反序列化可能是整数的JSON或C#中的字符串列表

Question

我C#用来反序列化一组JSON字符串.每个字符串都有一个属性employer_normalized,该属性应该包含a,string并且List<int>其中的值List始终为正.在某些情况下employer_normalized被设置为-1,所以我想重写这些案件的一些行为,设置employer_normalized到null.

这是我的班级:

public class EmployerNormalized
{
  public string company;
  public List<int> code;
}

好JSON

"employer_normalized": {
        "company": "self",
        "code": [
          "4581 ",
          "6732 ",
          "9121",
          "9999 ",
          "5947 ",
          "8322 ",
          "8351 ",
          "7335 ",
          "9999 ",
          "4225 ",
          "8399 "
        ]
      }

坏JSON

"employer_normalized": -1

我目前正在使用Json.NET我的JSON解析.解决这个问题的优雅解决方案是什么？如果是这样的话,最好将employer_normalized值设置null为-1？如果是这样,我该怎么做？

Answer 1

您可以使用自定义的 JsonConverter 来处理这种情况。无论您在哪里期望 a EmployerNormalized，转换器都可以检查该属性的值是否为-1null 并返回 null，否则正常反序列化它。

这是转换器的代码：

public class EmployerNormalizedConverter : JsonConverter
{
    public override bool CanConvert(Type objectType)
    {
        return (objectType == typeof(EmployerNormalized));
    }

    public override object ReadJson(JsonReader reader, Type objectType, object existingValue, JsonSerializer serializer)
    {
        JToken token = JToken.Load(reader);
        if (token.Type == JTokenType.Object)
        {
            return token.ToObject<EmployerNormalized>();
        }
        return null;
    }

    public override void WriteJson(JsonWriter writer, object value, JsonSerializer serializer)
    {
        throw new NotImplementedException();
    }
}

这是一个演示如何使用它的演示：

class Program
{
    static void Main(string[] args)
    {
        string json = @"
        {
            ""good"" : {
                ""company"": ""self"",
                ""code"": [
                    ""4581 "",
                    ""6732 "",
                    ""9121"",
                    ""9999 "",
                    ""5947 "",
                    ""8322 "",
                    ""8351 "",
                    ""7335 "",
                    ""9999 "",
                    ""4225 "",
                    ""8399 ""
                ]
            },
            ""bad"" : -1
        }";

        Wrapper wrapper =
            JsonConvert.DeserializeObject<Wrapper>(json, 
                new EmployerNormalizedConverter());

        DumpEmployer("good", wrapper.good);
        DumpEmployer("bad", wrapper.bad);
    }

    private static void DumpEmployer(string prop, EmployerNormalized emp)
    {
        Console.WriteLine(prop);
        if (emp != null)
        {
            Console.WriteLine("  company: " + emp.company);
            Console.WriteLine("  codes: " + 
                string.Join(", ", emp.code.Select(c => c.ToString())));
        }
        else
            Console.WriteLine("  (null)");
    }

    public class Wrapper
    {
        public EmployerNormalized good { get; set; }
        public EmployerNormalized bad { get; set; }
    }

    public class EmployerNormalized
    {
        public string company;
        public List<int> code;
    }
}

这是输出：

good
  company: self
  codes: 4581, 6732, 9121, 9999, 5947, 8322, 8351, 7335, 9999, 4225, 8399
bad
  (null)

重要提示：您可能会想用 来装饰该类EmployerNormalized，[JsonConverter(typeof(EmployerNormalizedConverter))]但如果这样做，那么当前形式的转换器将最终递归调用自身，直到出现错误StackOverflowException。如果您需要/想要使用该属性，则ReadJson需要更改转换器中方法的代码，以便手动创建该类的实例EmployerNormalized并单独填充其所有属性，而不是调用token.ToObject<EmployerNormalized>(). 这是一个替代版本，ReadJson可以避免递归问题：

public override object ReadJson(JsonReader reader, Type objectType, object existingValue, JsonSerializer serializer)
{
    JToken token = JToken.Load(reader);
    if (token.Type == JTokenType.Object)
    {
        EmployerNormalized employer = new EmployerNormalized();
        employer.company = token["company"].ToString();
        employer.code = token["code"].ToObject<List<int>>();
        return employer;
    }
    return null;
}