这里讨论的是ADO调用存储过程,按MSDN的例子,一般都是下面的调用方法(节选自MSDN):
HRESULT hr = S_OK;
// Define ADO object pointers.
// Initialize pointers on define.
// These are in the ADODB:: namespace.
_RecordsetPtr pRstByRoyalty = NULL;
_RecordsetPtr pRstAuthors = NULL;
_CommandPtr pcmdByRoyalty = NULL;
_ParameterPtr pprmByRoyalty = NULL;
_ConnectionPtr pConnection = NULL;
_bstr_t strCnn("Provider=sqloledb;Data Source=srv;"
"Initial Catalog=Pubs;User Id=sa;Password=;");
_bstr_t strMessage, strAuthorID;
int intRoyalty;
VARIANT vtRoyalty;
try
{
//Open a Connection.
TESTHR(pConnection.CreateInstance(__uuidof(Connection)));
hr = pConnection->Open(strCnn,"","",NULL);
pConnection->CursorLocation = adUseClient;
//Open Command Object with one Parameter
TESTHR(pcmdByRoyalty.CreateInstance(__uuidof(Command)));
pcmdByRoyalty->CommandText = "byroyalty";
pcmdByRoyalty->CommandType = adCmdStoredProc;
//Define Integer/variant.
vtRoyalty.vt = VT_I2;
vtRoyalty.iVal = intRoyalty;
pprmByRoyalty = pcmdByRoyalty->CreateParameter(
"percentage",adInteger,adParamInput,sizeof(int),vtRoyalty);
pcmdByRoyalty->Parameters->Append(pprmByRoyalty);
pprmByRoyalty->Value = vtRoyalty;
//Create Recordset by executing the command
pcmdByRoyalty->ActiveConnection = pConnection;
pRstByRoyalty = pcmdByRoyalty->Execute(
NULL,NULL,adCmdStoredProc);
//执行结果的处理,省略...
pRstByRoyalty->Close();
pRstAuthors->Close();
pConnection->Close();
}
catch(_com_error &e)
{
//意外处理...
}
}
必须指定执行SQL语句的CommandType是adCmdStoredProc,存储过程有参数就必须CreateParameter生成参数,这样调用存储过程就成了很麻烦的一件事情,必须针对不同的存储过程生成不同的参数。一个两个还可以接受,如果项目的业务逻辑比较复杂,需要依赖大量的存储过程,并且项目的需求或功能可能要经常变动或扩展,那就成了一个噩梦了。有没有通用的方法呢?当然有,只需要通用的查询就可以实现。
_ConnectionPtr pConnection = NULL;
m_Conn.CreateInstance(__uuidof(Connection));
m_Conn->ConnectionString = "Provider=sqloledb;Data Source=srv;"Initial Catalog=Pubs;User Id=sa;Password=;";
m_Conn->Open("","","",NULL);
_variant_t m_param;
_bstr_t m_bstr;
int index = 0,nFieldCount;
_RecordsetPtr m_Rs;
m_Rs.CreateInstance(__uuidof(Recordset));
m_Rs->Open("select * from titles",(IDispatch*)(m_Conn->m_Conn),adOpenStatic, adLockReadOnly,adCmdText);
nFieldCount = m_Rs->Fields->Count;
while(!m_Rs->EndOfFile)
{
for(index=0;index < nFieldCount;index ++)
{
m_param.Clear();
m_param = (short)index;
m_param = m_Rs->Fields->GetItem(m_param);
if(m_param.vt != VT_NULL)
m_bstr = m_param;
else
m_bstr = "";
printf("%s\n",(char*)m_bstr);
}
m_Rs->MoveNext();
}
m_Rs->Close();
m_Conn->Close();
以上一段就是普通的执行查询SQL语句的代码,简单起见,去掉了所有防护和判断代码,对于MS SQLServer,如果是执行返回结果集的存储过程,比如Pubs的byroyalty,只需要把上面代码中的"select * from titles"替换成"byroyalty 100"就可以了,即"过程名 参数1,参数2,..."形式。
这样一来,不需要生成参数,只需要改变SQL语句,就实现了MS SQLServer存储过程的调用,可以通用。但是对于返回参数的存储过程,这段代码就不行了,不过还是可以通过SQL语句解决:
“declare @q int exec checkpwd 'user','pwd',@ret = @q OUTPUT select @q”
checkpwd是一个存储过程,两个输入参数,用户名和密码,第三个是输出参数@ret,返回用户验证的结果。通过执行上面的SQL语句,把输出参数作为结果集返回,就可以适应上面的VC代码了。返回值的存储过程也是一样,执行“declare @q int,@return int exec @return = checkpwd 'user','pwd',@ret = @q OUTPUT select @q,@return”就可以了。
这里讲的是SQLServer的存储过程调用,这段代码同样可以调用ORACLE的存储过程,不过需要对SQL语句作出一些调整,留待下一篇再说了。