Visual C++中使用OLE DB读写SQL Server简明指南

 

©2005 黄友生。本文由原作者发布于MSN Space、CSDN。你可以保存、在非商业软件中使用、在非盈利性文章中引用本文中的部分或全部文字，但请注明作者及原文地址。要用于其它用途，请先联系作者（ [email protected] ）。作者不保证本文完全正确无误、不对因本文中的理论或代码缺陷造成的损失负责。

 

在需要对数据库进行操作时，OLE DB总是被认为是一种效率最高但最难的方法。但是以我最近使用OLE DB的经验看来，OLE DB的效率高则高矣，但却一点都不难。说它难恐怕主要是因为可参考的中文资料太少，为了帮助以后需要接触OLE DB的同行，我撰写了这篇文章。本文包含如下内容：

 

1. OLE DB写数据库；

2. OLE DB读数据库；

3. OLE DB对二进制数据(text、ntext、image等)的处理。

 

首先来看看对SQL Server进行写操作的代码，有一定VC基础的读者应该可以很顺利地看懂。OLE DB写数据库，就是这么简单！

 

注:

1.以下代码中使用的模板类EAutoReleasePtr<T>与ATL中的CComPtr<T>类似，是一个在析构时自动调用Release的类。CComPtr<T>的代码在ATLBASE.H中定义。

2.以下代码均在UNICODE环境下编译，因为执行的SQL语句必须是UNICODE的。设置工程为UNICODE的方法是:首先在project->settings->C/C++的属性页中的Preprocessor中，删除_MBCS写入UNICODE,_UNICODE。然后在link属性页中Category中选择output，在Entry-Point symbol 中添加wWinMainCRTStartup。

 

EAutoReleasePtr<IDBInitialize> pIDBInitialize;
HRESULT hResult = ConnectDatabase( &pIDBInitialize, _T("127.0.0.1"), _T(“sa”), _T("password") );
if( FAILED( hResult ) )

{
    //失败，可能是因为数据库没有启动、用户名密码错等等

    return;
}

EAutoReleasePtr<IOpenRowset> pIOpenRowset;
hResult = CreateSession( pIDBInitialize, &pIOpenRowset );
if( FAILED( hResult ) )
{

    //出错

    return;
}

EAutoReleasePtr<ICommand> pICommand;
EAutoReleasePtr<ICommandText> pICommandText;
hResult = CreateCommand( pIOpenRowset, &pICommand, &pICommandText );
if( FAILED( hResult ) )
{

   //出错

    return;
}

hResult = ExecuteSQL( pICommand, pICommandText, _T("USE PBDATA") );

if( FAILED( hResult ) )
{

    //如果这里失败，那就是SQL语句执行失败。在此处，就是PBDATA还未创建

    return;
}

 

// 创建表
ExecuteSQL( pICommand, pICommandText, _T("CREATE TABLE 2005_1(Volume real NOT NULL,ID int NOT NULL IDENTITY)") );

 

// 添加记录

ExecuteSQL( pICommand, pICommandText, _T("INSERT INTO 2005_1 VALUES(100.0)") );

 

//...

  

其中几个函数的代码如下:

 

HRESULT ConnectDatabase( IDBInitialize** ppIDBInitialize, LPCTSTR pszDataSource, LPCTSTR pszUserID, LPCTSTR pszPassword )
{
    ASSERT( ppIDBInitialize != NULL && pszDataSource != NULL && pszUserID != NULL && pszPassword != NULL );

 

    UINT uTimeout ＝ 15U; // 连接数据库超时（秒）
    TCHAR szInitStr[1024];

    VERIFY( 1023 >= wsprintf( szInitStr, _T("Provider=SQLOLEDB;Data Source=%s;Initial Catalog=master;User Id=%s;Password=%s;Connect Timeout=%u"), pszDataSource, pszUserID, pszPassword, uTimeout ) );
    //Initial Catalog=master指明连接成功后，"USE master"。

