CODEWITHSUNDEEP
javaserializationdeserialization
peter
peter

Reputation: 8473

Java : when do you want to customize the serialization process?

Under what circumstances and for what reasons should one consider customizing the serialization process? (The following methods are the methods that will be invoked by the JVM)

 private void writeObject(java.io.ObjectOutputStream out){
         //customize serialization
 }
 private void readObject(java.io.ObjectInputStream in){
         //customize de-serialization
 }

Upvotes: 2

Views: 225

Answers (6)

Pshemo
Pshemo

Reputation: 124295

One of advantages of customized serialization is possibility to add transient fields (like inherited from other classes) into serialization process. You can do it like this

class NotMyClass implements Serializable {
    String login;
    transient String password;
    //...    
}

//I want to serialize also transient password field
class MyClass extends NotMyClass {

    private void writeObject(ObjectOutputStream out) throws IOException {
        System.out.println("defaul serialization");
        out.defaultWriteObject();

        System.out.println("my serialization >>>> adding inherited transient value");
        out.writeObject(password);
    }

    private void readObject(ObjectInputStream in) throws IOException,
            ClassNotFoundException {
        System.out.println("defaul deserialization");
        in.defaultReadObject();
        System.out.println("my deserialization <<<< reading stored transient value");
        password = (String) in.readObject();
    }
}

Upvotes: 1

Dilum Ranatunga
Dilum Ranatunga

Reputation: 13374

One other great reason is when you want to reduce the amount of data getting serialized. Take for example a HashMap with a capacity of say 1024. Let's say it has just 3 entries so far. Serializing an Object array of 1024 would be very wasteful, when all it needs to do is serialize the three entries.

On a separate note, if the objects you are serializing have references to non-serializable services, you would first mark those members as transient (to prevent serialization), but then use the writeObject to capture enough information so that readObject can re-establish the references upon de-serialization.

Upvotes: 0

Juan Alberto L&#243;pez Cavallotti
Juan Alberto L&#243;pez Cavallotti

Reputation: 4704

I'm thinking some cases where you may want to control the serialization:

  • When you want to encrypt and later decrypt the serialized data.
  • You also may want the users or some other app to modify the serialized form of your objects so in that case you may want to write something more human friendly.
  • If you're transmitting data over a network you may want to adapt the data so the program on the other side may understand it.

Those are the reasons I can think of right now.

Hope it helps.

Upvotes: 2

Amir Afghani
Amir Afghani

Reputation: 38571

Read Advanced Serialization from SUN

  1. One common reason to override readObject and writeObject is to serialize the data for a superclass that is not Serializable itself.
  2. Validate Streams
  3. Encrypt serialized objects

Upvotes: 2

mahercbeaucoup
mahercbeaucoup

Reputation: 597

You might want to customize what you want to do at the time of a read or write-- you can add extra logging, you can perform computations, you might only want to read or write if certain conditions are met, etc. Really, you can do anything Java allows. For an example, see the section "Customize the Default Protocol" at http://java.sun.com/developer/technicalArticles/Programming/serialization/

Upvotes: 0

peter
peter

Reputation: 8473

The only reason I can think of is when you want your sub-class to void serialization while the super-class is already serializable, and then you can just simply throw exception there in both methods in your sub-class.

I am not sure if there's any other specific reason or not.

Upvotes: 0

Related Questions