.NET批量大数据插入性能分析及比较(4.使用DataAdapter批量插入)

#region 使用DataAdapter

public static bool ExecuteDataAdapterInsert(DataTable dt, int batchSize)

{

int count = dt.Rows.Count;

bool flag = false;

try

{

SqlConnection cn = new SqlConnection(connectionString);

SqlCommand cmd = new SqlCommand("Insert into TestTable(Id, Name) Values(@Id, @Name)", cn);

cmd.Parameters.Add("@Id", SqlDbType.Int, 4, "Id");

cmd.Parameters.Add("@Name", SqlDbType.NVarChar, 20, "Name");

cmd.UpdatedRowSource = UpdateRowSource.None;

SqlDataAdapter da = new SqlDataAdapter();

da.InsertCommand = cmd;

da.UpdateBatchSize = batchSize;

cn.Open();

da.Update(dt);

cn.Close();

flag = true;

}

catch (Exception ex)

{

LogHelper.Error(ex.Message);

return false;

}

return flag;

}

#endregion

结果如下：
Use SqlServer DataAdapter Insert;RecordCount:40000;BatchSize:10;Time:21569;
Use SqlServer DataAdapter Insert;RecordCount:40000;BatchSize:20;Time:19607;
Use SqlServer DataAdapter Insert;RecordCount:40000;BatchSize:50;Time:19647;
Use SqlServer DataAdapter Insert;RecordCount:40000;BatchSize:100;Time:18649;
Use SqlServer DataAdapter Insert;RecordCount:40000;BatchSize:200;Time:21823;

测试的结果让我比较失望，SqlDataAdapter会为每个修改行发送一条T-SQL命令，所以效率并不高。

而且再其他的方式中，都可以进行异步处理，SqlDataAdapter的Update()方法并不支持异步的结果，所以此方法不推荐