Re: 'DROP INDEX' kills stored rpocedures - Mailing list pgsql-general

From scott.marlowe
Subject Re: 'DROP INDEX' kills stored rpocedures
Date
Msg-id Pine.LNX.4.33.0304031050060.19813-100000@css120.ihs.com
Whole thread Raw
In response to 'DROP INDEX' kills stored rpocedures  (Vlad Krupin <vlad@echospace.com>)
Responses Re: 'DROP INDEX' kills stored rpocedures  (Tom Lane <tgl@sss.pgh.pa.us>)
Re: 'DROP INDEX' kills stored rpocedures  (Vlad Krupin <vlad@echospace.com>)
List pgsql-general
On Wed, 2 Apr 2003, Vlad Krupin wrote:

> this is my first post to this mailing list, so if this is the wrong
> list, direct me to the right one please.
> also, please cc: to me because I am not (yet) subscribed.
>
> 1. I understand that working with indexes is outside the scope of
> transaction. That is, if you start a transaction, then drop an index, at
> that very moment, before the transaction is committed, the index will be
> unavailable to any other concurrently running queries. I didn't find
> that in documentation, but a small experiment showed that to be true.

Don't make assumptions like that.  In postgresql, DDL is transactionable.

begin;
drop index test;
create index test on table (field);
commit;

will work just fine.


pgsql-general by date:

Previous
From: Joe Conway
Date:
Subject: Re: Help with array constraints
Next
From: Josh Berkus
Date:
Subject: OSS database needed for testing