Map, filter, fold and sequences

A teacher wants corrected scores, passing scores and a total. Map transforms every element into an output element. Filter keeps elements satisfying a condition. Reduce combines elements into a single accumulated result. These names describe distinct operations rather than interchangeable shortcuts.

Map transforms every element, filter keeps matching elements and fold combines values from a supplied initial accumulator. The lambda parameter it names a single argument when not explicitly named. AsSequence can defer intermediate operations until a terminal operation requests results.

Connect this to lambdas and higher-order functions and read-only and mutable lists.

Save the complete source as Main.kt and use the Kotlin/JVM environment. No separate input file is needed.

Kotlin/JVM · complete program

fun main() {
  val scores = listOf(60, 75, 90)
  println(scores.map { it + 5 })
  println(scores.filter { it >= 75 })
  println(scores.fold(0) { total, score -> total + score })
}

Expected output

[65, 80, 95]
[75, 90]
225

Follow the execution

The corrected list is [65, 80, 95], the passing original scores are [75, 90] and the folded total is 225. Each operation uses scores independently; creating one transformed list does not overwrite the original.

Try this in the Kotlin compiler by pasting the complete program.

Your task: Why does the passing list omit 60 despite creating corrected scores?

Kotlin/JVM · complete program

fun main() {
  val scores = listOf(60, 75, 90)
  println(scores.map { it + 5 })
  println(scores.filter { it >= 75 })
  println(scores.fold(0) { total, score -> total + score })
}

Worked answer

Filter reads the original scores, not the separate map result. Its condition applies directly to each original value.

Reduce has no supplied initial accumulator and needs handling for an empty collection; fold can return its initial value for empty input. Lazy sequences can avoid intermediate lists, but overhead and evaluation order mean they are not automatically faster for every task.

Filter reads the original scores, not the separate map result. Its condition applies directly to each original value. Reduce has no supplied initial accumulator and needs handling for an empty collection; fold can return its initial value for empty input.

Further reading: official Kotlin documentation for this topic.