Re: Spped of max
| От | Jean-Luc Lachance |
|---|---|
| Тема | Re: Spped of max |
| Дата | |
| Msg-id | 3CE26FB7.97AB2704@nsd.ca обсуждение исходный текст |
| Ответ на | Spped of max (Edmund Dengler <edmundd@eSentire.com>) |
| Ответы |
Re: Spped of max
|
| Список | pgsql-general |
Neil Conway wrote: >[...] > If you don't have a qualification (i.e. SELECT count(*) FROM x), there > are a couple ways to do it: use triggers to increment/decrement a > counter of the number of rows in the table, create a btree-indexed > SERIAL column and do an ORDER BY serial_column DESC LIMIT 1 on it (note > that this is fragile, your sequence could easily have holes), or if you > only need an approximation, you could use the pg_class attribute > "reltuples" for the OID of your table. My impression is that most > people choose the first method. > The real question is: Why is reltuples only an approximation?
В списке pgsql-general по дате отправления: