From: | Marcin Grondecki <ojciec(at)mtl(dot)pl> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | rules - optimizing |
Date: | 1998-12-16 16:04:47 |
Message-ID: | 3.0.6.32.19981216160447.007a6e10@mtl.pl |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Hi ppl
I've started to use rules. Something strange with one of them:
AIX 4.1
PostgreSQL 6.4
create word_list (id int4, word char(15), count int4 default 0);
CREATE
create unique index word_list_word_uidx on word_list (word);
CREATE
...
(something inserted into word_list)
create table temp (word char(15));
CREATE
create rule ins_word as on insert to temp do instead update word_list set
count=count+1 where word=new.word;
CREATE
I'd like to use "temp" relation as some kind of filter; now i could f.e.
import big chunk of
data from text file (import word) with simple "copy", and this should count
words in file
storing results in word_list.count.
And it works, but...
explain update word_list set count=count+1 where word='aix';
shows index scan (using word_list_word_uidx of course)
but
explain insert into temp (word) values ('aix');
gives seq scan as result.
Relation word_list has about 300 000 tuples, so it's rather boring to w8 1
or 2 s to insert ONE word.
Could i do smth to enforce index scan when using above rule?
BTW. sorry for my english :(
Marcin Grondecki
ojciec(at)mtl(dot)pl
***** I'm not a complete idiot, some parts are missing...
From | Date | Subject | |
---|---|---|---|
Next Message | Albert Chen | 1998-12-17 01:03:52 | Why PostgreSQL is better than other commerial softwares? |
Previous Message | Oleg Broytmann | 1998-12-16 08:55:46 | Re: [GENERAL] Date & Datetime problems |