使用杰克逊序列化/反序列化具有不同属性的Child属性

Question

I have to POJO class( A and B)

A{
   a_id;
   a_name;
    @JsonIdentityInfo(generator=ObjectIdGenerators.PropertyGenerator.class, 
     property="b_name")
   @JsonIdentityReference(alwaysAsId=true)
   B b;
}

B{
    b_id;
    b_name;
 }

To serialize object A to json I want b_name and while deserialize object AI want b_id;

In short, I want to pass b_name of child field with the parent object. While deserialize I will get b_id so it should bind to B object(b_id). so same child object should work with different property while serialize and deserialize.

from server= {a_id=1, a_name="abc", b="pqr"}
from client ={a_id=1,a_name="abc",b=1}
b is b_name at serialization and b_id at deserialization.

Is it possible?

Answer 1

Yes that should be possible. While serialization, if your b_id is going to be null, and you don't want it to be included in json, you can use :

@JsonInclude(Include.NON_NULL)
class B {
}

While deserializing, as you are not going to get b_name, you can have @JsonIgnoreProperties(ignoreUnknown = true) at class B level. It tells jackson to ignore any JSON properties that cannot be mapped to an existing java field.

If your requirement for inclusion/exclusion of fields is more complex then this, then you can use JsonSerializer and JsonDeserializer . Something like:

public class MySerializer extends JsonSerializer<A> {
    @Override
    public void serialize(A a, JsonGenerator jGen, SerializerProvider arg2)
            throws IOException, JsonProcessingException {
        jGen.writeStartObject();
        jGen.writeStringField("a_id", a.getA_id());
        jGen.writeStringField("b_name", a.getB().getB_name() );
        jGen.writeEndObject();
    }
}