The default maximum precision of numeric and decimal data types is 38. In Transact-SQL, numeric is functionally equivalent to the decimal data type. Use the decimal data type to store numbers with decimals when the data values must be stored exactly as specified.
The behavior of float and real follows the IEEE 754 specification on approximate numeric data types. Because of the approximate nature of the float and real data types, do not use these data types when exact numeric behavior is required, such as in financial applications, in operations involving rounding, or in equality checks. Instead, use the integer, decimal, money, or smallmoney data types. Avoid using float or real columns in WHERE clause search conditions, especially the = and <> operators. It is best to limit float and real columns to > or < comparisons.
