Java Basics

In lessons 5 and 6 of this series you've learned how to use some of the Java streams to read or write bytes, characters or numeric data. This lesson is about reading or writing entire Java objects into streams.

Let's say your application uses a class that looks like this:

class Employee {
String lName;
String fName;
double salary;
java.util.Date hireDate;
String address;
}

Now consider the following scenario: a program HeadQuarterEmpProcessor creates an instance of the object Employee. The values of its attributes (object's state) have to be saved in a file or some other stream. Later on, another program called BranchEmpProcessor needs to recreate the instance of this object Employee in memory.

We could have done it by using one of the streams like DataOutputStream, FileWriter or others. In this case both programs would need to know a format of the saved file (data types, order of the attributes and delimiters). Luckily, Java offers a more elegant way called object serialization, which greatly simplifies the process of objects exchange.

To send the entire object to a stream a program can use the class java.io.ObjectOutputStream, while the class java.io.ObjectInputStream knows how to get an object from a stream. To serialize an object means to convert it into a set of bytes and send it to a stream. To deserialize and object means to read these bytes from a stream and recreate the instance of the received object.

How to Make a Class Serializable

To make a class serializable, just declare that this class implements the interface Serializable:

class Employee implements java.io.Serializable {
String lName;
String fName;
double salary;
java.util.Date hireDate;
String address;
}

The good news is that Serializable interface does not force you to implement any methods, that's why modification of the class Employee was minimal.

All attributes of the class Employee must have either primitive data types, or represent objects that are also serializable.

How to Serialize an Object

To serialize an object into a stream perform the following actions:

• Open one of the output streams, for example FileOutputStream
• Chain it with the ObjectOutputStream
• Call the method writeObject() providing the instance of a Serializable object as an argument.
• Close the streams

The following example performs all these steps and creates a snapshot of the object Employee in the file called NewEmployee.ser

import java.io.*;
import java.util.Date;

public class HeadQuarterEmpProcessor {

public static void main(String[] args) {
Employee emp = new Employee();
emp.lName = "John";
emp.fName = "Smith";
emp.salary = 50000;
emp.address = "12 main street";
emp.hireDate = new Date();

FileOutputStream fOut=null;
ObjectOutputStream oOut=null;

try{
fOut= new FileOutputStream("c:\\NewEmployee.ser");
oOut = new ObjectOutputStream(fOut);
oOut.writeObject(emp); //serializing employee
System.out.println(
"An employee is serialized into c:\\NewEmployee.ser");
}catch(IOException e){
e.printStackTrace();
}finally{
try {
oOut.flush();
oOut.close();
fOut.close();
} catch (IOException e1) {
e1.printStackTrace();
}
}
}
}

If you do not want to serialize sensitive information such as salary, declare this variable using the keyword transient:

transient double salary;

The values of static and transient member variables are not serialized.

How to Deserialize an Object

To deserialize an object perform the following steps:

• Open an input stream
• Chain it with the ObjectInputStream
• Call the method readObject() and cast the returned object to the class that is being deserialized.
• Close the streams

The next example reads our file NewEmployee.ser and recreates the instance of the object Employee:

import java.io.*;

public class BranchEmpProcessor {

public static void main(String[] args) {
FileInputStream fIn=null;
ObjectInputStream oIn=null;

try{
fIn= new FileInputStream("c:\\NewEmployee.ser");
oIn = new ObjectInputStream(fIn);

//de-serializing employee
Employee emp = (Employee) oIn.readObject();

System.out.println("Deserialized " + emp.fName + " "
+ emp.lName + " from NewEmployee.ser ");
}catch(IOException e){
e.printStackTrace();
}catch(ClassNotFoundException e){
e.printStackTrace();
}finally{
try {
oIn.close();
fIn.close();
} catch (IOException e1) {
e1.printStackTrace();
}
}
}
}

The class BranchEmpProcessor will produce the following output:

Deserialized Smith John from NewEmployee.ser

Please note that we did not explicitly created an instance of the object Employee – JVM did it for us. Make sure that definition of the class Employee is available to JVM that reads the stream. In distributed applications it usually runs on a remote machine.

During the process of deserialization all transient variables will be initialized with default values according to their type, for example integer variables will have the value of zero.

Interface Externalizable

The method writeObject() sends all attributes of an object into a stream. This could lead to unnecessary large object footprint, especially if you need to serialize the values only of some of the instance variables. Java provides Externalizable interface that gives you more control over what is being serialized and it can produce smaller object footprint. Externalizable interface is a subclass of Serializable.

