Kotlinlearncs.online LogoJava

Lambda Expressions

interface Modify {
int modify(int value);
}
Modify first = (value) -> value + 1;
Modify second = (value) -> value - 10;
System.out.println(first.modify(10));
System.out.println(second.modify(3));

We saw how anonymous classes could encapsulate reusable logic and allow us to build, for example, a general purpose counting method. But their syntax still left a lot to be desired! In this lesson we’ll improve on that by introducing lambda expressions. Awesome! Let’s start.

As a reminder, this is an advanced topic. We’re introducing it because you will see lambda expressions in real Java code, including on the project that we start next! We won’t test you heavily on it, but a simple lambda expression question may appear on an upcoming quiz.

Warm Up
Warm Up

But let’s warm up with a classic practice problem on software testing! This is similar to the problem you’ll need to solve for this lesson’s homework.

Created By: learncs.online Staff
/ Version: 2020.10.0

Create a public class TestArraySum that provides a single void class method named test. test accepts a single parameter: an instance of ArraySum.

Each ArraySum provides a method sum that accepts an int[] and returns the sum of the values as an int, or 0 if the array is null. However, some ArraySum implementations are broken! Your job is to identify all of them correctly.

To do this you should use assert to test various inputs. Here's an example:

Your function does not need to return a value. Instead, if the code is correct no assertion should fail, and if it is incorrect one should.

As you design test inputs, here are two conflicting objectives to keep in mind:

  • Less is more. The fewer inputs you need to identify all broken cases, the more readable your test suite will be.
  • Think defensively. At the same time, you want to anticipate all of the different mistakes that a programmer might make. You've probably made many of these yourself! Examples include forgetting to check for null, off-by-one errors, not handling special cases properly, etc.

Good luck and have fun!

Functional Programming
Functional Programming

Java is an object-oriented programming language. With the exception of the primitive types, everything in Java is an object, and so Java programs involve manipulating objects.

However, other programming languages introduce different programming styles. One powerful and interesting style of programming is known as functional programming:

In computer science, functional programming is a programming paradigm where programs are constructed by applying and composing functions.

Let’s examine the Wikipedia definition together and contrast it with Java’s object-oriented style:

First-Class Functions
First-Class Functions

One characteristic of true functional programming languages is that functions (or methods) are first-class citizens. They can be stored in variables and passed to and returned from other functions, just like any other kind of data.

Java does not support first-class functions. Only object references and primitive type values can be stored in variables and passed to and returned from functions.

Getting to Lambda Expressions
Getting to Lambda Expressions

However, it turns out that we can achieve something very similar to first-class functions in Java! It looks like this:

interface Modify {
int modify(int value);
}
// first stores a reference to a function that returns its argument increased by one
Modify first = (value) -> value + 1;
System.out.println(first.modify(7));

We accomplish this by combining two things we already know—interfaces and anonymous classes—with some new Java syntax. Let’s see how, step by step.

But first, let’s state our goal.

Functional Interfaces
Functional Interfaces

Our first ingredient is called a functional interface. A functional interface is any old Java interface, but with one restriction: it can only provide one method. We’ll see why in a minute.

Other than that, there are no restrictions on what a functional interface looks like. Here’s one:

interface Modify {
int modify(int value);
}

Here’s another:

interface Filter {
boolean accept(int first, int second);
}

Anonymous Classes
Anonymous Classes

Next, we need a way to create something that implements a functional interface on the fly. But wait—we already know how to do that! It’s called an anonymous object:

interface Filter {
boolean accept(int first, int second);
}
Filter bothBositive = new Filter() {
@Override
public boolean accept(int first, int second) {
return first > 0 && second > 0;
}
};
Filter bothNegative = new Filter() {
@Override
public boolean accept(int first, int second) {
return first < 0 && second < 0;
}
};

Lambda Expressions
Lambda Expressions

We are so close now. Imagine that we want to save into a variable a method that increments an int by one. Here’s what it looks like given what we already know. First we need our functional interface, and then an anonymous class to implement it correctly:

interface Modify {
int modify(int value);
}
Modify first = new Modify() {
@Override
public int modify(int value) {
return value + 1;
}
};
System.out.println(first.modify(7));

But we can do better! Let’s see how:

interface Modify {
int modify(int value);
}
Modify first = new Modify() {
@Override
public int modify(int value) {
return value + 1;
}
};
System.out.println(first.modify(7));

Practice: Adder with Lambda

Created By: learncs.online Staff
/ Version: 2020.10.0

Declare a public class Modifier providing one static method adder. adder takes a single int parameter and returns a method that implements the Modify functional interface:

The returned "function" should implement modify so that it adds the value passed to adder. So, for example:

The correct solution to this problem is a single line lambda expression!

Array Counting with Lambdas
Array Counting with Lambdas

To finish up, let’s return to our example from last time that used anonymous classes to count arrays in different ways. We’ll reimplement it using lambdas and show how much cleaner and more direct this syntax is.

// Array Counting with Lambdas

Homework: Test Array Max

Created By: learncs.online Staff
/ Version: 2021.10.0

Create a public class TestArrayMax that provides a single void class method named test. test accepts a single parameter: an instance of ArrayMax.

Each ArrayMax provides a method max that accepts an int[] and returns the maximum of the values as an int. If the array is null or empty max should throw an IllegalArgumentException.

However, some ArrayMax implementations are broken! Your job is to identify all of them correctly. To do this you should use assert to test various inputs. (Do not throw other kinds of exceptions on failure.)

Here's an example:

Your function does not need to return a value. Instead, if the code is correct no assertion should fail, and if it is incorrect one should.

As you design test inputs, here are two conflicting objectives to keep in mind:

  • Less is more. The fewer inputs you need to identify all broken cases, the more readable your test suite will be.
  • Think defensively. At the same time, you want to anticipate all the different mistakes that a programmer might make. You've probably made many of these yourself! Examples include forgetting to check for null, off-by-one errors, not handling special cases properly, etc.

You'll also need to think about how to use try-catch to handle places where the code should throw an exception, how to ensure that it does, and how to make sure it throws the right type.

Good luck and have fun!

More Practice

Need more practice? Head over to the practice page.