Command Parameters
Parameters are marked in command text with the question mark character. For example, the following SQL statement is marked for a single input parameter:
{call SalesByCategory('Produce', ?)}
To improve performance by reducing network traffic, the SQL Server Native Client OLE DB provider does not automatically derive parameter information unless ICommandWithParameters::GetParameterInfo or ICommandPrepare::Prepare is called before executing a command. This means that the SQL Server Native Client OLE DB provider does not automatically:
Verify the correctness of the data type specified with ICommandWithParameters::SetParameterInfo.
Map from the DBTYPE specified in the accessor binding information to the correct SQL Server data type for the parameter.
Applications will receive possible errors or loss of precision with either of these methods if they specify data types that are not compatible with the SQL Server data type of the parameter.
To ensure this does not happen, the application should:
Ensure that pwszDataSourceType matches the SQL Server data type for the parameter if hard-coding ICommandWithParameters::SetParameterInfo.
Ensure that the DBTYPE value being bound to the parameter is of the same type as the SQL Server data type for the parameter if hard-coding an accessor.
Code the application to call ICommandWithParameters::GetParameterInfo so that the provider can obtain the SQL Server data types of the parameters dynamically. Note that this causes an extra network round trip to the server.
Note
The provider does not support calling ICommandWithParameters::GetParameterInfo for any SQL Server UPDATE or DELETE statement containing a FROM clause; for any SQL statement depending on a subquery containing parameters; for SQL statements containing parameter markers in both expressions of a comparison, like, or quantified predicate; or queries where one of the parameters is a parameter to a function. When processing a batch of SQL statements, the provider also does not support calling ICommandWithParameters::GetParameterInfo for parameter markers in statements after the first statement in the batch. Comments (/* */) are not allowed in the Transact-SQL command.
The SQL Server Native Client OLE DB provider supports input parameters in SQL statement commands. On procedure-call commands, the SQL Server Native Client OLE DB provider supports input, output, and input/output parameters. Output parameter values are returned to the application either on execution (only if there are no rowsets returned) or when all returned rowsets are exhausted by the application. To ensure that returned values are valid, use IMultipleResults to force rowset consumption.
The names of stored procedure parameters need not be specified in a DBPARAMBINDINFO structure. Use NULL for the value of the pwszName member to indicate that the SQL Server Native Client OLE DB provider should ignore the parameter name and use only the ordinal specified in the rgParamOrdinals member of ICommandWithParameters::SetParameterInfo. If the command text contains both named and unnamed parameters, all of the unnamed parameters must be specified before any named parameters.
If the name of a stored procedure parameter is specified, the SQL Server Native Client OLE DB provider checks the name to ensure that it is valid. The SQL Server Native Client OLE DB provider returns an error when it receives an erroneous parameter name from the consumer.
Note
To expose support for SQL Server XML and user-defined types (UDT), the SQL Server Native Client OLE DB provider implements a new ISSCommandWithParameters interface.