Splitting matrices into small matrices based on column values

I have two matrices, A and B where A is the name of each field and B is their corresponding productions as:

    A = [3 2 1 10;
         1 3 2 20;
         2 3 1 30;
         3 2 1 25;
         1 3 2 10;
         2 3 1 15]

and

B = [10 20  5 10;
      4  6 20 20;
      5 10 20 30;
     20 10 30 25;
      4 20 12 10;
     10 30 20 15]

The last column of each matrix represents their profit and they are matching. For example, filed 3 runs at 10, field 2 at 20 and field 3 runs at 5 to make profit of 10. Now I want to create matrices which shows only the matching order of each fields such as x = [3 2 1 10; 3 2 1 25] and their corresponding productions as [10 20 5 10;20 10 30 25]. In my model A and B sizes are 200,000 rows and I need to see how the profit is changing for each fields order. So, the result will be

AA_fields = [3 2 1 10; 3 2 1 25]

AA_Rate = [10 20 5 10; 20 10 30 25]

This is for the case of (3 2 1) and the rest will follow the same way with different matrices.

[Bob Thompson]

You can try working with the unique() command. I personally haven't used it yet, so I don't know if it will find the different sets and mark the accordingly, or if it will simply find only values which appear once.