the c and d were thrown in as column headers from the looks of things,
started with 2 columns of data, a and b, ended up with 4 columns of data, a, b, c and d.
its just data grouping in the end (all the data items in b where the A column is the same get grouped, etc).