    EAutoReleasePtr<IDataInitialize> pIDataInitialize;
    HRESULT hResult = ::CoCreateInstance( CLSID_MSDAINITIALIZE, NULL, CLSCTX_INPROC_SERVER,
            IID_IDataInitialize, ( void** )&pIDataInitialize );

    if( FAILED( hResult ) )

    {
        return hResult;
    }

 

    EAutoReleasePtr<IDBInitialize> pIDBInitialize;
    hResult = pIDataInitialize->GetDataSource( NULL, CLSCTX_INPROC_SERVER, ( LPCOLESTR )szInitStr,
              IID_IDBInitialize, ( IUnknown** )&pIDBInitialize );
    if( FAILED( hResult ) )
    {
        return hResult;
    }

    hResult = pIDBInitialize->Initialize( );
    if( FAILED( hResult ) )
    {
        return hResult;
    }

    * ppIDBInitialize = pIDBInitialize.Detach( );
    return S_OK;
}

 

HRESULT CreateSession( IDBInitialize* pIDBInitialize, IOpenRowset** ppIOpenRowset )
{
    ASSERT( pIDBInitialize != NULL && ppIOpenRowset != NULL );

    EAutoReleasePtr<IDBCreateSession> pSession;
    HRESULT hResult = pIDBInitialize->QueryInterface( IID_IDBCreateSession, ( void** )&pSession );
    if( FAILED( hResult ) )
    {
        return hResult;
    }

    EAutoReleasePtr<IOpenRowset> pIOpenRowset;
    hResult = pSession->CreateSession( NULL, IID_IOpenRowset, ( IUnknown** )&pIOpenRowset );
    if( FAILED( hResult ) )
    {
        return hResult;
    }

    * ppIOpenRowset = pIOpenRowset.Detach( );
    return S_OK;
}

 

HRESULT CreateCommand( IOpenRowset* pIOpenRowset, ICommand** ppICommand, ICommandText** ppICommandText )
{
    ASSERT( pIOpenRowset != NULL && ppICommand != NULL && ppICommandText != NULL );

    HRESULT hResult;
    EAutoReleasePtr<ICommand> pICommand;

    {
        EAutoReleasePtr<IDBCreateCommand> pICreateCommand;
        hResult = pIOpenRowset->QueryInterface( IID_IDBCreateCommand, ( void** )&pICreateCommand );
        if( FAILED( hResult ) )
        {
            return hResult;
        }
  
        hResult = pICreateCommand->CreateCommand( NULL, IID_ICommand, (IUnknown**)&pICommand );
        if( FAILED( hResult ) )
        {
            return hResult;
        }
    }

    EAutoReleasePtr<ICommandText> pICommandText;
    hResult = pICommand->QueryInterface( &pICommandText );
    if( FAILED( hResult ) )
    {
        return hResult;
    }

    * ppICommand = pICommand.Detach( );
    * ppICommandText = pICommandText.Detach( );

    return S_OK;
}

 

HRESULT ExecuteSQL( ICommand* pICommand, ICommandText* pICommandText, LPCTSTR pszCommand, LONG* plRowsAffected )
{
    ASSERT( pICommand != NULL && pICommandText != NULL && pszCommand != NULL && pszCommand[0] != 0 );

 

    HRESULT hResult = pICommandText->SetCommandText( DBGUID_DBSQL, ( LPCOLESTR )pszCommand );

    if( FAILED( hResult ) )
    {
        return hResult;
    }

    LONG lAffected;
    hResult = pICommand->Execute( NULL, IID_NULL, NULL, plRowsAffected == NULL ? &lAffected : plRowsAffected, ( IUnknown** )NULL );
    return hResult;
}

 

以上就是写数据库的全部代码了，是不是很简单呢？下面再来读的。

 

// 先用与上面代码中一样的步骤获取pICommand，pICommandText。此处省略

 

