pg noob <pgnube(at)gmail(dot)com> writes:
> Is there an efficient way to select the set of rows which have the max
> values across multiple columns?
I think SELECT DISTINCT ON might help you. The idea is to sort the rows
in an appropriate order then take the first row from each group. Look
at the "weather reports" example in PG's SELECT reference page for
inspiration.
regards, tom lane