本文以一则示例说明一般技术人员演示demo与实际应用开发之间的差距。
假设情况如下：一个项目当中需要用linq to sql 来作为数据访问层，如果未用过Linq to SQL 的话，一般大家会以DEMO示例 说明一张表的增、删、改、查的情况。
第一步 建好数据库表，假设建表脚本为
代码
[SequenceId] [int] IDENTITY(1,1) NOT NULL,
[MerId] [nvarchar](50) NOT NULL,
[MerDate] [nvarchar](50) NOT NULL,
[MerSeqId] [nvarchar](50) NULL,
[CardNo] [nvarchar](50) NOT NULL,
[UsrName] [nvarchar](50) NOT NULL,
[OpenBank] [nvarchar](50) NOT NULL,
[Prov] [nvarchar](50) NOT NULL,
[City] [nvarchar](50) NOT NULL,
[TransAmt] [decimal](8, 2) NOT NULL,
[Purpose] [nvarchar](100) NULL,
[Flag] [nvarchar](5) NULL,
[SubBank] [nvarchar](50) NULL,
[Version] [nvarchar](10) NULL,
[Submited] [bit] NULL,
[SubmitedResult] [nvarchar](50) NULL,
[SubmitedTime] [datetime] NULL,
[responsed] [bit] NULL,
[responseCode] [nvarchar](50) NULL,
[responseTime] [datetime] NULL,
[oracontrasted] [bit] NULL,
[oracontrastedTime] [datetime] NULL,
[oracontrastedResult] [nvarchar](50) NULL,
[EticketNo] [nvarchar](50) NULL,
[State] [nvarchar](50) NULL,
CONSTRAINT [PK_NetPay_Order] PRIMARY KEY CLUSTERED
(
[SequenceId] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
其中SequenceId是自增序列
第二步 打开Visual Studio.NET 2010,新建控制台项目TestLinqtoSQL ，然后依次点击Tool->Connect to Database 输入刚刚建好的数据库相关信息，在IDE 左侧看到数据库相关信息
第三步 新建Linq to SQL Classes ，右击项目依次点击add ->add item->linq to sql classes 并命名为DataAccess，在第二步当中看到的数据库信息找到第一步建好的表，直接拖到新建的linq to SQL Classes 上来。
第四步 前期工作完成，现在我们可以开始写代码了，我们要做的新增一条数据，我们会这样写
代码
NetPay_Order recorder = new NetPay_Order();
recorder.MerId = "0001";
recorder.MerDate = System.DateTime.Now.ToString("yyyyMMdd");
recorder.MerSeqId = null;
recorder.CardNo = "0002";
recorder.UsrName = "赵一村";
recorder.City = "苏州";
recorder.Prov = "江苏";
recorder.Purpose = "转账";
recorder.EticketNo = System.Guid.NewGuid().ToString();
recorder.OpenBank = "招商银行";
recorder.Flag = "00";
recorder.Version = "20100906";
recorder.Submited = false;
recorder.responsed = false;
recorder.oracontrasted = false;
_context.NetPay_Orders.InsertOnSubmit(recorder);
_context.SubmitChanges();
这样子，新增一笔数据就搞定了，用linq to SQL新增数据就这么简单。
可是这只能算是Demo 不足以实际就应用
遇到的问题1 批量操作与事务性，这里只是新增一笔数据，实际开发当中经常会有批量新增的事情，经常修改，代码会变成如下样子
{
DataAccessDataContext _context = new DataAccessDataContext();
foreach (NetPay_Order recorder in getOrders())
{
_context.NetPay_Orders.InsertOnSubmit(recorder);
}
_context.SubmitChanges();
}
private static NetPay_Order[] getOrders()
{
NetPay_Order[] recorders = new NetPay_Order[2];
NetPay_Order recorder1,recorder2;
recorder1 = new NetPay_Order();
recorder1.MerId = "1001";
recorder1.MerDate = System.DateTime.Now.ToString("yyyyMMdd");
recorder1.MerSeqId = null;
recorder1.CardNo = "0002";
recorder1.UsrName = "赵一村";
recorder1.City = "苏州";
recorder1.Prov = "江苏";
recorder1.Purpose = "转账";
recorder1.EticketNo = System.Guid.NewGuid().ToString();
recorder1.OpenBank = "招商银行";
recorder1.Flag = "00";
recorder1.Version = "20100906";
recorder1.Submited = false;
recorder1.responsed = false;
recorder1.oracontrasted = false;
recorders[0] = recorder1;
recorder2 = new NetPay_Order();
// recorder2.MerId = "1002";
recorder2.MerDate = System.DateTime.Now.ToString("yyyyMMdd");
recorder2.MerSeqId = null;
recorder2.CardNo = "0002";
recorder2.UsrName = "赵一村";
recorder2.City = "苏州";
recorder2.Prov = "江苏";
recorder2.Purpose = "转账";
recorder2.EticketNo = System.Guid.NewGuid().ToString();
recorder2.OpenBank = "招商银行";
recorder2.Flag = "00";
recorder2.Version = "20100906";
recorder2.Submited = false;
recorder2.responsed = false;
recorder2.oracontrasted = false;
recorders[1] = recorder2;
return recorders;
}
其中DataAccessDataContext _context = new DataAccessDataContext();
foreach (NetPay_Order recorder in getOrders())
{
_context.NetPay_Orders.InsertOnSubmit(recorder);
}
_context.SubmitChanges();
这几段代码中畎认情况linq to SQL 会为你提供事务功能
遇到的问题2 假如MerSeqId字段是根据MerDate+ SequenceId 两字字段组合起来作为流水号呢，我们知道程序热行到_context.NetPay_Orders.InsertOnSubmit(recorder)这一句时并没有新增数据库，而是要_context.SubmitChanges()以后才可以更新数据库，可谓有真正更新数据库的时候SequenceId 字段才会真正有值，此时照平常想法，代码会变成如下样子
DataAccessDataContext _context = new DataAccessDataContext();
foreach (NetPay_Order recorder in getOrders())
{
_context.NetPay_Orders.InsertOnSubmit(recorder);
_context.SubmitChanges();
recorder.MerSeqId = recorder.MerDate.ToString() + recorder.SequenceId.ToString();
_context.SubmitChanges();
}
这样子_context.SubmitChanges()以后再更新MerSeqId字段值然后再SubmitChanges一次，到这里，好像问题已经解决了。
可是这只能算是DEMO,还不足以实际应用
遇到的问题3 像上面的解决方案失去了事务性了，如果对事务没有要求的话，这样子问题也不大，而如果这段保存操作是作为WCF作务客户端调用呢，问题就有了，客户端提交一批数据上来，服务要保证其事务性，为此我们又要对程序改造成如下样子
if (_context.Connection != null)
_context.Connection.Open();
DbTransaction tran = _context.Connection.BeginTransaction();
try
{
foreach (NetPay_Order recorder in getOrders())
{
_context.NetPay_Orders.InsertOnSubmit(recorder);
_context.SubmitChanges();
recorder.MerSeqId = recorder.MerDate.ToString() + recorder.SequenceId.ToString();
_context.SubmitChanges();
}
}
catch (Exception ex)
{
tran.Rollback();
}
到这里已经解决了问题2，3了