On Fri, Feb 15, 2013 at 04:06:12PM -0600, Dave Rolsky wrote:
> On Fri, 15 Feb 2013, Bruce Momjian wrote:
>
> >On Wed, Feb 13, 2013 at 08:22:43PM +0000, autarch@urth.org wrote:
> >>The following bug has been logged on the website:
> >>
> >>Bug reference: 7873
> >>Logged by: Dave Rolsky
> >>Email address: autarch@urth.org
> >>PostgreSQL version: 9.2.3
> >>Operating system: Linux
> >>Description:
> >>
> >>When you pass the --clean option to pg_restore it tries to drop tables
> >>without checking if they exist. This results in lots of error output. If
> >>you're running pg_restore via an automated process it's very hard to
> >>distinguish between these "ok" errors and real errors.
> >>
> >>It should be using "DROP TABLE IF EXISTS" and the equivalent for
> >>constraints.
> >
> >Well, I think the question is whether you want error feedback for things
> >that don't exist. I don't really know the answer.
>
> Fair enough. It should probably an option to add "if exists", at
> least. I can't imagine I'm the only using this tool to ship database
> updates around to different machines, some of which may not have new
> tables. I'd really like to be able to know when the restore fails
> versus when it succeeds but is noisy.
All I can say is I don't remember anyone asking for this in the past.
--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +