Functions, control flow, and the apply family — R Module 5

R has the standard if/else and for/while constructs. But idiomatic R prefers vectorised operations and the apply family — explicit loops are usually a sign you're thinking in the wrong language.

if / else

rate <- 0.115
if (rate > 0.13) {
    stance <- "tight"
} else if (rate > 0.10) {
    stance <- "neutral"
} else {
    stance <- "loose"
}

# Vectorised: ifelse
rates <- c(0.08, 0.115, 0.14)
stance <- ifelse(rates > 0.10, "tight", "loose")
# c("loose", "tight", "tight")

for loop — usually unnecessary

for (i in 1:5) {
    print(i)
}

# But you almost never need this in R
rates <- c(0.07, 0.10, 0.12)

# Loop version (don't do this)
total <- 0
for (r in rates) {
    total <- total + r
}

# R way: vectorised
sum(rates)

Functions

compound <- function(principal, rate, years = 10) {
    principal * (1 + rate) ^ years
}

compound(1000, 0.10, 5)        # positional
compound(1000, rate = 0.10)    # default years = 10

The apply family

Instead of writing loops, R has a family of higher-order functions: sapply, lapply, mapply, apply. They take a function and apply it across a vector, list, or matrix. purrr (tidyverse) provides map, map_dbl, map_chr — same idea, more consistent.

rates <- list(c(0.07, 0.10, 0.12), c(0.08, 0.11), c(0.09))

sapply(rates, mean)   # 0.0967 0.0950 0.0900 — applies mean to each list element

# purrr version
library(purrr)
map_dbl(rates, mean)

When to vectorise

If you're writing a for loop in R that just builds a vector, you're probably doing it wrong. Either the operation is already vectorised (use it directly), or use sapply/map_dbl. Loops are reserved for genuinely sequential operations.

Exercise

Use sapply() to compute the length of each element in a list of c(c(1,2,3), c(4,5), c(6)).