Passing Functions

Part of the Logic & Flow section of Coddy's R journey. Lesson 28 of 64.

A function can take another function as an argument and call it. apply_twice() below works with any function of one argument:

apply_twice <- function(f, x) f(f(x))
print(apply_twice(\(x) x + 3, 10))
print(apply_twice(toupper, "ok"))

Output:

[1] 16
[1] "OK"

Built-in functions such as sum, max or toupper are values too, so pass them by name without parentheses. With parentheses they would be called instead:

summarize <- function(xs, how) how(xs)
print(summarize(c(4, 9, 2), max))
print(summarize(c(4, 9, 2), sum))

Output:

[1] 9
[1] 15

A function that takes a predicate, a function returning TRUE or FALSE, can count or select values by any rule:

count_if <- function(xs, keep) {
  n <- 0
  for (x in xs) if (keep(x)) n <- n + 1
  n
}
print(count_if(c(3, 8, 10, 5), \(x) x %% 2 == 0))
print(count_if(c(3, 8, 10, 5), \(x) x > 4))

Output:

[1] 2
[1] 3

The caller decides the behaviour and the function supplies the loop, so one function covers many cases:

transform_each <- function(xs, f) {
  out <- c()
  for (x in xs) out <- c(out, f(x))
  out
}
print(transform_each(1:4, \(x) x * 10))
print(transform_each(c("a", "b"), toupper))

Output:

[1] 10 20 30 40
[1] "A" "B"

Challenge

Easy

Complete apply_n(f, x, n). Call f on x, then on the result, n times in total, and return the final value. When n is 0, return x unchanged.

The supplied code defines three functions, double, inc and half. It reads the name of one of them, the start value and n, and prints the result of apply_n().

Try it yourself

apply_n <- function(f, x, n) {
  # Write your code here
  x
}

# Supplied input/output code: keep it as it is
input <- suppressWarnings(readLines(file("stdin")))
ops <- list(double = \(x) x * 2, inc = \(x) x + 1, half = \(x) x / 2)
cat(apply_n(ops[[input[1]]], as.numeric(input[2]), as.numeric(input[3])), sep = "\n")