Re: Average over time
От | Greg Sabino Mullane |
---|---|
Тема | Re: Average over time |
Дата | |
Msg-id | fffb090f2db247acc6c6f0c227322cba@biglumber.com обсуждение исходный текст |
Ответ на | Average over time (Jan Danielsson <jan.m.danielsson@gmail.com>) |
Список | pgsql-novice |
-----BEGIN PGP SIGNED MESSAGE----- Hash: RIPEMD160 > ...etc. Obviously, the avg_asum is the average for asum up to the dt > column's date. My gut feeling is that this can not be done -- but I > don't know why. Is it possible? Perhaps something like this: SELECT thisday, AVG(asum) AS avg_asum FROM (SELECT dt, SUM(amount) AS asum FROM transactions GROUP BY dt) AS x, (SELECT ('2007-01-01'::date+'1 day'::interval*q)::date AS thisday FROM generate_series(0,10) AS q) AS y WHERE dt <= thisday GROUP BY 1 ORDER BY 1; - -- Greg Sabino Mullane greg@turnstep.com PGP Key: 0x14964AC8 200703031916 http://biglumber.com/x/web?pk=2529DF6AB8F79407E94445B4BC9B906714964AC8 -----BEGIN PGP SIGNATURE----- iD8DBQFF6g/PvJuQZxSWSsgRAxxVAKC/wV0vh+b9ZTi3hEjEOjGK+4sniACfeSdU gqlKeLxqxylcZCCbY6OD1wQ= =srvv -----END PGP SIGNATURE-----
В списке pgsql-novice по дате отправления: