Functional Programming in Kotlin: Lazy Evaluation

Functional Programming in Kotlin: Lazy Evaluation

Table of contents

Lazy evaluation is an optimization technique commonly used in functional programming. The basic idea is simple: don't evaluate a function or expression until you actually need its result. This can save time and resources, especially when working with large data sets or complex calculations.

In Kotlin, lazy evaluation is implemented through the use of the lazy function. This function takes a lambda expression and returns a Lazy<T> instance, which represents a value that hasn't yet been evaluated. When you access the value for the first time, the lambda expression is evaluated and the result is stored for future use.

Here's an example:

val lazyValue: Lazy<Int> = lazy {
    println("Initializing lazy value")

// The lambda expression hasn't been evaluated yet
println("Before accessing lazy value")

// The lambda expression is evaluated and the result is stored
println("After accessing lazy value")

When you run this code, you'll see the following output:

Before accessing lazy value
Initializing lazy value
After accessing lazy value

As you can see, the lambda expression is only evaluated when you access the value property for the first time. The second time you access it, the stored result is returned immediately.

Lazy evaluation can be especially useful when working with large collections or expensive calculations. For example, if you need to perform a complex calculation on a large data set, you might want to use lazy evaluation to avoid calculating all the intermediate results until you actually need them.

Lazy evaluation can also be used to implement "memoization", which is a technique for caching the results of expensive function calls. Here's an example:

fun fib(n: Int): Int {
    val memo = mutableMapOf<Int, Int>()

    fun fibInternal(k: Int): Int = memo.getOrPut(k) {
        if (k < 2) k else fibInternal(k - 1) + fibInternal(k - 2)

    return fibInternal(n)

This code defines a function fib that calculates the nth Fibonacci number. It uses a mutable map to store the results of previous calculations, and the getOrPut function to look up a value in the map or calculate it if it's not already present.

The fibInternal function is defined inside the fib function, so it has access to the memo map. It uses recursion to calculate the Fibonacci number, but instead of calculating each intermediate result from scratch, it first checks the memo map to see if the value has already been calculated. If it has, it returns the cached result. If not, it calculates the result recursively and stores it in the memo map for future use.


Another advantage of lazy evaluation is that it can be used to handle infinite data sets. For example, suppose we have a function that generates an infinite sequence of prime numbers. With lazy evaluation, we can create a Sequence object that only generates the next prime number when it is needed, rather than generating all of the prime numbers at once.

fun primes(): Sequence<Int> {
    return generateSequence(2) { current ->
        var next = current + 1
        while (!isPrime(next)) {

val first10Primes = primes().take(10).toList()

In this example, the primes function generates an infinite sequence of prime numbers. We can use the take function to get the first 10 prime numbers from the sequence, and the toList function to convert the sequence to a list.

Wrapping Up

Overall, lazy evaluation is a powerful tool for improving the efficiency and performance of your functional programs, while also allowing you to work with more complex and interesting data structures.