HRESULT hResult = pICommandText->SetCommandText( DBGUID_DBSQL, ( LPCOLESTR )_T("SELECT Volume FROM 2005_1 WHERE ID = @@IDENTITY ") ); //取我们刚刚添加的那一条记录

if( FAILED( hResult ) )
{

    return;
}

 

LONG lAffected;
EAutoReleasePtr<IRowset> pIRowset;
hResult = pICommand->Execute( NULL, IID_IRowset, NULL, &lAffected, ( IUnknown** )&pIRowset );

if( FAILED( hResult ) )
{

    return;
}

EAutoReleasePtr<IAccessor> pIAccessor;
hResult = pIRowset->QueryInterface( IID_IAccessor, ( void** )&pIAccessor );

if( FAILED( hResult ) )
{

    return;
}

 

// 一个根据表中各字段的数值类型而定义的结构，用于存储返回的各字段的值

struct CLoadLastFromDB
{
     DBSTATUS dwdsVolume;
     DWORD     dwLenVolume;
     float          fVolume;

};

 

// 此处我们只查询了一个字段。如果要查询多个字段，CLoadLastFromDB中要添加相应的字段定义， 下面的dbBinding也要相应扩充。dbBinding[].iOrdinal要分别指向各个字段，dbBinding[].wType要 根据字段类型赋合适的值。

 

DBBINDING dbBinding[1];

dbBinding[0].iOrdinal            = 1;   // Volume 字段的位置，从 1 开始
dbBinding[0].obValue           = offsetof( CLoadLastFromDB, fVolume );
dbBinding[0].obLength         = offsetof( CLoadLastFromDB, dwLenVolume );
dbBinding[0].obStatus         = offsetof( CLoadLastFromDB, dwdsVolume );
dbBinding[0].pTypeInfo       = NULL;
dbBinding[0].pObject           = NULL;
dbBinding[0].pBindExt         = NULL;
dbBinding[0].dwPart            = DBPART_VALUE | DBPART_STATUS | DBPART_LENGTH;
dbBinding[0].dwMemOwner = DBMEMOWNER_CLIENTOWNED;
dbBinding[0].eParamIO       = DBPARAMIO_NOTPARAM;
dbBinding[0].cbMaxLen       = 0; 
dbBinding[0].dwFlags          = 0;
dbBinding[0].wType            = DBTYPE_R4; // float就是DBTYPE_R4，int就是DBTYPE_I4。参见MSDN
dbBinding[0].bPrecision       = 0;
dbBinding[0].bScale             = 0;

 

HACCESSOR hAccessor = DB_NULL_HACCESSOR;
DBBINDSTATUS dbs[1];
hResult = pIAccessor->CreateAccessor( DBACCESSOR_ROWDATA, 1, dbBinding, sizeof( CLoadLastDataFromDB ), &hAccessor, dbs );

if( FAILED( hResult ) )
{

    return;
}

ASSERT( dbs[0] == DBBINDSTATUS_OK );

ULONG uRowsObtained = 0;
HROW  hRows[1];                          // 这里我们只查询了最新的那一条记录
HROW* phRows = hRows;
CLoadLastFromDB rmd;

hResult = pIRowset->GetNextRows( NULL, 0, 1, &uRowsObtained, &phRows );
if( SUCCEEDED( hResult ) && uRowsObtained != 0U )
{
    hResult = pIRowset->GetData( phRows[0], hAccessor, &rmd );
    if( FAILED( hResult ) )
    {
        ASSERT( FALSE );
    }

    ASSERT( rmd.dwdsVolume == DBSTATUS_S_OK );

    // rmd.fVolume 就是我们要取的值
}

pIRowset->ReleaseRows( uRowsObtained, phRows, NULL, NULL, NULL );
pIAccessor->ReleaseAccessor( hAccessor, NULL );
pIAccessor.Release( );
pIRowset.Release( );

  

读操作也完成了，是不是仍然很简单呢？下面我们再来看看最麻烦的二进制数据（text、ntext、image等）的读写。要实现BLOB数据的读写，我们需要一个辅助的类，定义如下：

