On Mon, Apr 9, 2012 at 3:06 PM, Lukas Eklund <lukas@xxxxxxxxx> wrote: > On Mon, Apr 9, 2012 at 2:51 PM, Merlin Moncure <mmoncure@xxxxxxxxx> wrote: >> On Mon, Apr 9, 2012 at 9:57 AM, Lukas Eklund <lukas@xxxxxxxxx> wrote: >>> Is it safe to resize a numeric column using in pg_attribute without >>> altering the table? >>> >> >> One thing that's pretty weird is that dependent view columns keep the >> old typmod. That might cause some issues with things that depend on >> that value, for example client side tools. But you can always change >> those as well. >> >> I think this might work because the numeric storage doesn't change for >> this case. But it's still quite dangerous and avoiding view >> re-recreation is a pretty weak justification to go be updating system >> catalogs. You'd be better off spending some time developing a script >> to recreate views. The only reason why I would personally be >> considering this would be to avoid issues stemming from having to take >> out a lock on the table and I would be testing very, very carefully if >> I did so. >> >> merlin > > Ah. I must have not noticed that the typmod for views is not inherited > automatically. I'm okay with developing a script to recreate the 15 or > so views the depend on that table. What I'm trying to avoid is locking > that table for a substantial amount of time. Thanks for the advice! np -- note 9.2 will be much smarter about this: http://archives.postgresql.org/message-id/20120101003619.GA4395@xxxxxxxxxxxxxxxxxxxx merlin -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
