The SQL standard previsions three approximate numeric data types:
float(n)
real
double precision
The precistion of real is implementation defined. For example, in SQL Server, real is a synonym for float(24).
For most numbers, an approximate numeric data type stores an approximation of the number to be stored. Because the stored value is only an approximation of a value, this data type is problematic in where conditions when the = or <> operator is used.