DataContractJson反序列化报错：遇到意外字符'D'，求问题排查

错误信息 #

反序列化类型为Tutorial10.DataContractJsonExample.Customer的对象时出错，遇到意外字符'D'。

代码示例 #

internal class DataContractJsonSerialization
{
    public static void Run()
    {
        Customer customer = new Customer()
        {
            Name = "Sas",
            Age = 21,
            Email = "email@email.com",
            Weight = 75.4,
            IsAlive = true,
            Orders = new Order[] {
                new Order() {
                    Price = 400,
                    Name = "Cake"
                },
                new Order() {
                    Price = 20.5,
                    Name = "Apple"
                }
            }

        };
        WriteObject("DataContractJsonSerialization.json", customer);
        ReadObject("DataContractJsonSerialization.json");
    }
    public static void WriteObject(string fileName, Customer customer)
    {

        FileStream writer = new FileStream(fileName, FileMode.Create);
        DataContractJsonSerializer ser =
            new DataContractJsonSerializer(typeof(Customer));
        ser.WriteObject(writer, customer);
        writer.Close();

    }

    public static void ReadObject(string fileName)
    {
        var deserializedUser = new Customer();
        var ms = new MemoryStream(Encoding.UTF8.GetBytes(fileName));
        var ser = new DataContractJsonSerializer(deserializedUser.GetType());
        deserializedUser = ser.ReadObject(ms) as Customer;
        ms.Close();

    }
}

问题根源 #

问题出在ReadObject方法里：你把文件名字符串直接转换成字节流传给了MemoryStream，而不是读取文件中实际的JSON内容。文件名"DataContractJsonSerialization.json"的第一个字符是'D'，反序列化器误以为这是JSON数据的开头，自然会抛出"遇到意外字符'D'"的错误。

修复方案 #

修改ReadObject方法，读取文件的实际内容而非文件名：

方案一：使用FileStream直接读取文件

public static void ReadObject(string fileName)
{
    var deserializedUser = new Customer();
    // 用FileStream打开文件读取内容
    using (var fs = new FileStream(fileName, FileMode.Open))
    {
        var ser = new DataContractJsonSerializer(deserializedUser.GetType());
        deserializedUser = ser.ReadObject(fs) as Customer;
    }
}

方案二：读取文件字节数组到MemoryStream

public static void ReadObject(string fileName)
{
    var deserializedUser = new Customer();
    // 先读取文件的全部字节
    byte[] jsonContent = File.ReadAllBytes(fileName);
    using (var ms = new MemoryStream(jsonContent))
    {
        var ser = new DataContractJsonSerializer(deserializedUser.GetType());
        deserializedUser = ser.ReadObject(ms) as Customer;
    }
}

额外优化建议 #

建议使用using语句管理流资源，它会自动释放资源，避免手动调用Close()可能导致的资源泄漏，比如把序列化的WriteObject方法也改成更规范的写法：

public static void WriteObject(string fileName, Customer customer)
{
    using (FileStream writer = new FileStream(fileName, FileMode.Create))
    {
        DataContractJsonSerializer ser = new DataContractJsonSerializer(typeof(Customer));
        ser.WriteObject(writer, customer);
    }
}

内容的提问来源于stack exchange，提问作者NoOneCare