Re: Converting time interval to double precision of time unit
От | Tom Lane |
---|---|
Тема | Re: Converting time interval to double precision of time unit |
Дата | |
Msg-id | 27171.1269975326@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Re: Converting time interval to double precision of time unit (Alban Hertroys <dalroi@solfertje.student.utwente.nl>) |
Ответы |
Re: Converting time interval to double precision of time
unit
|
Список | pgsql-general |
Alban Hertroys <dalroi@solfertje.student.utwente.nl> writes: > On 30 Mar 2010, at 18:29, Mike Toews wrote: >> I'm using 8.3, and I'm trying to work with the interval type, and I >> can't seem to get things right. I've been all over the docs[1,2], and >> there is no mention on how this can be done. >> >> While I can get: >> SELECT '3 day 2 hour 34 minute'::interval >> >> .. how can then get the fractional hours of this time interval in >> double precision (or seconds, minutes, years, decades, etc.)? >> >> Do I really need to extract the time subcomponents and do the math myself? > You shouldn't try to do that. How do you expect to convert an interval type to a timestamp without having a timestamp tobase it on? It's a relative quantity with a variable value depending on it's base value. For a meaningful answer it requiresinformation about DST changes, different month lengths, leap years, etc, which it won't have if you don't tell whereyou're basing your interval off. > If instead you base your interval on a relevant base-timestamp, then you can simply extract epoch from the result, althoughthats in seconds and not (fractional) hours, but that's a linear relationship. I think what Mike is actually looking for is SELECT extract(epoch from interval '3 days 2 hours 34 minutes'); date_part ----------- 268440 (1 row) although your point about the uncertainty of the conversion for units of days or larger is certainly well-taken. regards, tom lane
В списке pgsql-general по дате отправления: