SQL 2005批量插入数据的二种方法

SQL 2005批量插入数据的二种方法

Posted on 2010-07-22 18:13 moss_tan_jun 阅读(2635) 评论(2) 编辑 收藏

　　　　

在SQL Server 中插入一条数据使用Insert语句，但是如果想要批量插入一堆数据的话，循环使用Insert不仅效率低，而且会导致SQL一系统性能问题。下面介绍SQL Server支持的两种批量数据插入方法：Bulk和表值参数(Table-Valued Parameters)。

运行下面的脚本，建立测试数据库和表值参数。

view plaincopy to clipboardprint?
--Create DataBase   
create database BulkTestDB;   
go   
use BulkTestDB;   
go   
--Create Table   
Create table BulkTestTable(   
Id int primary key,   
UserName nvarchar(32),   
Pwd varchar(16))   
go   
--Create Table Valued   
CREATE TYPE BulkUdt AS TABLE   
  (Id int,   
   UserName nvarchar(32),   
   Pwd varchar(16))  
--Create DataBase
create database BulkTestDB;
go
use BulkTestDB;
go
--Create Table
Create table BulkTestTable(
Id int primary key,
UserName nvarchar(32),
Pwd varchar(16))
go
--Create Table Valued
CREATE TYPE BulkUdt AS TABLE
  (Id int,
   UserName nvarchar(32),
   Pwd varchar(16))

下面我们使用最简单的Insert语句来插入100万条数据，代码如下：

view plaincopy to clipboardprint?
Stopwatch sw = new Stopwatch();   
  
SqlConnection sqlConn = new SqlConnection(   
    ConfigurationManager.ConnectionStrings["ConnStr"].ConnectionString);//连接数据库   
  
SqlCommand sqlComm = new SqlCommand();   
sqlComm.CommandText = string.Format("insert into BulkTestTable(Id,UserName,Pwd)values(@p0,@p1,@p2)");//参数化SQL   
sqlComm.Parameters.Add("@p0", SqlDbType.Int);   
sqlComm.Parameters.Add("@p1", SqlDbType.NVarChar);   
sqlComm.Parameters.Add("@p2", SqlDbType.VarChar);   
sqlComm.CommandType = CommandType.Text;   
sqlComm.Connection = sqlConn;   
sqlConn.Open();   
try  
{   
    //循环插入100万条数据，每次插入10万条，插入10次。   
    for (int multiply = 0; multiply < 10; multiply++)   
    {   
        for (int count = multiply * 100000; count < (multiply + 1) * 100000; count++)   
        {   
  
            sqlComm.Parameters["@p0"].Value = count;   
            sqlComm.Parameters["@p1"].Value = string.Format("User-{0}", count * multiply);   
            sqlComm.Parameters["@p2"].Value = string.Format("Pwd-{0}", count * multiply);   
            sw.Start();   
            sqlComm.ExecuteNonQuery();   
            sw.Stop();   
        }   
        //每插入10万条数据后，显示此次插入所用时间   
        Console.WriteLine(string.Format("Elapsed Time is {0} Milliseconds", sw.ElapsedMilliseconds));   
    }   
}   
catch (Exception ex)   
{   
    throw ex;   
}   
finally  
{   
    sqlConn.Close();   
}   
  
Console.ReadLine();  
            Stopwatch sw = new Stopwatch();

SqlConnection sqlConn = new SqlConnection(
                ConfigurationManager.ConnectionStrings["ConnStr"].ConnectionString);//连接数据库

SqlCommand sqlComm = new SqlCommand();
            sqlComm.CommandText = string.Format("insert into BulkTestTable(Id,UserName,Pwd)values(@p0,@p1,@p2)");//参数化SQL
            sqlComm.Parameters.Add("@p0", SqlDbType.Int);
            sqlComm.Parameters.Add("@p1", SqlDbType.NVarChar);
            sqlComm.Parameters.Add("@p2", SqlDbType.VarChar);
            sqlComm.CommandType = CommandType.Text;
            sqlComm.Connection = sqlConn;
            sqlConn.Open();
            try
            {
                //循环插入100万条数据，每次插入10万条，插入10次。
                for (int multiply = 0; multiply < 10; multiply++)
                {
                    for (int count = multiply * 100000; count < (multiply + 1) * 100000; count++)
                    {

sqlComm.Parameters["@p0"].Value = count;
                        sqlComm.Parameters["@p1"].Value = string.Format("User-{0}", count * multiply);
                        sqlComm.Parameters["@p2"].Value = string.Format("Pwd-{0}", count * multiply);
                        sw.Start();
                        sqlComm.ExecuteNonQuery();
                        sw.Stop();
                    }
                    //每插入10万条数据后，显示此次插入所用时间
                    Console.WriteLine(string.Format("Elapsed Time is {0} Milliseconds", sw.ElapsedMilliseconds));
                }
            }
            catch (Exception ex)
            {
                throw ex;
            }
            finally
            {
                sqlConn.Close();
            }

Console.ReadLine();

耗时图如下：

由于运行过慢，才插入10万条就耗时72390 milliseconds，所以我就手动强行停止了。

下面看一下使用Bulk插入的情况：

bulk方法主要思想是通过在客户端把数据都缓存在Table中，然后利用SqlBulkCopy一次性把Table中的数据插入到数据库

代码如下：

view plaincopy to clipboardprint?
public static void BulkToDB(DataTable dt)   
{   
    SqlConnection sqlConn = new SqlConnection(   
        ConfigurationManager.ConnectionStrings["ConnStr"].ConnectionString);   
    SqlBulkCopy bulkCopy = new SqlBulkCopy(sqlConn);   
    bulkCopy.DestinationTableName = "BulkTestTable";   
    bulkCopy.BatchSize = dt.Rows.Count;   
  
    try  
    {   
        sqlConn.Open();   
    if (dt != null && dt.Rows.Count != 0)   
        bulkCopy.WriteToServer(dt);   
    }   
    catch (Exception ex)   
    {   
        throw ex;   
    }   
    finally  
    {   
        sqlConn.Close();   
        if (bulkCopy != null)   
            bulkCopy.Close();   
    }   
}   
  
public static DataTable GetTableSchema()   
{   
    DataTable dt = new DataTable();   
    dt.Columns.AddRange(new DataColumn[]{   
        new DataColumn("Id",typeof(int)),   
        new DataColumn("UserName",typeof(string)),   
    new DataColumn("Pwd",typeof(string))});   
  
    return dt;   
}   
  
static void Main(string[] args)   
{   
    Stopwatch sw = new Stopwatch();   
    for (int multiply = 0; multiply < 10; multiply++)   
    {   
        DataTable dt = Bulk.GetTableSchema();   
        for (int count = multiply * 100000; count < (multiply + 1) * 100000; count++)   
        {   
            DataRow r = dt.NewRow();   
            r[0] = count;   
            r[1] = string.Format("User-{0}", count * multiply);   
            r[2] = string.Format("Pwd-{0}", count * multiply);   
            dt.Rows.Add(r);   
        }   
        sw.Start();   
        Bulk.BulkToDB(dt);   
        sw.Stop();   
        Console.WriteLine(string.Format("Elapsed Time is {0} Milliseconds", sw.ElapsedMilliseconds));   
    }   
  
    Console.ReadLine();   
}  
public static void BulkToDB(DataTable dt)
{
    SqlConnection sqlConn = new SqlConnection(
        ConfigurationManager.ConnectionStrings["ConnStr"].ConnectionString);
    SqlBulkCopy bulkCopy = new SqlBulkCopy(sqlConn);
    bulkCopy.DestinationTableName = "BulkTestTable";
    bulkCopy.BatchSize = dt.Rows.Count;

try
    {
        sqlConn.Open();
 if (dt != null && dt.Rows.Count != 0)
     bulkCopy.WriteToServer(dt);
    }
    catch (Exception ex)
    {
        throw ex;
    }
    finally
    {
        sqlConn.Close();
        if (bulkCopy != null)
            bulkCopy.Close();
    }
}

public static DataTable GetTableSchema()
{
    DataTable dt = new DataTable();
    dt.Columns.AddRange(new DataColumn[]{
        new DataColumn("Id",typeof(int)),
        new DataColumn("UserName",typeof(string)),
 new DataColumn("Pwd",typeof(string))});

return dt;
}

static void Main(string[] args)
{
    Stopwatch sw = new Stopwatch();
    for (int multiply = 0; multiply < 10; multiply++)
    {
        DataTable dt = Bulk.GetTableSchema();
        for (int count = multiply * 100000; count < (multiply + 1) * 100000; count++)
        {
            DataRow r = dt.NewRow();
            r[0] = count;
            r[1] = string.Format("User-{0}", count * multiply);
            r[2] = string.Format("Pwd-{0}", count * multiply);
            dt.Rows.Add(r);
        }
        sw.Start();
        Bulk.BulkToDB(dt);
        sw.Stop();
        Console.WriteLine(string.Format("Elapsed Time is {0} Milliseconds", sw.ElapsedMilliseconds));
    }