 

class CSequentialStream : public ISequentialStream    // BLOB 数据访问类
{
public:
    CSequentialStream( );
    virtual ~CSequentialStream( );

    virtual BOOL Seek( ULONG uPosition );
    virtual BOOL Clear( );
    virtual ULONG GetLength( ) { return m_uBufferUsed; };
    virtual operator void* const( ) { return m_pBuffer; };

    STDMETHODIMP_( ULONG ) AddRef( ) { return ++ m_uRefCount; };
    STDMETHODIMP_( ULONG ) Release( ) { ASSERT( m_uRefCount != 0U ); -- m_uRefCount; if( m_uRefCount == 0U ) { delete this; } return m_uRefCount; };
    STDMETHODIMP QueryInterface( REFIID riid, LPVOID* ppv );
    STDMETHODIMP Read( void __RPC_FAR* pv, ULONG cb, ULONG __RPC_FAR* pcbRead );
    STDMETHODIMP Write( const void __RPC_FAR* pv, ULONG cb, ULONG __RPC_FAR* pcbWritten );

 void    ResetPosition( ) { m_uPosition = 0U; };
 HRESULT PreAllocBuffer( ULONG uSize );

 

private:
    ULONG m_uRefCount;     // reference count
    void* m_pBuffer;           // buffer
    ULONG m_uBufferUsed;  // buffer used
    ULONG m_uBufferSize;   // buffer size
    ULONG m_uPosition;       // current index position in the buffer
};

 

实现如下：

 

CSequentialStream::CSequentialStream( ) : m_uRefCount( 0U ), m_pBuffer( NULL ), m_uBufferUsed( 0U ), m_uBufferSize( 0U ), m_uPosition( 0U )
{
    AddRef( );
}
 
CSequentialStream::~CSequentialStream( )
{
    Clear( );
}
 
HRESULT CSequentialStream::QueryInterface( REFIID riid, void** ppv )
{
    if( riid == IID_IUnknown || riid == IID_ISequentialStream )
    {
        * ppv = this;
        ( ( IUnknown* )*ppv )->AddRef( );
        return S_OK;
    }
    * ppv = NULL;
    return E_NOINTERFACE;
}

BOOL CSequentialStream::Seek( ULONG uPosition )
{
    ASSERT( uPosition < m_uBufferUsed );
    m_uPosition = uPosition;
    return TRUE;
}
 
BOOL CSequentialStream::Clear( )
{
    m_uBufferUsed = 0U;
    m_uBufferSize = 0U;
    m_uPosition = 0U;
    ( m_pBuffer != NULL ? CoTaskMemFree( m_pBuffer ) : 0 );
    m_pBuffer = NULL;

    return TRUE;
}

 

HRESULT CSequentialStream::PreAllocBuffer( ULONG uSize )
{
    if( m_uBufferSize < uSize )
    {
        m_uBufferSize = uSize;
        m_pBuffer = CoTaskMemRealloc( m_pBuffer, m_uBufferSize );
        if( m_pBuffer == NULL )
        {
            Clear( );
            return STG_E_INSUFFICIENTMEMORY;
        }
    }
    return S_OK;
}

 

HRESULT CSequentialStream::Read( void* pv, ULONG cb, ULONG* pcbRead )
{
    ( pcbRead != NULL ? ( * pcbRead = 0U ) : 0 );

    if( pv == NULL ) { return STG_E_INVALIDPOINTER; }
    if( cb == 0U ) { return S_OK; }

    ASSERT( m_uPosition <= m_uBufferUsed );
    ULONG uBytesLeft = m_uBufferUsed - m_uPosition;

 

    if( uBytesLeft == 0U ) { return S_FALSE; } //no more bytes

 

    ULONG uBytesRead = ( cb > uBytesLeft ? uBytesLeft : cb );
    memcpy( pv, ( BYTE* )m_pBuffer + m_uPosition, uBytesRead );

    m_uPosition += uBytesRead;
 
