如下面的例子，很容易看出问题：

如果类本身的无参构造函数， 就添加了一些元素，那么在序列化，再反序列化，会导致元素增加。

如果要避免，必须添加：

new JsonSerializerSettings() { ObjectCreationHandling = ObjectCreationHandling.Replace }

using Newtonsoft.Json;
using System;
using System.Collections.Generic;

class Program3
{
    static void Main(string[] args)
    {
        Product product = new Product();
        Console.WriteLine("Orginal Count:{0}", product.ListCustomer.Count);
        string json = JsonConvert.SerializeObject(product);  //序列化
        Product result1 = JsonConvert.DeserializeObject<Product>(json);
        Console.WriteLine("Result Count:{0}", result1.ListCustomer.Count);
        Product result2 = JsonConvert.DeserializeObject<Product>(json,
                    new JsonSerializerSettings() { ObjectCreationHandling = ObjectCreationHandling.Replace });
        Console.WriteLine("Result2 Count:{0}", result2.ListCustomer.Count);
        Console.WriteLine("--------------------------------");
        P2 p = new P2();
        p.ListCustomer = new List<int>();
        p.ListCustomer.AddRange(new int[]{ 1,2,3,4 });
        Console.WriteLine("Orginal Count:{0}", p.ListCustomer.Count);
        string json2 = JsonConvert.SerializeObject(p);  //序列化
        P2 p2Result = JsonConvert.DeserializeObject<P2>(json2);
        Console.WriteLine("Result Count:{0}", p2Result.ListCustomer.Count);

        Console.WriteLine("END");
        Console.Read();
    }

    public class Product
    {
        public Product()
        {
            ListCustomer = new List<int> { 1, 2, 3, 4 };
        }
        public List<int> ListCustomer { get; set; }
    }

    public class P2
    {
        public List<int> ListCustomer { get; set; }
    }
}