NVL is a scalar function whose return value data type depends on the data type associated with the arguments passed to it.

- If the input arguments are numeric types, the function determines which argument has the highest precedence, converts the other argument to that data type, and returns that data type. For details about the order of precedence, see “Compatible Types” in
Teradata Vantage™ SQL External Routine Programming
, B035-1147
.
If that data type is DECIMAL/NUMERIC and the precision and scale of the input arguments are different, the precision and scale of the return type is set to achieve the maximum precision possible. For example, if the input arguments are DECIMAL(6,3), DECIMAL(7,4), and DECIMAL(8,7), the return type would need three digits to the left of the decimal point and seven digits to the right of the decimal point to avoid any reduction in precision.

In cases where it is not possible to maintain the maximum precision, the data is rounded according to the DBS Control Record RoundHalfWayMagUp field. For example, if the input arguments are DECIMAL(32, 8) and DECIMAL(30, 28), the return type will be DECIMAL(38,14). This will allow for 24 digits to the left of the decimal point (required for the DECIMAL(32,8) argument), and 14 digits to the right of the decimal point.

If the data type is fixed point NUMBER and the precision is less than or equal to 38, the precision and scale of the return type are calculated with the same method used for DECIMAL/NUMERIC. However, if the precision is greater than 38, the return type is changed to NUMBER(*) to avoid loss of accuracy. If the data type is floating point NUMBER, the return type is NUMBER(*).

- If the two arguments are character data types, the function converts the second argument to the data type of the first argument and returns the type as VARCHAR.
- If all input character types are LATIN, the result is LATIN. If any input is not LATIN, the function converts all input to Unicode and the return character set is Unicode.

You can also pass arguments with data types that can be converted to the above types using the implicit data type conversion rules that apply to UDFs.