    ( pcbRead != NULL ? ( * pcbRead = uBytesRead ) : 0 );

    return ( cb != uBytesRead ? S_FALSE : S_OK );
}
 
HRESULT CSequentialStream::Write( const void* pv, ULONG cb, ULONG* pcbWritten )
{
    if( pv == NULL ) { return STG_E_INVALIDPOINTER; }
    ( pcbWritten != NULL ? ( * pcbWritten = 0U ) : 0 );

    if( cb == 0U ){ return S_OK; }

 

    ASSERT( m_uPosition <= m_uBufferUsed );
    if( m_uBufferSize < m_uPosition + cb )
    {
        m_uBufferSize = m_uPosition + cb;
        m_pBuffer = CoTaskMemRealloc( m_pBuffer, m_uBufferSize );
        if( m_pBuffer == NULL )
        {
            Clear( );
            return STG_E_INSUFFICIENTMEMORY;
        }
    }

    m_uBufferUsed = m_uPosition + cb;
    memcpy( ( BYTE* )m_pBuffer + m_uPosition, pv, cb );
    m_uPosition += cb;

    ( pcbWritten != NULL ? ( * pcbWritten = cb ) : 0 );
    return S_OK;
}

 

下面我们开始往一个包含ntext字段的表中添加记录。假设这个表（News）的结构为：ID int NOT NULL IDENTITY、Title nchar(80)、 Contents ntext。

 

// 先将记录添加进去，ntext字段留空。我们稍后再更新ntext的内容。

HRESULT hResult = ExecuteSQL( pICommand, pICommandText, _T("INSERT INTO News VALUES('TEST','')") );

 

DBPROP dbProp;
dbPropSet.guidPropertySet = DBPROPSET_ROWSET;
dbPropSet.cProperties        = 1;
dbPropSet.rgProperties       = &dbProp;

DBPROPSET dbPropSet;
dbPropSet.rgProperties[0].dwPropertyID     = DBPROP_UPDATABILITY;
dbPropSet.rgProperties[0].dwOptions         = DBPROPOPTIONS_REQUIRED;
dbPropSet.rgProperties[0].dwStatus           = DBPROPSTATUS_OK;
dbPropSet.rgProperties[0].colid                  = DB_NULLID;
dbPropSet.rgProperties[0].vValue.vt           = VT_I4;
V_I4( &dbPropSet.rgProperties[0].vValue ) = DBPROPVAL_UP_CHANGE;
   
EAutoReleasePtr<ICommandProperties> pICommandProperties;
hResult = pICommandText->QueryInterface( IID_ICommandProperties, ( void** )&pICommandProperties );

 

// 设置 Rowset 属性为“可以更新某字段的值”
hResult = pICommandProperties->SetProperties( 1, &dbPropSet );

 

hResult = pICommandText->SetCommandText( DBGUID_DBSQL, ( LPCOLESTR )L"SELECT Contents FROM News WHERE ID = @@IDENTITY" );

 

LONG lAffected;
EAutoReleasePtr<IRowsetChange> pIRowsetChange;
hResult = pICommand->Execute( NULL, IID_IRowsetChange, NULL, &lAffected, ( IUnknown** )&pIRowsetChange );

EAutoReleasePtr<IAccessor> pIAccessor;
hResult = pIRowsetChange->QueryInterface( IID_IAccessor, ( void** )&pIAccessor );
 
struct BLOBDATA
{
    DBSTATUS           dwStatus;
    DWORD              dwLength;
    ISequentialStream* pISeqStream;
};

 

// 有关DBOBJECT、DBBINDING的设置，建议参考MSDN，很容易懂。

DBOBJECT dbObj;
dbObj.dwFlags = STGM_READ;
dbObj.iid         = IID_ISequentialStream;

 

