Posted by Ed Murphy on 02/22/07 17:01

Mark Probert wrote:

> There is a table that contains the following:
>
> aID bID cID dID
> --------------------------
> 6 9 28 1
> 6 2 28 2
> 6 11 28 3
> 6 1 27 5
> 6 9 27 5
> 6 11 27 5
>
> I am trying to get a result of:
>
> aID bID cID dID
> --------------------------
> 6 9 28 1
> 6 2 28 2
> 6 11 28 3
> 6 1 27 5
>
> So, something like "the set off all values over bID which have the
> greatest cID". Any ideas?

select t.aID, t.bID, t.cID, t.dID
from the_table t
join (
select bID, max(cID) max_cID
from the_table
group by bID
) s on t.bID = s.bID and t.cID = s.max_cID

[Back to original message]