Learn about serialization and deserialization in Java using the `Serializable` interface.

Serialization: Lesson

Great job! You’ve learned a lot about serialization. 

In this lesson we've discussed how:
- Java makes serialization easy using the `Serializable` interface.
- The `Serializable` interface is a _marker_ interface and requires no methods to be implemented.
- To use `FileOutputStream` and `ObjectOutputStream` to serialize objects to a file.
- To use `FileInputStream` and `ObjectInputStream` to deserialize objects from a file.
- `static` and `transient` fields are not serializable.
- The JVM handles serializing objects with reference type fields. 
- To implement `writeObject()` to customize the serialization of an object to handle non-serializable fields.
- To implement `readObject()` to customize the deserialization of an object with non-serializable fields. 

Knowing these topics will improve your mastery of Java and allow you to persist your objects past program termination or transport them over a network.


Review

Knowing how to serialize our objects is great, but it becomes more useful when we can turn serialized objects back into Java objects with deserialization. As the name suggests, deserialization does the opposite of serialization and transforms a stream of bytes into a Java object. 

Like with serialization, we’ll make use of helper classes `FileInputStream`, which helps us read a file, and `ObjectInputStream` which helps us read a serialized stream of bytes. 

Assuming we have the file **persons.txt** we created in the last exercise, let's understand how to do this by looking at some code:

```java
public class Person implements Serializable {
  public static void main(String[] args) throws FileNotFoundException, IOException, ClassNotFoundException {

    FileInputStream fileInputStream = new FileInputStream("persons.txt");
    ObjectInputStream objectInputStream = new ObjectInputStream(fileInputStream);
		
    Person michaelCopy = (Person) objectInputStream.readObject();
    Person peterCopy = (Person) objectInputStream.readObject();
  }
}
```
In the example above we:
- Initialized `FileInputStream` and `ObjectInputStream` in `main()` which will read a file and transform a stream of bytes into a Java object.
- Created a `Person` object named `michaelCopy` by using `objectInputStream.readObject()` to read a serialized object. 
- Created a `Person` object named `peterCopy` by using `objectInputStream.readObject()` to read a serialized object. 

It's important to note that deserialization creates a _copy_ of the original object. This means that the deserialized object shares the same field values as the original object but is located in a different place in memory. Also, The deserialized objects will be in the order that they were serialized in and we need to explicitly typecast the `readObject()` return value back into the type of object we serialized. 

The JVM ensures it deserializes the object using the correct class file by comparing the `serialVersionUID` in the class file with the one in the serialized object. If a match is not found an `InvalidClassException` is thrown. Also, `readObject()` throws a `ClassNotFoundException` when the `class` of the serialized object cannot be found.

Let's practice deserializing objects from a file. 

Deserializing an Object to a File

Now that we've learned about the `Serializable` interface and how to implement it, let's take a look at how to serialize an object to a file. To do this we'll need to use the helper classes, `FileOutputStream`, which will help us write to a file,  and `ObjectOutputStream`, which will help us write a serializable object to an output stream.  

Let's look at how to do this with some code:
```java
public class Person implements Serializable {
  private String name;
  private int age;
  private static final long serialVersionUID = 1L; 

  public Person(String name, int age) {
    this.name = name;
    this.age = age;
  }  

  public static void main(String[] args) throws FileNotFoundException, IOException{
    Person michael = new Person("Michael", 26);
    Person peter = new Person("Peter", 37);
		
    FileOutputStream fileOutputStream = new FileOutputStream("persons.txt");
    ObjectOutputStream objectOutputStream = new ObjectOutputStream(fileOutputStream);
		
    objectOutputStream.writeObject(michael);
    objectOutputStream.writeObject(peter);	
  }
} 
```
In the example above we:
- Added a constructor to the `Person` class we defined in the previous lesson. 
- Defined `main()` and initialized two `Person` objects - `michael` and `peter`.
- Initialized a `FileOutputStream` object in `main()` which will create and write a stream of bytes to a file named _persons.txt_.
- Initialized an `ObjectOutputStream` object in `main()` which will help serialize an object to a specified output stream. 
- Used `objectOutputStream.writeObject()` in `main()` to serialize the `michael` and `peter` objects to a file.

After the execution of the above program, the _persons.txt_ will contain a stream of bytes that has the type and value information of the fields for the `michael` and `peter` objects respectively.  

Let's practice serializing an object to a file.

Serializing Objects to a File

As we've been serializing objects, we've not only had primitive type fields, but also reference types. For example: 
```java
public class Person implements Serializable {
  private String name;
  private int age;
  private static final long serialVersionUID = 1L; 
```
In the example above, we've defined two primitive types, `age` and `serialVersionUID`, and a reference `String` type  `name`. 
For reference types to be serializable, they must also implement the `Serializable` interface, which the `String` class does. If the reference type did not implement `Serializable`, then we would get a `NotSerializableException` thrown. When the JVM encounters a reference type, it will try to serialize the reference type first before trying to serialize the encapsulating object. 

Let's look at class definitions and go through the steps of serializing a `Person` object:
```java
public class Address implements Serializable{
  private String streetName;
}

public class Person implements Serializable{
  private Address address;
}
```
In the example above, when serializing a `Person` object the JVM will:
- Check that the `Person` object is serializable and try to serialize the `Address` type.
- Check that the `Address` object is serializable and try to serialize the `String` type.
- Check that the `String` object is serializable and serialize the value. 
- Finish serializing the `Address` object after serializing `streetName`.
- Finish serializing the `Person` object after serializing `address`.

Our serializable classes will often have a combination of reference and primitives types, but the JVM will perform the same steps as above. If an exception were to be thrown at any point while serializing a reference type, a `NotSerializableException` would be thrown and propagated up. 

Let's practice serializing objects with reference-type fields.


Serializing Associated Fields

As we've learned about serialization, we've discussed how the JVM defines a default way to serialize objects when their classes implement the `Serializable` interface. Can we modify this default process? We can by implementing the methods `readObject()` and `writeObject()` in our class!

Let’s look at some code that implements `readObject()` and `writeObject()`:
```java
public class DateOfBirth {
  private int month;
  private int day;
  private int year;

  // constructors and getters
}

public class Person implements Serializable {
  private String name;
  private DateOfBirth dateOfBirth;

  private void writeObject(java.io.ObjectOutputStream stream) throws IOException{
    stream.writeObject(this.name);
    stream.writeInt(this.dateOfBirth.getMonth());
    stream.writeInt(this.dateOfBirth.getDay());
    stream.writeInt(this.dateOfBirth.getYear());
  }

  private void readObject(java.io.ObjectInputStream stream) throws IOException, ClassNotFoundException{
    this.name = (String) stream.readObject();
    int month = (int) stream.readInt();
    int day = (int) stream.readInt();
    int year = (int) stream.readInt();
    this.dateOfBirth = new DateOfBirth(month, day, year);
  }	
}
```
In the example above:
- We have two classes: `Person` which implements `Serializable` and `DateOfBirth` which does not.
- `Person` has a reference field of type `DateOfBirth`.
- If we were to use the default serialization process, we would get a `NotSerializableException` because `DateOfBirth` does not implement `Serializable`.
- We implement `writeObject()`, which must throw `IOException`, to serialize a `DateOfBirth` object by manually serializing all of its fields separately. We also serialize the serializable `String` field.
- We implement `readObject()`, which must throw `IOException` and `ClassNotFoundException`, to deserialize a `Person` object by reading the `int` fields that are a part of `DateOfBirth` and creating a new `DateOfBirth` object with the provided constructor. This new object is used to set the `dateOfBirth` field in `Person`.

Often, the default process of serialization is enough as long as all references implement `Serializable`. The implementation of `readObject()` and `writeObject()` is useful when we have a reference field that does not or cannot implement `Serializable`. We could also potentially handle `static` field values if we needed to persist them. This, however, is not good practice as a `static` field should belong to a class and not an object.

Let's practice implementing our custom serialization and deserialization methods.

Custom Serialization

Previously, we learned that to serialize an object, we needed our class to implement the `Serializable` interface. Notice that although we implemented the interface, we did not need to implement any methods. You may be thinking, "If we don't need to implement any methods, why do we need our class to implement `Serializable`?"

