| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Grzegorz(dot)Przezdziecki(at)PolskieSklepy(dot)pl |
| Cc: | somee(at)poczta(dot)onet(dot)pl, pgsql-general(at)postgresql(dot)org |
| Subject: | Re: change attnum in pg_catalog.pg_attribute |
| Date: | 2005-04-05 18:41:32 |
| Message-ID: | 26417.1112726492@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Grzegorz =?iso-8859-2?q?Prze=BCdziecki?= <Grzegorz(dot)Przezdziecki(at)PolskieSklepy(dot)pl> writes:
>> It won't work and will likely crash the backend.
> Even if I will do like this and for ever??
> 1. attnum will be unique and greater than 0
> 2. no holes in attnum numeric
> 3. Make change
> 4. restart postgresql
Yes, unless you also TRUNCATE the table, because attnum ordering tells
about the physical data layout within the table rows. Furthermore,
altering only pg_attribute is unlikely to represent a complete update
of the catalogs --- attnums also show up in pg_index, pg_attrdef,
pg_depend, pg_rewrite if there are any rules or views mentioning the
columns, and probably some other places I forgot.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Grzegorz Przeździecki | 2005-04-05 18:44:24 | Re: change attnum in pg_catalog.pg_attribute |
| Previous Message | David Rysdam | 2005-04-05 18:29:55 | Re: Check for existence of index |