Use the approximate numeric types, float, double precision, and real, for numeric data that can tolerate rounding. The approximate numeric types are especially suited to data that covers a wide range of values. They support all aggregate functions and all arithmetic operations.