In Java, the JVM defines a default way for classes that implement `Serializable` to have their objects serialized. The interface provides run-time information about the object field's type and value for serialization and it takes care of converting it into a stream of bytes. 

Although there is no need to implement any methods for serialization, it is important for the implementing class to provide a `serialVersionUID`: 
```java
public class Person implements Serializable {
  private String name;
  private int age;
  private static final long serialVersionUID = 1L; 
} 
```
In the example above, the `serialVersionUID`, a `static final long` type number, acts as an identifier for the JVM to choose the proper class to convert a stream of bytes back into an object (we'll cover this process in-depth later). Our serializable class can get a `serialVersionUID` in one of the following ways:
- You can have the JVM define one for you. This is not ideal because, depending on the JVM you have, you'll get a different value and this may cause deserialization to fail.
- You can have your IDE generate one for you. This is better than the first option but you'll need to ensure that your IDE has this feature.
- You can define the `serialVersionUID` explicitly in the class. This is the preferred option because you don't need to rely on the JVM or your IDE to ensure proper deserialization.  

Although the JVM uses the `serialVersionUID` to locate the proper class, it does not store the class file as part of the serialization. We need to ensure we load the class file into our program (if it's not there already). With a default process of serializing objects and the `serialVersionUID`, Java makes serialization easy to implement. 

Let's practice adding a `serialVersionUID` to our serializable class.

Serializable Interface

When serializing objects, we often need to handle static class fields or exclude non-static fields in the serialization. Recall that the JVM defines a default way of serializing objects; this default includes ignoring static class fields, which belong to a class and not an object. The JVM also serializes all fields in an object, even those marked `private` and `final`.

Although the JVM implicitly serializes non-static fields, we can instruct the JVM to ignore them using the `transient` keyword. A field marked as `transient` will have its value ignored during serialization and instead receive the default value for that type of field. 

Let's visualize this with some code:
```java
public class Person implements Serializable {
  private String name;
  private int age;
  private static int numberOfPeople = 10;
  private transient int yearBorn;  
} 
```
In the example above, we defined a `static` field called `numberOfPeople` with a value of `10`. Since it belongs to the class, it will not be included in the serialization process. We also defined a `transient` field called `yearBorn` of type `int`. The JVM will ignore the initialized value of this field and instead serialize the field with its default value for its type (`0` in this case).

Let's discuss some interesting aspects of deserializing objects with `transient` and `static` fields:
- The deserialized (copy) object will not get the default value for a `static` field (in our example the value `0`). Since a `static` field belongs to the class and not the object, a deserialized `static` field will receive the value it corresponds to in the current class.
- A constructor is not called during deserialization for the deserialized type object. Object fields are set using _reflection_.
- A constructor is only called for the first non-serializable class in the parent hierarchy of the deserialized object. 

We often want to serialize all non-static fields in an object. However, sometimes we may need to serialize `transient` fields if a field's value is dependent on other fields or, more importantly, if we have a reference field that is not serializable (we'll talk about this later).

Let's practice using different types of serializable and non-serializable fields.

Serializable Fields

Often when creating Java applications and working with objects, we find the need to persist them. Specifically, we need to be able to store an object's state (member fields) in files, in memory, or in a database. Java provides the `Serializable` interface to do just that. 

Remember that an interface describes the behavior a class should have and by having our class implement `Serializable` it can be serialized by the Java Virtual Machine (JVM). 

Let's look at how this is done:
```java
public class Person implements Serializable {
  private String name;
  private int age;
} 
```
In the example above we've defined a `Person` class that implements `Serializable`. By implementing `Serializable`, the JVM will know how and what to do when a `Person` object needs to be serialized (we'll look at this more later).

Notice that we didn't need to override a method and implement it for us to be able to serialize a `Person` object. This is because `Serializable` is a _marker_ interface. A marker interface provides run time information to the JVM about the class and does not have any methods associated with it. 

Over the course of this lesson, we’ll continue learning about the `Serializable` interface and how serialization works in Java. 
