Content-Type: text/plain;
  charset="iso-8859-1"
From: Richard Huxton <dev@archonet.com>
Organization: Archonet Ltd
To: "Ries van Twisk" <ries@jongert.nl>, <pgsql-sql@postgresql.org>
Subject: Re: Checking my HD space  (noarchive)
Date: Tue, 4 Mar 2003 09:20:52 +0000
User-Agent: KMail/1.4.3
References: <000201c2e1a3$180cdc10$f100000a@IT001>
In-Reply-To: <000201c2e1a3$180cdc10$f100000a@IT001>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Message-Id: <200303040920.52946.dev@archonet.com>

On Monday 03 Mar 2003 4:36 pm, Ries van Twisk wrote:
> Hi All,
>
> My bare text table is in size around 6Mb while the footprint of my databa=
se
> on HD (using du) is around 24Mb (after a vacuum -f)
> I have normalized the tables on places where I could do it.
> For your reference it's a squid2.2 access.log textfile.
> How can I check witch tables/indexes takes up so much space?

Your database files have a name equal to the object's oid. So if you have a=
=20
table with oid=3D1234 look for a file named "1234". You can figure out oids=
=20
manually, but there's a tool "oid2name" in the contrib folder of the source=
=20
distribution which can help you.

A 3x increase over a raw text file isn't unexpected once you allow for=20
structure overheads, indexes etc.

> And I did understand that postgresql can compress text but it seems that
> this is not the default case.
> can anybody tell me somthing about it and/or point my to a url??

I think you're talking about TOAST - check the manuals and techdocs, there'=
s a=20
section about it somewhere. It's automatic for large textfields, and was=20
originally developed to work around PG's old 8k rowsize limit.

HTH

--=20
  Richard Huxton