XStream序列化空值

本文介绍了XStream序列化空值的处理方法，对大家解决问题具有一定的参考价值，需要的朋友们下面随着小编来一起学习吧！

问题描述

假设我有

class Student
{
String name;
int    age;
String teacher;
}

然后：

public class App1
{
    public static void main(String[] args)
    {
        Student st = new Student();
        st.setName("toto");

        XStream xs = new XStream();

        xs.alias("student",Student.class);

        System.out.println(xs.toXML(st));
    }

}

给我：

<student>
  <name>toto</name>
  <age>0</age>
</student>

有没有办法处理空值？我的意思是：

Is there a way for dealing null values ? I mean :

<student>
  <name>toto</name>
  <age>0</age>
  <teacher></teacher>
</student>

如果我这样做可能

st.setTeacher("");

但如果老师为空则不会。

but not if teacher is null.

我尝试使用自定义转换器，但似乎空值不会发送到转换器。

I tried with a custom converter but it seems the null values are not sent to the converter.

推荐答案

我正在使用XStream 1.4.7，自定义字段名称的@XStreamAlias注释，自定义转换器的@XStreamConverter（表示日期和其他自定义bean）。但是，甚至没有调用空值的自定义转换器。
我的目标是序列化对象的所有字段，包括null字段，我不需要解组XML。

I'm using XStream 1.4.7, @XStreamAlias annotation for custom field names, @XStreamConverter for custom converters (to represent dates and other custom beans). However, a custom converter for null values was not even called.My goal was to serialize all the fields of the object, including the null ones, I didn't need to unmarshal XML.

我设法做到了通过创建自定义ReflectionConverter。
我从XStream库扩展了ReflectionConverter并覆盖了doMarshal方法。我唯一改变的是为null info.values调用writeField方法：

I managed to do that by creating a custom ReflectionConverter.I extended ReflectionConverter from XStream library and overrode doMarshal method. The only thing I changed was calling writeField method for null info.values:

new Object() {
{
    for (Iterator fieldIter = fields.iterator(); fieldIter.hasNext();) {
        FieldInfo info = (FieldInfo) fieldIter.next();
        if (info.value != null) {
            //leave the code unchanged
            ...
        } else {
            //add this to write null info.value to xml
            Log.info("MyCustomReflectionConverter -> serialize null field: " + info.fieldName);
            writeField(info.fieldName, null, info.type, info.definedIn, info.value);
        }
    }
    //... leave the rest of the code unchanged
}
};

之后，我创建了类似的xStream实例（将转换器注册到非常低的位置非常重要优先）：

After that, I created xStream instance like that (it is very important to register your converter with very low priority):

StaxDriver driver = new StaxDriver(new NoNameCoder()) {
    @Override
    public StaxWriter createStaxWriter(XMLStreamWriter out) throws XMLStreamException {
        // the boolean parameter controls the production of XML declaration
        return createStaxWriter(out, false);
    }
};

XStream xStream = new XStream(driver);
xStream.autodetectAnnotations(true);//needed to process aliases
//register MyCustomReflectionConverter
MyCustomReflectionConverter reflectionConverter = new MyCustomReflectionConverter (xStream.getMapper(), new SunUnsafeReflectionProvider());
xStream.registerConverter(reflectionConverter, XStream.PRIORITY_VERY_LOW);

感谢Mark Nabours的解决方案

Thanks to Mark Nabours for his solution here

希望它有所帮助。有人找到了更好的解决方案吗？

Hope it helps. Has anyone found a better solution for this?

这篇关于XStream序列化空值的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！

Xstream