How to Design logic for grouping of variables

Suppose we have a variable called 'group' which contains grouping information

group = [ {1}, {2}, ..., {D} ]

We have to maintain a counter for every pair (i, j) where i, j are in the range [1 D] and for all those pairs whose count(i, j) >= 5, these pairs are merged.

Example:

D = 5
group = [ {1}, {2}, {3}, {4}, {5} ]

(1, 2) = 7
(2, 3) = 10
(3, 4) = 2
(4, 5) = 20

hence pairs are merged and the resultant group is

group = [ {1 2 3}, {4 5} ]

in my problem D can have a value of 1000, what is the efficient way to implement this logic

[Stephen]

I get the feeling that this is part of a larger task, which possibly could be efficiently solved using accumarray. You should read these: http://xyproblem.info/

https://meta.stackexchange.com/questions/66377/what-is-the-xy-problem