C# Random类的正确应用方法

 /// <summary>
  /// 错误的Random构建。
  /// </summary>
  public static void Bad_Random()
  {
    //正确做法应当将 Random构建防止循环外。
    //Random创建间隔时间极短的情况下，随机算法序列会基本一致，倒是随机性也是一致的
    //var r = new Random();
    for (int i = 0; i < 10; i++)
    {
      var r = new Random();
      var val = r.Next(1, 100);
      Console.WriteLine(val);
    }
  }

  /// <summary>
  /// 生成一个10位随机数
  /// 设定了一定的复杂性，保证单线程下随机数不重复
  /// </summary>
  /// <param name="random">Random.</param>
  /// <returns>随机数.</returns>
  private static string GenerateRandomStr(Random random)
  {
    string source = "ABCDEFGHIKLMNOPQRTUVWXYZabcdefghiklmnopqrtuvwxyz";
    int length = 10;
    var list = Enumerable.Repeat(source, length)
       .Select(s => s[random.Next(s.Length)]).ToArray();
    return new string(list);
  }
  /// <summary>
  /// 单线程基本可以保证唯一性
  /// </summary>
  public static void Good_Random_In_SingleThread()
  {
    //正确做法应当将 Random构建防止循环外。
    //Random创建间隔时间极短的情况下，随机算法序列会基本一致，倒是随机性也是一致的
    var r = new Random();
    ConcurrentBag<string> list = new ConcurrentBag<string>();
    for (int i = 0; i < 20000; i++)
    {
      var val = GenerateRandomStr(r);
      list.Add(val);
    }

    Console.WriteLine($"单线程下重复数据有：{20000 - list.Distinct().Count()}");
  }

  /// <summary>
  /// 多线程下的Random构建。
  /// Bad案例，Random非线程安全
  /// 多线程高并发情况下，会出现概率重复
  /// </summary>
  public static void Bad_Random_In_MultThreads()
  {
    var r = new Random(unchecked((int)DateTime.Now.Ticks));
    ConcurrentBag<string> list = new ConcurrentBag<string>();

    var t1 = Task.Run(() =>
    {
      for (int i = 0; i < 10000; i++)
      {
        var val = GenerateRandomStr(r);
        list.Add(val);
      }
    });

    var t2 = Task.Run(() =>
    {
      for (int i = 0; i < 10000; i++)
      {
        var val = GenerateRandomStr(r);
        list.Add(val);
      }
    });

    Task.WaitAll(t1, t2);

    Console.WriteLine($"线程1和线程2的重复数据有：{20000 - list.Distinct().Count()}");
  }

 //利用ThreadLocal 实现每个线程下Random独有
  //再通过seed原子性变更，保证每个Random的seed不同而生成的随机数列也不同
  private static int seed = 100;
  private static ThreadLocal<Random> threadLocal = new ThreadLocal<Random>(() => new Random(Interlocked.Increment(ref seed)));

  /// <summary>
  /// 多线程下的Random构建。
  /// </summary>
  public static void Good_Random_In_MultThreads()
  {
    ConcurrentBag<string> list = new ConcurrentBag<string>();

    var t1 = Task.Run(() =>
    {
      for (int i = 0; i < 10000; i++)
      {
        var val = GenerateRandomStr(threadLocal.Value);
        list.Add(val);
      }
    });

    var t2 = Task.Run(() =>
    {
      for (int i = 0; i < 10000; i++)
      {
        var val = GenerateRandomStr(threadLocal.Value);
        list.Add(val);
      }
    });

    Task.WaitAll(t1, t2);

    Console.WriteLine($"[ThreadLocal模式]线程1和线程2的重复数据有：{20000 - list.Distinct().Count()}");
  }