Does R allows operators to be compound expressions?

Question

In Harold Abelson's "Structure and Interpretation of Computer Programs", which uses the Lisp programming language, one of the exercises asks you to evaluate the following code:

(define (a-plus-abs-b a b)
((if (> b 0) + -) a b))

One way to do this in R is of course:

APlusAbsB <- function(a, b){
  if(b > 0) a + b else a - b
}

Then I am repeating the as and bs.

My question is: does R allow operators to be compound expressions (e.g., if (> b 0) + -))? For example, is there a way to do something like

APlusAbsB <- function(a, b){  ## this doesn't work
  if(b > 0) "+" else "-" (a, b)    ## gives "+" or "-"
}

Answer 1

Try

APlusAbsB <- function(a, b){
  (if(b > 0) `+` else `-`)(a, b)
}

Here the if will return a function and then you call it with the parenthesis syntax. For example

APlusAbsB(1,5)
# [1] 6
APlusAbsB(1,-1)
# [1] 2

Answer 2

Thanks to brilliant programmers I have three alternatives. I ran a benchmark test and Alternatives 1 and 2 are almost exactly the same and faster than the Alternative 3 that uses ifelse.

APlusAbsB1 <- function(a, b){(if(b > 0) `+` else `-`)(a, b)}
APlusAbsB2 <- function(a, b) {`if`(`>`(b, 0), `+`, `-`)(a, b)}
APlusAbsB3 <- function(a, b) { ifelse(b > 0, `+`, `-`)(a, b)}

microbenchmark(APlusAbsB1(sample(10, 1), sample(-10:10, 1)),
               APlusAbsB2(sample(10, 1), sample(-10:10, 1)),
               APlusAbsB3(sample(10, 1), sample(-10:10, 1)), times=10000)
Unit: microseconds
                                         expr   min    lq median    uq     max neval
 APlusAbsB1(sample(10, 1), sample(-10:10, 1)) 3.925 4.529  4.831 5.133 912.546 10000
 APlusAbsB2(sample(10, 1), sample(-10:10, 1)) 3.925 4.529  4.830 5.132 854.587 10000
 APlusAbsB3(sample(10, 1), sample(-10:10, 1)) 4.831 5.736  5.737 6.340 851.568 10000