Date: Sun, 2 Apr 2006 12:31:49 +0400
From: =?utf-8?B?0JrQu9GO0YfQvdC40LrQvtCyINCQLtChLg==?= <alexs@analytic.mv.ru>
To: pgsql-performance@postgresql.org
Subject: Trigger vs Rule
Message-ID: <20060402083149.GA62166@mail.analytic.mv.ru>
Mail-Followup-To: pgsql-performance@postgresql.org
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
User-Agent: Mutt/1.4.2.1i

Hi all.

There are two tables:

create table device_types (
id int,
name varchar
);
about 1000 rows

create table devices (
id int,
type int REFERENCES device_types(id),
name varchar,
data float
);
about 200000 rows

And about 1000 functions:
create function device_type1(int) returns ..
create function device_type2(int) returns ..
...
create function device_type1000(int) returns ..


What is faster?

One trigger with 1000 ELSE IF
if old.type=1 then 
    select device_type1(old.id);
else if old.type=2 then
	select device_type2(old.id);
...
else if old.type=1000 then
	select device_type1000(old.id);
end if;

Or 1000 rules
create rule device_type1 AS ON update to devices 
	where old.type=1  
    DO select device_type1(old.id);
create rule device_type2 AS ON update to devices 
    where old.type=2
    DO select device_type2(old.id);
...
create rule device_type1000 AS ON update to devices 
    where old.type=1000
    DO select device_type1000(old.id);

thx.

-- 
С уважением,
Ключников А.С.