Even if the function did a select from d, it could still have plenty of duplicates. To remove that possibility you would have to use the distinct clause which is also generally less efficient then a group by. "Bruno Wolff III" <bruno@xxxxxxxx> wrote in message news:20050219192033.GA24244@xxxxxxxxxxx > On Sat, Feb 19, 2005 at 14:02:34 -0500, > Oisin Glynn <me@xxxxxxxxxxxxxx> wrote: > > > > But the where clause defines the result of the aggregate function (in this > > case the SUM)? > > Not really. > > > Is the only reason for needing the GROUP BY CLAUSE is because the aggregate > > function demands it? > > Note that there is also a join to a table d. So that values in d are > being summed up based on some connection from d to the other 3 tables. > > > If so could something like the following work where we pass the where clause > > conditions into the function and it performs the aggregate function and > > returns.. I am guessing this would be extremely inefficient? > > > > select A.*,B.*,C.*,my_cheating_sum(a.id,b.id,c.id) from a,b,c, > > where some conditions; > > If that function did a select from d, you could make this work, but it > would likely be much slower than doing it in one SQL statement. > > ---------------------------(end of broadcast)--------------------------- > TIP 7: don't forget to increase your free space map settings > ---------------------------(end of broadcast)--------------------------- TIP 7: don't forget to increase your free space map settings
