foo_pkey'::regclass;
objoid | classoid | objsubid | description
--------+----------+----------+--------------
478892 | 1259 | 0 | test comment
(1 row)

I suspect you were expecting the comment to be displayed in some place it
isn't, but that's hard to discuss intelligently when you didn't say where
you expected it to show up. It is there in, eg, \di+ output:

regression=# \di+ foo_pkey
List of relations
Schema | Name | Type | Owner | Table | Size | Description
--------+----------+-------+----------+-------+------------+--------------
public | foo_pkey | index | postgres | foo | 8192 bytes | test comment
(1 row)

regards, tom lane