Java Keywords Addendum: The Java Record

Since originally my Java series was based on Java 8, it did not include a new keyword introduced later on. For that reason, I decided to post an addition to the Java Keyword series to include the Java record keyword. Introduced in Java 14, the purpose of this keyword is to eliminate all the boilerplate code when creating a Java POJO. For example,

public class Student {
  private String name;
  private int id;
  
  public Student(String name, int name) {
    this.name = name;
    this.id = id;
  }
  
  public String getName() {
    return name;
  }
  
  public int getId() {
    return id;
  }
  
  public void setName(String name) {
    this.name = name;
  }
  
  public void setId(int id) {
    this.id = id;
  }
}
can be replaced simply with a record that looks like this:

public record Student(String name, int id){ }
And not only it replaces the boilerplate code I showed you, it also automatically overrides Object#equals(Object), Object#hashCode(), and Object#toString().

Usage

Here is a simpe example of a record declaration as a .java file.

Student.java


public record Student(String name, int id){ }
Main.java

public class Main {
  public static void main(String[] args) {
    Student s1 = new Student("John", 1);
    Student s2 = new Student("Mary", 2);
    Student s3 = new Student("Hector", 3);
    Student s4 = s2;
    
    System.out.println("Student 1's name: " + s1.name());
    System.out.println("Student 2's id: " + s2.id());
    System.out.println("Student 3: " + s3);
    System.out.println("Student 2 is equal to 4: " + s2.equals(s4));
  }
}
The first thing you would notice is that, using a record object is basically the same as using a class object. A Java record is a class. It is a class without boilerplate code. A much more simplified form of a class. It does have caveats, which are outlined in the "Restrictions" section below.

Another thing you should notice is that the getter methods do not have the prefix "get" in them. It is a direct derivation of the variable name. Is this a good thing or a bad thing? I am somewhat neutral, but leaning towards this convention. I am an old developer who is quite used to the getter/setter naming convention. But even I know this can get somewhat confusing when it comes to boolean fields. It seems to me that the developer community is almost split in the middle when it comes to the prepending boolean getter methods with "is" versus "get". To me, not prepending seems more sensible. Simpler. If you want to know the getter method name for a particular field, simply look for that field in the constructor.

The output of this example is as shown below:


Student 1's name: John
Student 2's id: 2
Student 3: Student[name=Hector, id=3]
Student 2 is equal to 4: true
The first two print out statements illustrate the use of the getter methods for the two fields in the class. The third print out illustrates the use of the default overridden Object#toString() method. The fourth print out statement illustrate the use of the implementation of Object#equals(Object) and Object#hashCode(). All of that packaged in this simple record declaration

public record Student(String name, int id){ }
This looks trivial when you look at a simple POJO with a two fields. But just think for a second how much time (compounded) it saves you over the life of a project. Not long ago, I was workinng alone in a project where I needed to create, initially, about 4 to 5 POJOs. It would have taken me maybe close to an hour to write all these by hand. Even with IDEs like IntelliJ or Eclipse, it would have taken me many minutes to create the classes and declared the fields, then go to the IDE menu to create getters and setters, override Object#equals(Object) and Object#hashCode(), and then override Object#toString(). It took me seconds to create the records.

Restrictions

There are some caveats that you must be aware when using records.
  1. Java records cannot be extended. They are implicitly final. For this reason, they cannot be abstract.
  2. Java records are immutable. Therefore, their inner values cannot be changed after instantiation.
  3. Because records are immutable, they contain no setter methods. Values must be passed through the constructor
For me, these constraints are more of a benefit than a constraint. Consider the immutable nature of a record. This means that they are inherently thread-safe. This benefit makes it easier to share this object across mulitple threads without worrying about changing internal values, proper ways to synchronize, slowing down the application when applying locks, etc. However, with every benefit, there are some drawbacks. For example, if the data being held by the object changes frequently, this means you will need to create a new record each and every time the data changes which can present challenges when garbage collecting. As a developer, it is up to you to decide what is the best approach for your application. In some cases, using records will greatly improve your application. In other cases, it might be of a hinderance to use them. Arm yourself with what you learn here and decide for yourself.

I hope this short article was of some benefit to you. For my next topic, I will get into customizing Java records and implementing interfaces with records. Good luck and stay tuned!

Comments

Popular posts from this blog

Combining State and Singleton Patterns to Create a State-Machine

The Beauty of the Null Object Pattern

Exception Handling: File CRUD Operations Example