Newtonsoft 特性[JsonIgnore ] 对于继承属性的神奇效果

阅读前请有点基础

            [JsonIgnore]
            public DateTime CreateTimccc { get; set; }

一般用Newtonsoft 序列化类时候,如果不要序列化这个属性,在上面加这个特性就好了(ps.这个特性和Newtonsoft和Text.Json的名称重复,注意不要搞错)

定义子类和父类,用隐藏基类属性的方式

        class CommonRelationTest : CommonRelationba
        {
            public new  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            public  int RelationBusinessState { get; set; } = 1;
        }

测试代码

       Console.WriteLine(JsonConvert.SerializeObject(new CommonRelationTest()));
       Console.WriteLine(JsonConvert.SerializeObject(new CommonRelationba()));

结果如下

 

接下来,代码修改一下,加入[JsonIgnore]特性

        class CommonRelationTest : CommonRelationba
        {
            public new  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            [JsonIgnore]
            public int RelationBusinessState { get; set; } = 1;
        }

        class CommonRelationTest : CommonRelationba
        {
            [JsonIgnore]
            public new  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            public  int RelationBusinessState { get; set; } = 1;
        }

代码,很神奇吧,好像和预期的不一样,预期上面应该是输出 {} ,再改一下

        class CommonRelationTest : CommonRelationba
        {
            [JsonIgnore]
            public new  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            [JsonIgnore]
            public int RelationBusinessState { get; set; } = 1;
        }

这个原因是 JsonIgnore attribute on shadowed properties · Issue #463 · JamesNK/Newtonsoft.Json (github.com) 里面有解释

主要就是要区分new属性和原本,不然会降低灵活性,如果正好有人需要序列化基类的属性呢?(需求就是这样,我就不要评判)

如果要实现想要的需求,可以通过虚方法去实现

如果你虚方法用这种写法,那么结果是

        class CommonRelationTest : CommonRelationba
        {
            [JsonIgnore]
            public new int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            public virtual int RelationBusinessState { get; set; } = 1;
        }

   class CommonRelationTest : CommonRelationba
        {           
            public new int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            [JsonIgnore]
            public virtual int RelationBusinessState { get; set; } = 1;
        }

接下来写下真的虚方法

        class CommonRelationTest : CommonRelationba
        {
            public override  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            [JsonIgnore]
            public virtual int RelationBusinessState { get; set; } = 1;
        }

        class CommonRelationTest : CommonRelationba
        {
            [JsonIgnore]
            public override  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            public virtual int RelationBusinessState { get; set; } = 1;
        }

        class CommonRelationTest : CommonRelationba
        {
            [JsonIgnore]
            public override  int RelationBusinessState { get; set; } = 2;
        }
        class CommonRelationba
        {
            [JsonIgnore]
            public virtual int RelationBusinessState { get; set; } = 1;
        }

 注意对比下不同, 果然,代码很神奇哦,

我理解是对于new的属性newtonsoft会先解析本类的属性,如果这个new的属性正好被JsonIgnore,那么newtonsoft 回去找基类的属性(等于是有2个相同名称的属性)

而用虚方法重写,newtonsoft 就不会往基类寻找属性了(等于是只有1个属性)