Keep highest value by multiple groups

[Anne Todd]

With data organized like below, I want to keep only the row containing the highest -counter- value within respondent-category groups. So for respondent 100, category 0, the highest -counter- value would be 3, and that's the row I'd want to keep within that 100-0 grouping. What is the most efficient way to do this? Thanks in advance.


* Example generated by -dataex-. For more info, type help dataex
clear
input int respondent byte(category counter score)
100 0 1 3
100 0 2 1
100 0 3 10
100 1 1 4
100 1 2 20
100 1 3 14
100 1 4 4
100 2 1 13
100 2 2 2
101 0 1 3
101 0 2 4
101 0 3 2
101 0 4 3
101 1 1 17
101 1 2 2
101 2 1 12
101 2 2 11
101 2 3 3
end
[/CODE]

[Clyde Schechter]

by respondent category (counter), sort: keep if _n == _N