Re: time problem with postgres ODBC driver (fwd)
От | Tom Lane |
---|---|
Тема | Re: time problem with postgres ODBC driver (fwd) |
Дата | |
Msg-id | 22764.1013539055@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Re: time problem with postgres ODBC driver (fwd) (Laurette Cisneros <laurette@nextbus.com>) |
Ответы |
Re: time problem with postgres ODBC driver (fwd)
|
Список | pgsql-odbc |
Laurette Cisneros <laurette@nextbus.com> writes: > On Tue, 12 Feb 2002, Tom Lane wrote: >> It sounds like ODBC (or the client application) is misinterpreting the >> datatype of col2 as being a "date" or "date/time" type not "time of >> day". There is a translation between PG internal types and the ODBC >> standard's notion of types, so one possible explanation is that there's >> something getting lost in translation. > What's even more interesting is that casting doesn't help - > col2::time still returns the current date. No, of course it wouldn't --- the column coming out of the backend is PG's "time" type either way. I'm sure that the unwanted conversion to a date or datetime value (with, evidently, implicit fill-in of today's date) is happening on the client side. I don't know if our ODBC code should be blamed or if it's a client-application bug. A quick look at the ODBC sources makes it appear that the ODBC code reports ODBC type code "SQL_TIME" for a PG "time" column, which seems a reasonable mapping to me, but I'm no ODBC expert. > Us too. The only difference is that that ODBC driver was an older version > than the new one that was downloaded and started this problem Hmm, so perhaps the problem could have been triggered by a recent "fix". I looked at 7.1 and 7.2 ODBC sources and they seemed about the same in this respect. Do you know what version the GB driver was, exactly? regards, tom lane
В списке pgsql-odbc по дате отправления: