如果一次要向数据库服务器提交多条记录 , 通常会执行多次Insert命令 , 这样就为要插入的每个记录执行一次与数据库服务器的往返 , 这就给服务器增加了压力 , 效率也大大的降低了...
.Net FrameWork 2.0 新增功能 Bulk Copy 可以很快将大量数据加载到数据库中 , 现在利用这一新功能来实现上述功能.
这里从 MS Sql Server 2000 的 NorthWind 的 Orders 表加载数据到 DateTable 模拟要向数据库服务器提交的多条记录集 . 用 Tempdb 库来模拟目标数据库服务器 .
先在 Tempdb 建一个表 temp_orders
USE TEMPDB
CREATE TABLE TEMP_ORDERS
(
TEMP_ORDERID INT,
TEMP_CUSTOMERID NCHAR(5),
TEMP_ORDERDATE DATETIME,
TEMP_SHIPNAME NVARCHAR(40)
)
下面为模拟程序CREATE TABLE TEMP_ORDERS
(
TEMP_ORDERID INT,
TEMP_CUSTOMERID NCHAR(5),
TEMP_ORDERDATE DATETIME,
TEMP_SHIPNAME NVARCHAR(40)
)
protected void Page_Load(object sender, EventArgs e)
{
#region 从NorthWind的Orders表获取要插入的数据
DataTable dtNorthWindOrders = new DataTable();
using ( SqlConnection northWindConnection = new SqlConnection( "Data Source=.;Initial Catalog=NorthWind;Integrated Security=True" ) )
{
using ( SqlDataAdapter northWindAdapter = new SqlDataAdapter( "SELECT ORDERID,CUSTOMERID,ORDERDATE,SHIPNAME FROM ORDERS" , northWindConnection ) )
{
northWindAdapter.Fill( dtNorthWindOrders );
}
}
#endregion
using ( SqlConnection tempdbConnection = new SqlConnection( "Data Source=.;Initial Catalog=Tempdb;Integrated Security=True" ) )
{
tempdbConnection.Open( );
using ( SqlTransaction tran = tempdbConnection.BeginTransaction( ) )
{
SqlBulkCopy bulkCopyOrders = new SqlBulkCopy( tempdbConnection , SqlBulkCopyOptions.Default , tran );
bulkCopyOrders.DestinationTableName = "TEMP_ORDERS";
//将数据源表字段和目标表的字段做个映射
bulkCopyOrders.ColumnMappings.Add( "ORDERID" , "TEMP_ORDERID" );
bulkCopyOrders.ColumnMappings.Add( "CUSTOMERID" , "TEMP_CUSTOMERID" );
bulkCopyOrders.ColumnMappings.Add( "ORDERDATE" , "TEMP_ORDERDATE" );
bulkCopyOrders.ColumnMappings.Add( "SHIPNAME" , "TEMP_SHIPNAME" );
bulkCopyOrders.BulkCopyTimeout = 1000;
//每处理10行触发一个事件向页面上输出一个消息
bulkCopyOrders.SqlRowsCopied += new SqlRowsCopiedEventHandler( onRowsCopy );
bulkCopyOrders.NotifyAfter = 10;
try
{
bulkCopyOrders.WriteToServer( dtNorthWindOrders );
tran.Commit( );
}
catch ( Exception ex )
{
Response.Write( ex.ToString( ) );
}
finally
{
dtNorthWindOrders = null;
}
}
}
}
private void onRowsCopy ( object Sender , SqlRowsCopiedEventArgs args )
{
Response.Write("已复制:<font color=red>"+ args.RowsCopied.ToString( ) + "</font><br />" );
}
通过SQL SERVER 事件探察器发现执行的SQL为:{
#region 从NorthWind的Orders表获取要插入的数据
DataTable dtNorthWindOrders = new DataTable();
using ( SqlConnection northWindConnection = new SqlConnection( "Data Source=.;Initial Catalog=NorthWind;Integrated Security=True" ) )
{
using ( SqlDataAdapter northWindAdapter = new SqlDataAdapter( "SELECT ORDERID,CUSTOMERID,ORDERDATE,SHIPNAME FROM ORDERS" , northWindConnection ) )
{
northWindAdapter.Fill( dtNorthWindOrders );
}
}
#endregion
using ( SqlConnection tempdbConnection = new SqlConnection( "Data Source=.;Initial Catalog=Tempdb;Integrated Security=True" ) )
{
tempdbConnection.Open( );
using ( SqlTransaction tran = tempdbConnection.BeginTransaction( ) )
{
SqlBulkCopy bulkCopyOrders = new SqlBulkCopy( tempdbConnection , SqlBulkCopyOptions.Default , tran );
bulkCopyOrders.DestinationTableName = "TEMP_ORDERS";
//将数据源表字段和目标表的字段做个映射
bulkCopyOrders.ColumnMappings.Add( "ORDERID" , "TEMP_ORDERID" );
bulkCopyOrders.ColumnMappings.Add( "CUSTOMERID" , "TEMP_CUSTOMERID" );
bulkCopyOrders.ColumnMappings.Add( "ORDERDATE" , "TEMP_ORDERDATE" );
bulkCopyOrders.ColumnMappings.Add( "SHIPNAME" , "TEMP_SHIPNAME" );
bulkCopyOrders.BulkCopyTimeout = 1000;
//每处理10行触发一个事件向页面上输出一个消息
bulkCopyOrders.SqlRowsCopied += new SqlRowsCopiedEventHandler( onRowsCopy );
bulkCopyOrders.NotifyAfter = 10;
try
{
bulkCopyOrders.WriteToServer( dtNorthWindOrders );
tran.Commit( );
}
catch ( Exception ex )
{
Response.Write( ex.ToString( ) );
}
finally
{
dtNorthWindOrders = null;
}
}
}
}
private void onRowsCopy ( object Sender , SqlRowsCopiedEventArgs args )
{
Response.Write("已复制:<font color=red>"+ args.RowsCopied.ToString( ) + "</font><br />" );
}
insert bulk TEMP_ORDERS ([TEMP_ORDERID] Int, [TEMP_CUSTOMERID] NChar(5) COLLATE Chinese_PRC_CI_AS, [TEMP_ORDERDATE] DateTime, [TEMP_SHIPNAME] NVarChar(40) COLLATE Chinese_PRC_CI_AS)
通过运行程序可以看出这个速度是相当的快 , 使用这个方法的最大优点是 : 减少对数据库的访问次数 .WriteToServer不仅可以处理 DataTable 对象 , 还可以处理 DataReader , DataRow 对象数组 .