Hi,
I am running a query which makes a join on 3 tables:
- generals (contains 200 000 rows)
- category_generals (contains 200 000 rows)
- generals_topics (contains 15 000 000 rows)
The query is wrote this way:
SELECT gt.id_topic
FROM generals g,
category_generals cs,
generals_topics gt
WHERE g.media = 't'
AND g.id_general = cg.id_general
AND gt.id_general = g.id_general
AND cg.id_category = 15
ORDER BY gt.id_topic DESC
OFFSET 0
LIMIT 20;
The column id_general of table generals is a primary key. All the other
columns used to make the joins are indexes.
The query is slow but it works fine as it completes in less than 1 second.
The problem is that if I change the filter value of id_category from 15 to 3
the query will take more than 7 minutes to complete! The only difference
between id_category 3 and 15 is that there is about 4000 rows in the first
one (id_category = 3) and 2000 rows in the second one (id_category = 15).
An explain give me the following result:
QUERY PLAN
---------------------------------------------------------------------------------------------------------------------------
Limit (cost=0.00..9677.68 rows=20 width=4)
-> Nested Loop (cost=0.00..61006657.19 rows=126077 width=4)
-> Nested Loop (cost=0.00..59991538.61 rows=252145 width=12)
-> Index Scan Backward using generals_topics_pkey on
generals_topics gt (cost=0.00..615679.86 rows=14750423 width=8)
-> Index Scan using ix_category_generals_id_general on
category_generals cs (cost=0.00..4.01 rows=1 width=4)
Index Cond: ("outer".id_general = cs.id_general)
Filter: (id_category = 3)
-> Index Scan using generals_id_topic_key on generals g
(cost=0.00..4.01 rows=1 width=4)
Index Cond: (g.id_general = "outer".id_general)
Filter: media
As you can see, every rows of generals_topics table is scanned and I don't
understand why? How can I do to make postgresql to work only with the tuples
resulting from the join conditions? Is it a configuration problem ?
Thanks in advance for any help.
--
Olivier Sirven
Elma Ingénierie Informatique
3, rue d'Uzès
F-75002 - Paris - France
http://www.elma.fr
Tel: +33-1-44882744
Fax: +33-1-44882747
Email: olivier@xxxxxxx
