--- Igor Tandetnik <[EMAIL PROTECTED]> wrote: > woj <[EMAIL PROTECTED]> wrote: > > Now, when I run a query: > > SELECT Mieszalnia.IdMat, Mieszalnia.Partia, Mieszalnia.Kont, > > Mieszalnia.Uk, Max(Mieszalnia.Data) FROM Mieszalnia; > > I always get: > > IdMat Partia Kont Uk Data > > 6 33333 33333 sl1 1172135769 > > > > In this result there is indeed max from Data field but rest of the > > fields fit not... > > When a SELECT statement involves aggregate functions, all column > references in the SELECT must be either parameters to some aggregate > functions, or else be also mentioned in GROUP BY clause. SQLite allows, > as an extension, departure from this rule, but the row from which values > for columns that are neither aggregated nor grouped by are taken is > random and unpredictable. > > Specifically, in the query you show, there's no guarantee that values > for IdMat, Partia and so on would be taken from the same row from which > Max(Data) comes. Even if SQLite really wanted to help you out here, it > is impossible in general. Consider: > > SELECT IdMat, Max(Data), Min(Data) from Mieszalnia; > > Which value of IdMat would you expect to see in response to such a > query? Should it come from the row with the largest value of Data, with > the smallest, or some other?
This thread describes how the non-aggregate values in a group by are selected by SQLite: http://www.mail-archive.com/sqlite-users@sqlite.org/msg17770.html ____________________________________________________________________________________ Yahoo! Music Unlimited Access over 1 million songs. http://music.yahoo.com/unlimited ----------------------------------------------------------------------------- To unsubscribe, send email to [EMAIL PROTECTED] -----------------------------------------------------------------------------
