I would like to create a formula that a) compares four columns of data (see below) Sep Oct name units name units apple 2 apple 3 pear 3 pear 7 orange 4 banana 6 banana 3 toffee 5 then b) merges the two "names" column into one column, dropping any duplicates but still retaining the two unit columns (for months Sep and Oct) Sep Oct name units units apple 2 3 pear 3 7 orange 4 0 banana 3 6 toffee 0 6 then c) creates a third column that compares "Sep units" against "Oct units" and produces the total in the "difference" column Sep Oct name units units difference apple 2 3 1 pear 3 7 4 orange 4 0 -4 banana 3 6 3 toffee 0 6 6