string connetionString = null;
SqlConnection connection;
SqlCommand command;
SqlDataAdapter adpter = new SqlDataAdapter();
DataSet ds = new DataSet();
XmlReader xmlFile;
string sql = null;
int ID = 0;
string Name = null, Text = null, Screenname = null;
connetionString = "myconnection";
connection = new SqlConnection(connetionString);
xmlFile = XmlReader.Create("my.XML", new XmlReaderSettings());
ds.ReadXml(xmlFile);
int i = 0;
connection.Open();
for (i = 0; i <= ds.Tables[0].Rows.Count - 1; i++)
{
ID = Convert.ToInt32(ds.Tables[0].Rows[i].ItemArray[0]);
Text = ds.Tables[0].Rows[i].ItemArray[1].ToString().Replace("'", "''");
Name = ds.Tables[0].Rows[i].ItemArray[2].ToString().Replace("'", "''");
Screenname = ds.Tables[0].Rows[i].ItemArray[3].ToString().Replace("'", "''");
//sql = "insert into nicktest values(" + ID + ",'" + Text + "'," + Name + "," + Screenname + "," + DateTime.Now.ToString() + ")";
sql = "If Exists(Select * from niktest2 Where Id = ID) " +
" BEGIN " +
" update niktest2 set Name = '" + Text + "' , Screenname = '" + Name + "', Profimg= '" + Screenname + "', InsertDateTime= '" + DateTime.Now.ToString() + "' where Id=ID" +
" END " +
" ELSE " +
" BEGIN " +
" insert into niktest2(Id,Name,Screenname,Profimg,InsertDateTime) values('" + ID + "','" + Text + "','" + Name + "','" + Screenname + "' ,'" + DateTime.Now.ToString() + "')" +
" END ";
command = new SqlCommand(sql, connection);
adpter.InsertCommand = command;
adpter.InsertCommand.ExecuteNonQuery();
}
}
运行此代码后,即使我的xml文件有更多数据,第一行也会更新。 我想将所有数据插入到数据库中,并在xml文件中为其分配id。 请帮忙..
答案 0 :(得分:4)
只要插入一行,就会出现这种情况:
If Exists(Select * from niktest2 Where Id = ID)
因此,您将执行更新,而不是插入,因此您只能在数据库中获得一行。
由于您使用的是SQL Server 2008,我会采用完全不同的方法,使用Parameterised queries,MERGE和table valued parameters。
第一步是创建表值参数(我不得不猜测你的类型:
CREATE TYPE dbo.nicktestTableType AS TABLE
(
Id INT,
Name VARCHAR(255),
Screenname VARCHAR(255),
Profimg VARCHAR(255)
);
然后,您可以编写MERGE
语句以插入数据库:
MERGE nicktest WITH (HOLDLOCK) AS t
USING @NickTestType AS s
ON s.ID = t.ID
WHEN MATCHED THEN
UPDATE
SET Name = s.Name,
Screenname = s.Screenname,
Profimg = s.Profimg,
InsertDateTime = GETDATE()
WHEN NOT MATCHED THEN
INSERT (Id, Name, Screenname, Profimg, InsertDateTime)
VALUES (s.Id, s.Name, s.Screenname, s.Profimg, GETDATE());
然后,您可以将数据表作为参数传递给查询:
using (var command = new SqlCommand(sql, connection))
{
var parameter = new SqlParameter("@NickTestType", SqlDbType.Structured);
parameter.Value = ds.Tables[0];
parameter.TypeName = "dbo.nicktestTableType";
command.Parameters.Add(parameter);
command.ExecuteNonQuery();
}
如果您不想做出如此大的改动,那么您至少应该使用参数化查询,因此您的SQL将是:
IF EXISTS (SELECT 1 FROM nicktest WHERE ID = @ID)
BEGIN
UPDATE nicktest
SET Name = @Name,
ScreenName = @ScreeName,
InsertDateTime = GETDATE()
WHERE ID = @ID;
END
ELSE
BEGIN
INSERT (Id, Name, Screenname, Profimg, InsertDateTime)
VALUES (@ID, @Name, @Screenname, @Profimg, GETDATE());
END
或者最好仍然使用MERGE
作为HOLDLOCK
表提示阻止(或至少大大降低了)race condition的可能性:
MERGE nicktest WITH (HOLDLOCK) AS t
USING (VALUES (@ID, @Name, @ScreenName, @ProfImg)) AS s (ID, Name, ScreenName, ProfImg)
ON s.ID = t.ID
WHEN MATCHED THEN
UPDATE
SET Name = s.Name,
Screenname = s.Screenname,
Profimg = s.Profimg,
InsertDateTime = GETDATE()
WHEN NOT MATCHED THEN
INSERT (Id, Name, Screenname, Profimg, InsertDateTime)
VALUES (s.Id, s.Name, s.Screenname, s.Profimg, GETDATE());
虽然使用表值参数,但效率远低于第一种解决方案 然后你的c#会是这样的:
for (i = 0; i <= ds.Tables[0].Rows.Count - 1; i++)
{
using (var command = new SqlCommand(sql, connection))
{
command.Parameters.AddWithValue("@ID", ds.Tables[0].Rows[i][0]);
command.Parameters.AddWithValue("@Name", ds.Tables[0].Rows[i][1]);
command.Parameters.AddWithValue("@ScreeName", ds.Tables[0].Rows[i][2]);
command.Parameters.AddWithValue("@ProfImg", ds.Tables[0].Rows[i][3]);
command.ExecuteNonQuery();
}
}