Another stupid question :)
От | Dmitri Touretsky |
---|---|
Тема | Another stupid question :) |
Дата | |
Msg-id | 189177874429.20021204060241@listsoft.ru обсуждение исходный текст |
Список | pgsql-novice |
Good time of the day! There are three tables: CREATE big ( id int4 PRIMARY KEY, ref1 int4, ref2 int4, ... ); CREATE ref1 ( id int4 PRIMARY KEY, ... ); CREATE ref2 ( id int4 PRIMARY KEY, ... ); Tables ref1 and ref2 are small, like 10-15 records. Table big is large. Often I need to get data from all tables. Which way is better in terms on performance: SELECT big.*, ref1.*, ref2.* FROM big, ref1, ref2 WHERE big.ref1 = ref1.id AND big.ref2 = ref2.id AND ... OR make 3 separate queries like, say SELECT * FROM big WHERE ... and then in the loop select corresponding data from small tables? And more general question: how queries affect performance? Which way is better: run two queries which will return, say, 5 records each, or run one query which will return, say, 20 records and then drop unneeded records in my backend application? Best regards, Dmitri ( mailto:dmitri@listsoft.ru ) New SOFT daily (RUS): http://www.listsoft.ru/ (ENG): http://www.listsoft.com/ Articles, tips : http://www.diskovod.ru/ --- Ошибок не прощают женщины и тетрис на 9й скорости.
В списке pgsql-novice по дате отправления: