Data can be converted from one numeric type to another. If the new type is an exact numeric whose precision or scale is not sufficient to hold the data, errors can occur.
For example, if you provide a float or numeric value as an argument to a built-in function that expects an integer, the value of the float or numeric is truncated. However, Adaptive Server does not implicitly convert numerics that have a fractional part but returns a scale error message. For example, Adaptive Server returns error 241 for numerics that have a fractional part and error 257 if other datatypes are passed.
Use the arithabort and arithignore options to determine how Adaptive Server handles errors resulting from numeric conversions.
The arithabort and arithignore options have been redefined for release 10.0 or later. If you use these options in your applications, examine them to be sure they are still producing the desired behavior.