| From: | Andy S <gatekeeper(dot)mail(at)gmail(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | pgsql-bugs(at)lists(dot)postgresql(dot)org |
| Subject: | Re: BUG #16763: CREATE TABLE IF NOT EXISTS fails with "relation exists" |
| Date: | 2020-12-04 17:43:38 |
| Message-ID: | CAFAcjJOCE=TCeiRX_=2qKzHkLiQjhqmzVk9XeAHEXBDH8Erq3A@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs |
On Fri, Dec 4, 2020 at 8:30 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> IF NOT EXISTS doesn't attempt to be bulletproof: it just checks at the
> start of the command to see if the object name is already there. So
> it's not sufficient to guard concurrent creations. You can call that
> a bug if you like, but it's quite unlikely to change anytime soon.
>
> This is then at least a documentation bug since docs provide less
information for the user then it's needed to properly decide on strategy.
> You might consider using advisory locks [1] to keep your various sessions
> from trying to do this at the exact same time.
>
> Thanks, I'll update the DDLs to wrap the queries emitted with those locks.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Ellis | 2020-12-05 17:30:56 | Small query yields "error: no relation entry for relid 11" |
| Previous Message | Tom Lane | 2020-12-04 17:30:06 | Re: BUG #16763: CREATE TABLE IF NOT EXISTS fails with "relation exists" |