一、前言

工厂方法模式，定义一个用于创建对象的接口，让子类决定实例化哪一个类。工厂方法使一个类的实例化延迟到其子类。

二、工厂方法模式结构图

三、实例代码

1.1运算类

    /// <summary>
    /// 运算类
    /// </summary>
    class Operation
    {
        private double _numberA = 0;
        private double _numberB = 0;

        public double NumberA
        {
            get { return _numberA; }
            set { _numberA = value; }
        }

        public double NumberB
        {
            get { return _numberB; }
            set { _numberB = value; }
        }

        /// <summary>
        /// 得到运算结果
        /// </summary>
        /// <returns></returns>
        public virtual double GetResult()
        {
            double result = 0;
            return result;
        }
    }

    /// <summary>
    /// 加法类
    /// </summary>
    class OperationAdd : Operation
    {
        public override double GetResult()
        {
            double result = 0;
            result = NumberA + NumberB;
            return result;
        }
    }

    /// <summary>
    /// 减法类
    /// </summary>
    class OperationSub : Operation
    {
        public override double GetResult()
        {
            double result = 0;
            result = NumberA - NumberB;
            return result;
        }
    }
    /// <summary>
    /// 乘法类
    /// </summary>
    class OperationMul : Operation
    {
        public override double GetResult()
        {
            double result = 0;
            result = NumberA * NumberB;
            return result;
        }
    }
    /// <summary>
    /// 除法类
    /// </summary>
    class OperationDiv : Operation
    {
        public override double GetResult()
        {
            double result = 0;
            if (NumberB == 0)
                throw new Exception("除数不能为0。");
            result = NumberA / NumberB;
            return result;
        }
    }

1.2工厂方法

    /// <summary>
    /// 工厂方法
    /// </summary>
    interface IFactory
    {
        Operation CreateOperation();
    }

    /// <summary>
    /// 专门负责生产“+”的工厂
    /// </summary>
    class AddFactory : IFactory
    {
        public Operation CreateOperation()
        {
            return new OperationAdd();
        }
    }

    /// <summary>
    /// 专门负责生产“-”的工厂
    /// </summary>
    class SubFactory : IFactory
    {
        public Operation CreateOperation()
        {
            return new OperationSub();
        }
    }

    /// <summary>
    /// 专门负责生产“*”的工厂
    /// </summary>
    class MulFactory : IFactory
    {
        public Operation CreateOperation()
        {
            return new OperationMul();
        }
    }

    /// <summary>
    /// 专门负责生产“/”的工厂
    /// </summary>
    class DivFactory : IFactory
    {
        public Operation CreateOperation()
        {
            return new OperationDiv();
        }
    }

1.3客户端调用

   class Program
    {
        static void Main(string[] args)
        {
            IFactory operFactory = new AddFactory();
            Operation oper = operFactory.CreateOperation();
            oper.NumberA = 1;
            oper.NumberB = 2;
            double result=oper.GetResult();

            Console.WriteLine(result);

            Console.Read();
        }
    }

四、简单工厂和工厂方法的比较

客户端需要判断，就用简单工厂，如在客户端经过下拉列表选择运算符号传入工厂创建函数的参数中。

客户端不需要判断，就用工厂方法，如客户端此时知道自己要创建哪个具体的类。

简单工厂模式最大的优点是在工厂中包含了必要的逻辑判断，根据客户端的条件动态实例化相关的类，对于客户端来说，去除了与具体产品的依赖。但是你增加一个“求M数的N次方”的方法，我们就必须在运算工厂类中参加一个case的判断，修改了原有的类，这个就相当于违背了开放-封闭的原则。所以用到了工厂方法就不用修改原来的类，直接为新功能作扩展。