它也可以用于创建作用域，例如:

class LoggerScope:IDisposable {
   static ThreadLocal<LoggerScope> threadScope = 
        new ThreadLocal<LoggerScope>();
   private LoggerScope previous;

   public static LoggerScope Current=> threadScope.Value;

   public bool WithTime{get;}

   public LoggerScope(bool withTime){
       previous = threadScope.Value;
       threadScope.Value = this;
       WithTime=withTime;
   }

   public void Dispose(){
       threadScope.Value = previous;
   }
}


class Program {
   public static void Main(params string[] args){
       new Program().Run();
   }

   public void Run(){
      log("something happend!");
      using(new LoggerScope(false)){
          log("the quick brown fox jumps over the lazy dog!");
          using(new LoggerScope(true)){
              log("nested scope!");
          }
      }
   }

   void log(string message){
      if(LoggerScope.Current!=null){
          Console.WriteLine(message);
          if(LoggerScope.Current.WithTime){
             Console.WriteLine(DateTime.Now);
          }
      }
   }

}

Microsoft documentation states that using has a double function (https://msdn.microsoft.com/en-us/library/zhdeatwt.aspx), both as a directive and in statements. As a statement, as it was pointed out here in other answers, the keyword is basically syntactic sugar to determine a scope to dispose an IDisposable object. As a directive, it is routinely used to import namespaces and types. Also as a directive, you can create aliases for namespaces and types, as pointed out in the book "C# 5.0 In a Nutshell: The Definitive Guide" (http://www.amazon.com/5-0-Nutshell-The-Definitive-Reference-ebook/dp/B008E6I1K8), by Joseph and Ben Albahari. One example:

namespace HelloWorld
{
    using AppFunc = Func<IDictionary<DateTime, string>, List<string>>;
    public class Startup
    {
        public static AppFunc OrderEvents() 
        {
            AppFunc appFunc = (IDictionary<DateTime, string> events) =>
            {
                if ((events != null) && (events.Count > 0))
                {
                    List<string> result = events.OrderBy(ev => ev.Key)
                        .Select(ev => ev.Value)
                        .ToList();
                    return result;
                }
                throw new ArgumentException("Event dictionary is null or empty.");
            };
            return appFunc;
        }
    }
}

这是一种明智的做法，因为滥用这种做法会损害代码的清晰性。在DotNetPearls (http://www.dotnetperls.com/using-alias)中有一个关于c#别名的很好的解释，也提到了优点和缺点。

public class ClassA:IDisposable
{
    #region IDisposable Members
    public void Dispose()
    {
        GC.SuppressFinalize(this);
    }
    #endregion
}

public void fn_Data()
{
    using (ClassA ObjectName = new ClassA())
    {
        // Use objectName
    }
}

当你使用using时，它将调用using作用域末端对象上的Dispose()方法。因此，在Dispose()方法中可以有相当多出色的清理代码。

要点:

如果您实现了IDisposable，请确保在Dispose()实现中调用GC.SuppressFinalize()，否则自动垃圾收集将尝试出现并在某个时刻Finalize它，如果您已经Dispose()d了它，这至少会浪费资源。

对我来说，“using”这个名字有点令人困惑，因为它可以是一个导入Namespace的指令，也可以是一个用于错误处理的语句(就像这里讨论的那样)。

为错误处理取一个不同的名字会很好，而且可能是一个更明显的名字。

比如:

using (var conn = new SqlConnection("connection string"))
{
   conn.Open();

    // Execute SQL statement here on the connection you created
}

这个SqlConnection将被关闭，而不需要显式地调用. close()函数，并且即使抛出异常也会发生，而不需要try/catch/finally。