pg.ddx.io  pgsql-admin@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Chris Hoover <revoohc@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Alvaro Herrera <alvherre@commandprompt.com>
Cc: pgsql-admin@postgresql.org Admin <pgsql-admin@postgresql.org>
Subject: Re: 8.2.4 Won't Build 8.1 Functional Indexes
Date: Wed, 6 Jun 2007 22:16:35 -0400
Message-ID: <1d219a6f0706061916k73deaa6bv2dd9e800e77a6f24@mail.gmail.com> (raw)
In-Reply-To: <17138.1181173286@sss.pgh.pa.us>
References: <1d219a6f0706061112p355fc6cexb30a6b4b4bf187c3@mail.gmail.com>
	<20070606195459.GH11800@alvh.no-ip.org>
	<17138.1181173286@sss.pgh.pa.us>

Well, the one index:
CREATE INDEX acceptedbilling_to_date_accepted_billing_dt_idx
  ON acceptedbilling
  USING btree
  (to_date(accepted_billing_dt::text, 'yyyymmdd'::text));.

Reject:
ERROR: functions in index expression must be marked IMMUTABLE
SQL state: 42P17

Is actually a date stored in a varchar field (stupid I know, but it's a
continual fight between the me (the dba), management, and our COBOL
programmers).  I was indexing it as a date so that I can treat it as a date
in some of the non-COBOL programs/utilities that we have written.

The second index is rather stupid, it was an early index before I figured
out how to split a timestamp.

Anyway, is there a way to make the first index work?  Otherwise we end up
with a seq scan on our billing table which is very painful.

Thanks,

Chris


On 6/6/07, Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> Alvaro Herrera <alvherre@commandprompt.com> writes:
> > You can't do this because to_date and other functions are not immutable.
> > 8.2 seems to be more picky about this -- the date conversions of
> > timestamptz columns are dependent on the current timezone.
>
> The reason 8.2 is more picky is that the function is less immutable
> thanks to the addition of locale-dependent functionality:
> http://archives.postgresql.org/pgsql-committers/2006-11/msg00264.php
>
> I gather that the underlying column is timestamp without tz, or it would
> never have worked in 8.1 either.  That being the case, these index
> definitions seem pretty darn stupid anyway --- why aren't you just
> indexing on date_trunc or a plain cast to date?
>
>                         regards, tom lane
>

view thread (6+ messages)  latest in thread

Message-ID: <1d219a6f0706061916k73deaa6bv2dd9e800e77a6f24@mail.gmail.com>
Permalink:  ../1d219a6f0706061916k73deaa6bv2dd9e800e77a6f24@mail.gmail.com/
Also on:    postgresql.org/message-id/1d219a6f0706061916k73deaa6bv2dd9e800e77a6f24@mail.gmail.com

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: pgsql-admin@postgresql.org
  Cc: revoohc@gmail.com, tgl@sss.pgh.pa.us, alvherre@commandprompt.com
  Subject: Re: 8.2.4 Won't Build 8.1 Functional Indexes
  In-Reply-To: <1d219a6f0706061916k73deaa6bv2dd9e800e77a6f24@mail.gmail.com>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox