UDFs support two types of parameter passing conventions.
| Parameter Passing Convention | Description | Usage |
|---|---|---|
| Parameter Style SQL | Provides a way to pass nulls as input arguments and return a null as the result. |
|
| Parameter Style TD_GENERAL | Does not accept null input arguments and does not return a null result. |
|
The parameter passing convention you use to code a UDF must correspond to the parameter passing specification in the CREATE FUNCTION statement for the UDF.
| IF CREATE FUNCTION specifies … | AND you are writing … | THEN use this syntax for function parameters … |
|---|---|---|
| PARAMETER STYLE TD_GENERAL | a scalar function | Scalar Function Parameter Style TD_GENERAL Syntax |
| an aggregate function | Aggregate Function Parameter Style TD_GENERAL Syntax | |
| PARAMETER STYLE SQL or omits the PARAMETER STYLE option | a scalar function | Scalar Function Parameter Style SQL Syntax |
| an aggregate function | Aggregate Function Parameter Style SQL Syntax | |
| a table function | Table Function Syntax |
Teradata provides a special UDF input parameter data type called the dynamic user-defined type (UDT). UDFs can specify up to eight dynamic UDT input parameters. A dynamic UDT can accommodate up to 128 attributes, where the data types of the attributes are determined at runtime.
If you consider each attribute of a dynamic UDT as an input parameter, then the number of effective input parameters increases to 1144, where:
| (8 dynamic UDT parameters * 128 attributes) + (120 parameters of other data types) = 1144 |
For details on using dynamic UDTs, see Defining Functions that Use UDT Types.