Entity Framework 6 Recipes 2nd Edition（12-3）译 -> 数据库连接日志

12-3. 数据库连接日志

问题

你想为每次与数据库的连接和断开记录日志

解决方案

EF为DbContext的连接公开了一个StateChange 事件.我们需要处理这个事件, 为每次与数据库的连接和断开记录日志.

假设我们的模型如Figure 12-3所示. 在 Listing 12-3代码里, 我们创建一些Donation 实例,然后把它们保存到数据库. 这里的代码实现override SaveChanges() 方法为我们的StateChange 事件提供切入点.

Figure 12-3. The model with the Donation entity

 

Listing 12-3. Code to Implement Logging of Open and Close of a Database Connection

class Program

{

static void Main(string[] args)

{

RunExample();

}

static void RunExample()

{

using (var context = new EFRecipesEntities())

{

context.Donations.Add(new Donation

{

DonorName = "Robert Byrd",

Amount = 350M

});

context.Donations.Add(new Donation

{

DonorName = "Nancy McVoid",

Amount = 250M

});

context.Donations.Add(new Donation

{

DonorName = "Kim Kerns",

Amount = 750M

});

Console.WriteLine("About to SaveChanges()");

context.SaveChanges();

}

using (var context = new EFRecipesEntities())

{

var list = context.Donations.Where(o => o.Amount > 300M);

Console.WriteLine("Donations over $300");

foreach (var donor in list)

{

Console.WriteLine("{0} gave {1}", donor.DonorName,

donor.Amount.ToString("C"));

}

}

Console.WriteLine("Press any key to close...");

Console.ReadLine();

}

}

public partial class EFRecipesEntities

{

public override int SaveChanges()

{

this.Database.Connection.StateChange += (s, e) =>

{

var conn = (DbConnection)s;

Console.WriteLine("{0}: Database: {1}, State: {2}, was: {3}",

DateTime.Now.ToShortTimeString(), conn.Database,

e.CurrentState, e.OriginalState);

};

return base.SaveChanges();

}

}

上述Listing 12-3代码输出如下:

About to SaveChanges()

09:56 : Database: EFRecipes, State: Open, was: Closed

09:56: Database: EFRecipes, State: Closed, was: Open

Donations over $300

Robert Byrd gave $350.00

Kim Kerns gave $750.00

原理

我们实现override SaveChanges() 方法,为我们的StateChange事件提供切入点,该事件处理程序接收两个参数:事件的发送者(sender)和StateChangeEventArgs.第二个参数提供了访问连接的当前状态和原始状态, 我们为这两个与数据库相关的状态做日志记录.如果你特点注意日志内容的顺序,你会注意到在第二个using块里,与数据库的连接在foreach执行要查询时重新启动,而不是在写查询语句时. 这个演示说明了一个重要的概念:查询只在必要时才进行.在我们的例子中,查询在遍历期间才执行

附：创建示例用到的数据库的脚本文件