Re: TOAST compression

From: "Luke Lonergan" <llonergan(at)greenplum(dot)com>
To: "Josh Berkus" <josh(at)agliodbs(dot)com>, pgsql-hackers(at)postgresql(dot)org
Cc: "Skype Technologies OY" <hannu(at)skype(dot)net>, "Jim C(dot) Nasby" <jnasby(at)pervasive(dot)com>, "Neil Conway" <neilc(at)samurai(dot)com>
Subject: Re: TOAST compression
Date: 2006-02-27 06:15:06
Message-ID: C027D8EA.1DE01%llonergan@greenplum.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Josh,

On 2/26/06 8:04 PM, "Josh Berkus" <josh(at)agliodbs(dot)com> wrote:

> Check out SET STORAGE.

I just altered the MIVP data generator in Bizgres MPP to produce the usual
15 column table but with a 6K row size. You'd only expect a few tens of
bytes variance around the 6K, and the data is randomly chosen words from a
seed file of about 14.5K.

I generated a file with 100,000 rows, total size is 600MB. My tests are
with Bizgres MPP 2.1 GA (based on Postgres 8.1.3) on 4 machines with 8
primary segments.

So, with a column storage type of PLAIN, here's the filesystem resident
size:
mpptestdb=# select relname,8*relpages/128 as MB from pg_class where
relname='bigtable1';;
relname | mb
-----------+-----
bigtable1 | 789

And here are a couple of selects (obviously coming from the 64GB of RAM):
mpptestdb=# select count(*) from bigtable1;
count
--------
100000
(1 row)

Time: 45.685 ms
mpptestdb=# select count(*) from bigtable1;
count
--------
100000
(1 row)

Time: 25.024 ms

Here it is with the default EXTENDED:
mpptestdb=# select relname,8*relpages/128 as MB from pg_class where
relname='bigtable1';
relname | mb
-----------+----
bigtable1 | 20

Now I'm stuck - I can get the toast table OID, but my attempts to find out
it's size are futile.

Suggestions?

- Luke

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Luke Lonergan 2006-02-27 06:47:08 Re: TOAST compression
Previous Message Josh Berkus 2006-02-27 05:03:58 Re: constraints and sql92 information_schema compliance