DBBINDING dbBinding;
dbBinding.iOrdinal   = 1;                              // BLOB 字段的位置，从 1 开始
dbBinding.obValue    = offsetof( BLOBDATA, pISeqStream );
dbBinding.obLength   = offsetof( BLOBDATA, dwLength );
dbBinding.obStatus   = offsetof( BLOBDATA, dwStatus );
dbBinding.pTypeInfo  = NULL;
dbBinding.pObject    = &dbObj;
dbBinding.pBindExt   = NULL;
dbBinding.dwPart     =  DBPART_VALUE | DBPART_STATUS | DBPART_LENGTH;
dbBinding.dwMemOwner = DBMEMOWNER_CLIENTOWNED;
dbBinding.eParamIO   = DBPARAMIO_NOTPARAM;
dbBinding.cbMaxLen   = 0;
dbBinding.dwFlags    = 0;
dbBinding.wType      = DBTYPE_IUNKNOWN;
dbBinding.bPrecision = 0;
dbBinding.bScale     = 0;

 

HACCESSOR hAccessor = DB_NULL_HACCESSOR;
DBBINDSTATUS dbs;
hResult = pIAccessor->CreateAccessor( DBACCESSOR_ROWDATA, 1, &dbBinding, sizeof( BLOBDATA ), &hAccessor, &dbs );

EAutoReleasePtr<IRowset> pIRowset;
hResult = pIRowsetChange->QueryInterface( IID_IRowset, ( void** )&pIRowset );

ULONG uRowsObtained = 0;
HROW* phRows = NULL;
hResult = pIRowset->GetNextRows( NULL, 0, 1, &uRowsObtained, &phRows );

CSequentialStream* pss = new CSequentialStream;
pss->PreAllocBuffer( 1024 );                           // 预先分配好内存，并读入数据

pss->Write( pszSomebuffer, 512, NULL );        // pss->Write可以连续调用

pss->Write( pszSomebuffer+512, 512, NULL );
pss->ResetPosition( );

BLOBDATA bd;
bd.pISeqStream = ( ISequentialStream* )pss;
bd.dwStatus    = DBSTATUS_S_OK;
bd.dwLength    = pss->GetLength( );

 

// 将 BLOB 数据写入到数据库
hResult = pIRowsetChange->SetData( phRows[0], hAccessor, &bd );

 

pIAccessor->ReleaseAccessor( hAccessor, NULL );
pIRowset->ReleaseRows( uRowsObtained, phRows, NULL, NULL, NULL );

 

// pss was released by pIRowsetChange->SetData.

 

这样，我们就完成了一条记录的添加。读取BLOB字段的代码跟上面的完全类似，只要把

  hResult = pIRowset->GetNextRows( NULL, 0, 1, &uRowsObtained, &phRows );

后面的那些改成下面的代码即可。

 

BLOBDATA bd;
hResult = pIRowset->GetData( phRows[0], hAccessor, &bd );
if( bd.dwStatus == DBSTATUS_S_ISNULL )
{
    // 此字段为空

}
else if( bd.dwStatus != DBSTATUS_S_OK || bd.pISeqStream == NULL )
{
    // 失败

}
else
{
    // 从系统分配的 ISequentialStream 接口读入 BLOB 数据

    BYTE szReadBuffer[1024];

    for( ULONG uRead = 0U; ; )
    {
        if( FAILED( bd.pISeqStream->Read( szReadBuffer, 1024, &uRead ) ) )
        {
            break;
        }
        //szReadBuffer中就包含了BLOB字段的数据
        if( uRead != 1024 )
        {
            break;
        }
    }

    bd.pISeqStream->Release( );

}
pIAccessor->ReleaseAccessor( hAccessor, NULL );
pIRowset->ReleaseRows( uRowsObtained, phRows, NULL, NULL, NULL );

 

至此，要讲的已全部讲完，希望对你能有所帮助。文中贴出的代码都是可以复制使用的，只是某些地方需要加入返回值判断、错误处理代码。