This interface defines 2 methods: readExternal() and writeExternal() and you have to implement these methods in the class that will be serialized (Employee). In these methods you'll have to write code that reads/writes only the values of the attributes you are interested in. Programs that perform serialization and deserialization have to write and read these attributes in the same sequence.

The following class Employee2 serializes only the values of the last name and salary.

import java.io.ObjectOutput;
import java.io.ObjectInput;
class Employee2 implements Externalizable {
String lName;
String fName;
double salary;
java.util.Date hireDate;
String address;

public void writeExternal(ObjectOutput stream)
throws java.io.IOException {
// Serializing only salary and last name
stream.writeDouble(salary);
stream.writeUTF(lName); // String encoded in UTF-8 format
}

public void readExternal(ObjectInput stream)
throws java.io.IOException {
salary = stream.readDouble();
lName = stream.readUTF();
}

}

The class HeadQuaterEmpProcessor2 shows how to externalize the object Employee2:

import java.io.*;
import java.util.Date;

public class HeadQuarterEmpProcessor2 {

public static void main(String[] args) {
Employee2 emp = new Employee2();
emp.fName = "John";
emp.lName = "Smith";
emp.salary = 50000;
emp.address = "12 main street";
emp.hireDate = new Date();

FileOutputStream fOut=null;
ObjectOutputStream oOut=null;

try{
fOut= new FileOutputStream("c:\\NewEmployee2.ser");
oOut = new ObjectOutputStream(fOut);
emp.writeExternal(oOut); //serializing employee
System.out.println(
"An employee is serialized into c:\\NewEmployee2.ser");

}catch(IOException e){
e.printStackTrace();
}finally{
try {
oOut.flush();
oOut.close();
fOut.close();
} catch (IOException e1) {
e1.printStackTrace();
}
}
}
}

Unlike with Serializable interface, we had to write a little more code to implement Externalizable interface, but the size of the file NewEmployee2.ser is only 21 bytes, whereas the file NewEmployee.ser has 207 bytes. First of all, we serialized the values of only two attributes, and the other reason is that files created using Externalizable interface contain data only, while files created by default Java serialization contain class metadata that include attribute names.

The next code snippet shows you how to recreate an externalized object:

fIn= new FileInputStream("c:\\NewEmployee2.ser");
oIn = new ObjectInputStream(fIn);

Employee2 emp = new Employee2();
emp.readExternal(oIn);

Serializaion in the Real World

In some types of applications you have to write the code to serialize objects, but in many cases serialization is performed behind the scenes by various server-side containers. These are some of the typical uses of serialization:

• To persist data for future use.
• To send data to a remote computer using such client/server Java technologies as RMI or socket programming.
• To "flatten" an object into array of bytes in memory.
• To exchange data between applets and servlets.
• To store user session in Web applications.
• To activate/passivate enterprise java beans.
• To send objects between the servers in a cluster.

When you use serialization in time-critical applications, for example real- time stock trading systems, the size of the serialized objects should be minimal. Keep in mind that variables with longer names produce larger footprints during serialization, and this may substantially slow down your application. Think of a high volume of trade orders that is being serialized. I remember working on the application where a class TradeOrder had about a hundred member variables. After renaming the variables into meaningless v1, v2, and so on, the size of one TradeOrder instance was reduced by a thousand bytes. And we are talking about serializing of thousands orders over the network!

If performance is your primary goal, use Externalizable interface instead of Serializable. Yes, you'll have to write code to serialize each attribute, but this may speed up serialization process substantially.

While applets can connect to a remote computer using socket or RMI programming (these technologies will be explained in the future lessons of this series), HTTP protocol and such Java classes as URL and URLConnection simplify network programming. With an HTTP protocol, applets can receive or send not only a text, but also binary objects using Java Serialization.

When an EJB container decides to passivate (unload from memory) so-called stateful session bean, JVM persists its state in a safe place (usually on a disk). Later on, when this bean will be activated again, all its variables will be automatically deserialized by the EJB container.

It's not too difficult for JVM to convert a primitive integer variable into four bytes for serialization, but it's not as simple in case of classes containing variables with references to other objects. The process of converting such complex object into a sequence of bytes is called marshalling and the process of reconstructing of the objects from these bytes is called unmarshalling and Java does this job for you.

Even though we have not learned yet how to create Web applications, I still want to mention that objects used for tracking of the user sessions should be serializable, otherwise you may not be able to deploy these application in a cluster of servers.

Java serialization is a simple but powerful feature of the language, and you'll definitely will have a chance to use it in your applications.