Vectorized conditional statement in R and Racket

In a previous post, I wrote that Racket’s if is similar to ifelse in R. That’s not quite right. This is a short post to clarify the comparison. The more accurate description is that this Racket code

(if (test-expr)
    true-expr
    false-expr)

is the same as this R code.

if (test_expr){
  true_expr
}else{
  false_expr
}

In contrast, R’s ifelse function is vectorized meaning that the same operation is applied to multiple elements of a vector. Below, we use ifelse to return a vector of the same length as the original vector with all negative values replaced by zero.

> a = c(-999, 2, -999, 4, 5, 6, 7, -999, 9, 10)
> ifelse(a < 0, 0, a)
 [1]  0  2  0  4  5  6  7  0  9 10

In Racket, we use map and an anonymous function, specified with lambda, to apply if to the elements of a list.

> (define a '(-999 2 -999 4 5 6 7 -999 9 10))
> (map (lambda (x) (if (< x 0) 0 x)) a)
'(0 2 0 4 5 6 7 0 9 10)

To apply if to a vector, we use vector-map.

> (define b #(-999 2 -999 4 5 6 7 -999 9 10))
> (vector-map (lambda (x) (if (< x 0) 0 x)) b)
'#(0 2 0 4 5 6 7 0 9 10)

It is also possible to write similar code in R using sapply. The curly braces can be omitted when if else is written inline.

> sapply(a, function(x) if (x < 0) 0 else x)
 [1]  0  2  0  4  5  6  7  0  9 10

R and Racket have a shared heritage in Scheme and Lisp that can yield some strinkingly similar code.

Avatar
Travis Hinkelman
Ecological Modeler