Visual C++中使用OLE DB读写SQL Server简明指南

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 );