Console.ReadLine();
}

耗时图如下：

可见，使用Bulk后，效率和性能明显上升。使用Insert插入10万数据耗时72390，而现在使用Bulk插入100万数据才耗时17583。

最后再看看使用表值参数的效率，会另你大为惊讶的。

表值参数是SQL Server 2008新特性，简称TVPs。对于表值参数不熟悉的朋友，可以参考最新的book online，我也会另外写一篇关于表值参数的博客，不过此次不对表值参数的概念做过多的介绍。言归正传，看代码：

view plaincopy to clipboardprint?
public static void TableValuedToDB(DataTable dt)   
{   
    SqlConnection sqlConn = new SqlConnection(   
      ConfigurationManager.ConnectionStrings["ConnStr"].ConnectionString);   
    const string TSqlStatement =   
     "insert into BulkTestTable (Id,UserName,Pwd)" +   
     " SELECT nc.Id, nc.UserName,nc.Pwd" +   
     " FROM @NewBulkTestTvp AS nc";   
    SqlCommand cmd = new SqlCommand(TSqlStatement, sqlConn);   
    SqlParameter catParam = cmd.Parameters.AddWithValue("@NewBulkTestTvp", dt);   
    catParam.SqlDbType = SqlDbType.Structured;   
    //表值参数的名字叫BulkUdt，在上面的建立测试环境的SQL中有。   
    catParam.TypeName = "dbo.BulkUdt";   
    try  
    {   
      sqlConn.Open();   
      if (dt != null && dt.Rows.Count != 0)   
      {   
          cmd.ExecuteNonQuery();   
      }   
    }   
    catch (Exception ex)   
    {   
      throw ex;   
    }   
    finally  
    {   
      sqlConn.Close();   
    }   
}   
  
public static DataTable GetTableSchema()   
{   
    DataTable dt = new DataTable();   
    dt.Columns.AddRange(new DataColumn[]{   
      new DataColumn("Id",typeof(int)),   
      new DataColumn("UserName",typeof(string)),   
      new DataColumn("Pwd",typeof(string))});   
  
    return dt;   
}   
  
static void Main(string[] args)   
{   
    Stopwatch sw = new Stopwatch();   
    for (int multiply = 0; multiply < 10; multiply++)   
    {   
        DataTable dt = TableValued.GetTableSchema();   
        for (int count = multiply * 100000; count < (multiply + 1) * 100000; count++)   
        {           
            DataRow r = dt.NewRow();   
            r[0] = count;   
            r[1] = string.Format("User-{0}", count * multiply);   
            r[2] = string.Format("Pwd-{0}", count * multiply);   
            dt.Rows.Add(r);   
        }   
        sw.Start();   
        TableValued.TableValuedToDB(dt);   
        sw.Stop();   
        Console.WriteLine(string.Format("Elapsed Time is {0} Milliseconds", sw.ElapsedMilliseconds));   
    }   
  
    Console.ReadLine();   
}  
public static void TableValuedToDB(DataTable dt)
{
    SqlConnection sqlConn = new SqlConnection(
      ConfigurationManager.ConnectionStrings["ConnStr"].ConnectionString);
    const string TSqlStatement =
     "insert into BulkTestTable (Id,UserName,Pwd)" +
     " SELECT nc.Id, nc.UserName,nc.Pwd" +
     " FROM @NewBulkTestTvp AS nc";
    SqlCommand cmd = new SqlCommand(TSqlStatement, sqlConn);
    SqlParameter catParam = cmd.Parameters.AddWithValue("@NewBulkTestTvp", dt);
    catParam.SqlDbType = SqlDbType.Structured;
    //表值参数的名字叫BulkUdt，在上面的建立测试环境的SQL中有。
    catParam.TypeName = "dbo.BulkUdt";
    try
    {
      sqlConn.Open();
      if (dt != null && dt.Rows.Count != 0)
      {
          cmd.ExecuteNonQuery();
      }
    }
    catch (Exception ex)
    {
      throw ex;
    }
    finally
    {
      sqlConn.Close();
    }
}

public static DataTable GetTableSchema()
{
    DataTable dt = new DataTable();
    dt.Columns.AddRange(new DataColumn[]{
      new DataColumn("Id",typeof(int)),
      new DataColumn("UserName",typeof(string)),
      new DataColumn("Pwd",typeof(string))});

return dt;
}

static void Main(string[] args)
{
    Stopwatch sw = new Stopwatch();
    for (int multiply = 0; multiply < 10; multiply++)
    {
        DataTable dt = TableValued.GetTableSchema();
        for (int count = multiply * 100000; count < (multiply + 1) * 100000; count++)
        {        
            DataRow r = dt.NewRow();
            r[0] = count;
            r[1] = string.Format("User-{0}", count * multiply);
            r[2] = string.Format("Pwd-{0}", count * multiply);
            dt.Rows.Add(r);
        }
        sw.Start();
        TableValued.TableValuedToDB(dt);
        sw.Stop();
        Console.WriteLine(string.Format("Elapsed Time is {0} Milliseconds", sw.ElapsedMilliseconds));
    }

Console.ReadLine();
}