DECIMAL and FLOAT are both used to store numerical values. But they have the following main differences:

* DECIMAL(p,s) stores values with the decimal point fixed at the position of s (scale) digits from the right. The total number of decimal digits is also fixed as p (precesion).

* FLOAT(n) stores values with the decimal point floating based on the value. The number of bits used to store the mantissa part is fixed as n.

* If the input value of DECIMAL(p,s) has more digits after the decimal point than the scale s, the value will be rounded to the scale s.

* If the input value of FLOAT(n) has more total digits (mantissa) than what n bits can store, the value will be rounded to fit the storage size.

* If the input value of DECIMAL(p,s) has more digits before the decimal point than p-s, SQL Server will give you an over-flow error.

* If the input value of FLOAT(n) is too big that the exponential part goes over the positive limit, SQL Server will give you an over-flow error.

* If the input value of FLOAT(n) is too small that the exponential part goes over the negative limit, SQL Server will give you an under-flow error.