C Data Type Definition
typedef signed char DECIMAL1;
typedef short DECIMAL2;
typedef int DECIMAL4;
typedef signed char NUMERIC1;
typedef short NUMERIC2;
typedef int NUMERIC4;
typedef struct
{
unsigned int low;
int high;
} DECIMAL8, NUMERIC8;
typedef struct
{
unsigned int int1;
unsigned int int2;
unsigned int int3;
int int4;
} DECIMAL16, NUMERIC16;
Usage
The size of the SQL type determines which C type to use.
| FOR DECIMAL(n,m) or NUMERIC(n,m), where … | Use one of these C types … |
|---|---|
| 1 ≤ n ≤ 2 | DECIMAL1 or NUMERIC1 |
| 3 ≤ n ≤ 4 | DECIMAL2 or NUMERIC2 |
| 5 ≤ n ≤ 9 | DECIMAL4 or NUMERIC4 |
| 10 ≤ n ≤ 18 | DECIMAL8 or NUMERIC8 |
| 18 < n | DECIMAL16 or NUMERIC16 |
The size of the SQL type also determines the range of values for a DECIMAL input argument or return argument. For example, if a CREATE FUNCTION statement defines a DECIMAL(1,0) input argument, the range of values for the argument is -9 to 9. A value outside the valid range of values produces a numeric overflow error. For details on DECIMAL types, see Teradata Vantage™ - Data Types and Literals, B035-1143.
For a DECIMAL type that can be represented as variable length, the best practice is to specify the DECIMAL so that it can handle the largest value that the C code is willing to deal with.
This example uses DECIMAL in a UDF definition and C function declaration.
| SQL Function Definition | Equivalent C Function Declaration |
|---|---|
CREATE FUNCTION F1 ( A DECIMAL(5,0) ) RETURNS DECIMAL(5,0) ...; |
void f1( DECIMAL4 *a,
DECIMAL4 *result,
... )
{ ... }
|