I got a result very similar to yours:

select count(*) as rank, max(s1.number) as number

from(

select number from NMaxMin group by number

) as s1

inner join

(

select number from NMaxMin group by number

) as s2 on s1.number <= s2.number

group by s1.number

having count(*) = 5

(Just change the join condition to seach for top or bottom values.)

Edited by - pegasus on 11/20/2003 9:07:45 PM