[posted and mailed, please reply in news]
Giovanni Azua (bravegag@hotmail.com) writes:
[quoted text, click to view] > I need to return back to TSQL two numeric values from an
> Extended Stored Procedure developed in C. As result my C
> dll program produces two float values, the TSQL side expects
> to have exactly: numeric(10, 5) and numeric.
>
> Given that the structure to fill-up is DBNUMERIC defined
> as:
>
> typedef struct dbnumeric // Numeric (and decimal)
> {
> BYTE precision; // Precision
> BYTE scale; // Scale
> BYTE sign; // 1 = Positive, 0 = Negative
> BYTE val[MAXNUMERICLEN]; // Padded little endian value
> } DBNUMERIC;
>
> I do not have any clue how to convert the float C datatype
> to DBNUMERIC, specifically how to convert the float to an
> array of bytes "BYTE val[MAXNUMERICLEN]".
A quick glance makes me believe that srv_convert is able to do the task.
There is some documentation on how to work with DBNUMERIC.
Else you can use the IDataConvert object from OLE DB:
static IDataConvert * data_convert_ptr = NULL;
// Call this once when you DLL is loaded, and keep the object.
ret = CoCreateInstance(CLSID_OLEDB_CONVERSIONLIBRARY,
NULL, CLSCTX_INPROC_SERVER,
IID_IDataConvert,
(void **) &data_convert_ptr);
// Here is a sample routine. Don't bother about the SV, that's a Perl
// thing which holds the double value.
BOOL SV_to_decimal(SV * sv,
BYTE precision,
BYTE scale,
DB_NUMERIC &decimalval)
{
HRESULT ret;
double dbl = SvNV(sv);
ret = data_convert_ptr->DataConvert(
DBTYPE_R8, DBTYPE_NUMERIC, sizeof(double), NULL,
&dbl, &decimalval, NULL, DBSTATUS_S_OK, NULL,
precision, scale, 0);
return SUCCEEDED(ret);
}
Don't really remember which include files that has which, but these
are the ones that I include in my code (which access SQL Server through
SQLOLEDB, and has nothing to do with XPs):
#include <cguid.h>
#include <oledb.h>
#include <oledberr.h>
#include <msdasc.h>
#include <msdadc.h>
#include <msdaguid.h>
--
Erland Sommarskog, SQL Server MVP, esquel@sommarskog.se
Books Online for SQL Server SP3 at
