GROUP by Customer. GROUP is used with aggregate functions. For example, to list the total amount of all the Customer transactions in a given result set, GROUP by Customer and SUM the amount to produce a result set w one row for every Customer and the an aggregate amount next to it. Now, if you take away the SUM -- let's say the only column is Customer name -- and you leave it to GROUP BY Customer, the result set is effectively a unique list of Customers implicated by the queriy criteriais simply the list of Customers of the result set at hand -- still only one row for every Customer, no repetition/duplicates (aka "unique" or "distinct" values).