OK, this does not work: select max(t1."When"), t1."Pt_Id", t2."DateOfBirth" from "PtStaffAccess" t1, "Person" t2 where t1."Pt_Id" = t2.id group by t1."Pt_Id"; But this does: select max(t1."When"), t1."Pt_Id", min(t2."DateOfBirth") from "PtStaffAccess" t1, "Person" t2 where t1."Pt_Id" = t2.id group by t1."Pt_Id"; Now the error message was clear, and I think PG is following the standard here. But I have a question just for my own education. It seems to me, given that "Person".id is declared as the primary key, it should possible to deduce that there is no way that the 1st query could ever have multiple values of "DateOfBirth" to choose from when building a result row. Am I missing something? Or am I right, that this is something that SQL could do but simply doesn't, for whatever reason, historical, complexity... In fact, what's even more surprising to me, is that if I change the grouping to the other side of the join, it still doesn't work: select max(t1."When"), t2.id, t2."DateOfBirth" from "PtStaffAccess" t1, "Person" t2 where t1."Pt_Id" = t2.id group by t2.id; Come on, I'm grouping on the primary key and it thinks that there might be multiple values for the other columns? -- Scott Ribe scott_ribe@xxxxxxxxxxxxxxx http://www.killerbytes.com/ (303) 722-0567 voice -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
