Vertical Partitioning with TOAST

From: Junji TERAMOTO <teramoto(dot)junji(at)lab(dot)ntt(dot)co(dot)jp>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Vertical Partitioning with TOAST
Date: 2005-12-01 08:59:08
Message-ID: 438EBB5C.9070304@lab.ntt.co.jp
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi all,

I wrote a experimental patch for a vertical partitioning
function.

I decided to use the code of TOAST to create the function
easily. In a word, the row that the user specified is forcedly
driven out with TOAST.

The performance gain of 10% was seen by driving out c_data of the
customer table in the DBT-2 benchmark in our environment.

The mechanism of TOAST is an overdesigned system to use it for a
vertical partitioning. Because the overhead of processing is large,
the performance might down according to the environment.

There are seriously a lot of things that should be considered if
a vertical partitioning is mounted.
For instance, TOAST index is omitted, and ctid is used for link.

Your comments are welcome. Thanks.

---
How To Use
---
Use "ALTER TABLE" command.
http://www.postgresql.org/docs/8.1/static/sql-altertable.html

ALTER TABLE name ALTER COLUMN column SET STRAGE FORCEEXTERNAL;

I do not understand whether "FORCEEXTERNAL" is an appropriate
word. Please teach when there is a better word...

--
Junji Teramoto

Attachment Content-Type Size
patch_forceexternal.diff text/plain 4.6 KB

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Richard Huxton 2005-12-01 10:06:12 Re: [HACKERS] Upcoming PG re-releases
Previous Message Tatsuo Ishii 2005-12-01 07:13:05 Re: A couple of proposed pgbench changes