原文链接:http://apps.hi.baidu.com/share/detail/16973788
最近在将excel中的文件导入到数据库中,用程序进行编写,由于数据量较大所以速度很慢,后来采用了SqlBulkCopy类,解决了速度的问题,我就insert语句,sqldataadapter.update(dataset,tablename);sqlbulkcopy.WriteToServer(datatable);三个方法的性能进行比较:
1、生成测试的datatable表,表结构如下:
UniqueID(主键,自动增长) | CompanyName | CompanyCode | Address | Owner | Memo
共6个字段。
SqlConnection sqlconnection = new SqlConnection(connectionString);
SqlDataAdapter sqldataadapter = new SqlDataAdapter("select * from Table_1 where 1=2", sqlconnection);
DataSet dataset = new DataSet();
sqldataadapter.Fill(dataset, "Table_1");
DataTable datatable = dataset.Tables[0];
//生成20000条记录
for (int i = 0; i < 20000; i++)
{
DataRow datarow = datatable.NewRow();
datarow["CompanyName"] = "companyname"+string.Format("{0:0000}",i);
datarow["CompanyCode"] = "companycode" + string.Format("{0:0000}", i);
datarow["Address"] = "address" + string.Format("{0:0000}", i);
datarow["Owner"] = "owner" + string.Format("{0:0000}", i);
datarow["Memo"] = "memo" + string.Format("{0:0000}", i);
datatable.Rows.Add(datarow);
}
2、使用sqlcommand.executenonquery()方法插入
foreach (DataRow datarow in datatable.Rows)
{
string sql = "INSERT INTO [Table_1]
([CompanyName],[CompanyCode],[Address],[Owner],[Memo])" +
"VALUES('" + datarow["CompanyName"].ToString() + "'" +
",'" + datarow["CompanyCode"].ToString() + "'" +
",'" + datarow["Address"].ToString() + "'" +
",'" + datarow["Owner"].ToString() + "'" +
",'" + datarow["Memo"].ToString() + "')";
using (SqlConnection sqlconn = new SqlConnection(connectionString))
{
sqlconn.Open();
SqlCommand sqlcommand = new SqlCommand(sql, sqlconn);
sqlcommand.ExecuteNonQuery();
sqlconn.Close();
}
}
插入20000条记录时间:00:00:29.7336000
3、使用sqldataadapter.update(dataset,tablename);
SqlCommand insertcommand = new SqlCommand("INSERT INTO [Table_1]([CompanyName],[CompanyCode],[Address],[Owner],[Memo])" +
"VALUES(@CompanyName, @CompanyCode,@Address,@Owner,@Memo)",new SqlConnection(connectionString));
insertcommand.Parameters.Add("@CompanyName", SqlDbType.NChar, 50, "CompanyName");
insertcommand.Parameters.Add("@CompanyCode", SqlDbType.NChar, 25, "CompanyCode");
insertcommand.Parameters.Add("@Address", SqlDbType.NChar, 255, "Address");
insertcommand.Parameters.Add("@Owner", SqlDbType.NChar, 25, "Owner");
insertcommand.Parameters.Add("@Memo", SqlDbType.NChar, 255, "Memo");
sqldataadapter.InsertCommand = insertcommand;
sqldataadapter.Update(dataset, "Table_1");
插入20000条记录时间:00:00:22.8938000
4、使用sqlbulkcopy.writetoserver(datatable)
SqlBulkCopy sqlbulkcopy = new SqlBulkCopy(connectionString, SqlBulkCopyOptions.UseInternalTransaction);
sqlbulkcopy.DestinationTableName = "Table_1";//数据库中的表名
sqlbulkcopy.WriteToServer(dataset.Tables[0]);
插入20000条记录时间:00:00:00.3276000
所以说速度是sqlbulkcopy最快,sqldataadapter.update()次之,sqlcommand.ExecuteNonQuery()最慢。
补充:听说在sql2008中还有一种批量写入的方法,具体情况不太清楚。