Condividi tramite


Recupero di dati di grandi dimensioni

Si applica a: SQL Server Azure SQL DatabaseIstanza gestita di SQL di AzureAzure Synapse Analytics AnalyticsPlatform System (PDW)

Scaricare il driver OLE DB

I consumer devono in genere isolare il codice che crea un oggetto di archiviazione del driver OLE DB per SQL Server da altro codice che gestisce i dati a cui non si fa riferimento tramite un puntatore di interfaccia ISequentialStream.

In questo articolo vengono descritte le funzionalità disponibili con le funzioni seguenti:

  • IRowset:GetData

  • IRow::GetColumns

  • ICommand::Execute

Il consumer deve recuperare una singola riga di dati in una chiamata al metodo GetNextRows quando la proprietà DBPROP_ACCESSORDER, nel gruppo di proprietà del set di righe, è impostata su DBPROPVAL_AO_SEQUENTIAL o DBPROPVAL_AO_SEQUENTIALSTORAGEOBJECTS. I dati BLOB infatti non vengono memorizzati nel buffer. Se il valore di DBPROP_ACCESSORDER è impostato su DBPROPVAL_AO_RANDOM, il consumer può recuperare più righe di dati in GetNextRows.

OLE DB Driver per SQL Server non recupera dati di grandi dimensioni da SQL Server fino a quando non riceve una richiesta a tale scopo dal consumer. Il consumer deve associare tutti i dati di tipo short in una funzione di accesso e quindi utilizzare una o più funzioni di accesso temporanee per recuperare valori di dati di grandi dimensioni come richiesto.

Esempio

In questo esempio viene recuperato un valore di dati di grandi dimensioni da una singola colonna:

HRESULT GetUnboundData  
    (  
    IRowset* pIRowset,  
    HROW hRow,  
    ULONG nCol,   
    BYTE* pUnboundData  
    )  
    {  
    UINT                cbRow = sizeof(IUnknown*) + sizeof(ULONG);  
    BYTE*               pRow = new BYTE[cbRow];  
  
    DBOBJECT            dbobject;  
  
    IAccessor*          pIAccessor = NULL;  
    HACCESSOR           haccessor;  
  
    DBBINDING           dbbinding;  
    ULONG               ulbindstatus;  
  
    ULONG               dwStatus;  
    ISequentialStream*  pISequentialStream;  
    ULONG               cbRead;  
  
    HRESULT             hr;  
  
    // Set up the DBOBJECT structure.  
    dbobject.dwFlags = STGM_READ;  
    dbobject.iid = IID_ISequentialStream;  
  
    // Create the DBBINDING, requesting a storage-object pointer from  
    // The OLE DB Driver for SQL Server.  
    dbbinding.iOrdinal = nCol;  
    dbbinding.obValue = 0;  
    dbbinding.obStatus = sizeof(IUnknown*);  
    dbbinding.obLength = 0;  
    dbbinding.pTypeInfo = NULL;  
    dbbinding.pObject = &dbobject;  
    dbbinding.pBindExt = NULL;  
    dbbinding.dwPart = DBPART_VALUE | DBPART_STATUS;  
    dbbinding.dwMemOwner = DBMEMOWNER_CLIENTOWNED;  
    dbbinding.eParamIO = DBPARAMIO_NOTPARAM;  
    dbbinding.cbMaxLen = 0;  
    dbbinding.dwFlags = 0;  
    dbbinding.wType = DBTYPE_IUNKNOWN;  
    dbbinding.bPrecision = 0;  
    dbbinding.bScale = 0;  
  
    if (FAILED(hr = pIRowset->  
        QueryInterface(IID_IAccessor, (void**) &pIAccessor)))  
        {  
        // Process QueryInterface failure.  
        return (hr);  
        }  
  
    // Create the accessor.  
    if (FAILED(hr = pIAccessor->CreateAccessor(DBACCESSOR_ROWDATA, 1,  
        &dbbinding, 0, &haccessor, &ulbindstatus)))  
        {  
        // Process error from CreateAccessor.  
        pIAccessor->Release();  
        return (hr);  
        }  
  
    // Read and process BLOCK_SIZE bytes at a time.  
    if (SUCCEEDED(hr = pIRowset->GetData(hRow, haccessor, pRow)))  
        {  
        dwStatus = *((ULONG*) (pRow + dbbinding.obStatus));  
  
        if (dwStatus == DBSTATUS_S_ISNULL)  
            {  
            // Process NULL data  
            }  
        else if (dwStatus == DBSTATUS_S_OK)  
            {  
            pISequentialStream = *((ISequentialStream**)   
                (pRow + dbbinding.obValue));  
  
            do  
                {  
                if (SUCCEEDED(hr =  
                    pISequentialStream->Read(pUnboundData,  
                    BLOCK_SIZE, &cbRead)))  
                    {  
                    pUnboundData += cbRead;  
                    }  
                }  
            while (SUCCEEDED(hr) && cbRead >= BLOCK_SIZE);  
  
            pISequentialStream->Release();  
            }  
        }  
    else  
        {  
        // Process error from GetData.  
        }  
  
    pIAccessor->ReleaseAccessor(haccessor, NULL);  
    pIAccessor->Release();  
    delete [] pRow;  
  
    return (hr);  
    }  

Vedere anche

Oggetti BLOB e OLE
Uso di tipi valore di grandi dimensioni