Evaluate elipsis (dots) multiple times, substitute arguments

Question

Context

I am using in R, the "elipsis" or "dots" that wrap function calls

main_function <- function(...)

If I want to evaluate once, I do

main_function <- function(...) {
    res = list(...)}

It works fine

Problem

fun_A <- function(arg_A){
    print(paste("I am A", paste0(round(runif(arg_A, 0,1), 2),collapse = ", ")))
}

fun_B <- function(arg_B){
    print(paste("I am B", paste0(round(runif(arg_B, 1,2), 2),collapse = ", ")))
}

Here the result is evaluated once and replicate 3 times :

main_fun_wrong <- function(..., times) {
    res = list(...)
    replicate(times, eval(res))
}

main_fun_wrong(fun_A(1), fun_B(2), times = 3)

Here it works :

main_fun <- function(..., times) {
    calls = match.call(expand.dots = FALSE)$`...`
    replicate(times, lapply(1:length(calls), function(num) eval(calls[[num]])), simplify = F)
}

main_fun(fun_A(1),fun_B(2), times = 3)

But now if arg_A is an object rather than a value, it will fail finding the arg_A and arg_B in the environment.

main_fun_problem <- function(arg_A, arg_B) {
    main_fun(fun_A(arg_A),fun_B(arg_B), times = 3)
}

main_fun_problem(1,2)

I got an error :

Error in fun_A(arg_A) : object 'arg_A' not found

I do not know what R do when it find list(...) the first time in first example but I just want to repeat it multiple times.

@nrussell Thx.I did not provided example but rephrased it to make it more clear — zakrapovic, Apr 28 '16 at 13:15
`replicate(10, eval(calls))` won't evaluate 10 times, it just replicates 10 times the results from single evaluation — jangorecki, Apr 28 '16 at 13:30
@jangorecki indeed but the problem remain if I wrap it in a for loop — zakrapovic, Apr 28 '16 at 13:31
you should also make example reproducible, so provided funA and funB — jangorecki, Apr 28 '16 at 13:33
this cannot be seen from your current example, funA and funB should increment a counter or prints to console — jangorecki, Apr 28 '16 at 13:34
you may also read the source of `microbenchmark::microbenchmark` function as it does exactly what you are looking for — jangorecki, Apr 28 '16 at 17:39
@jangorecki actually replicate does evaluate multiple time in the solution. I may have a look at microbenchmark source eventually. Thanks — zakrapovic, Apr 29 '16 at 14:19
I think this has more to do with your `lapply` + `<<-` attempt than the ellipsis. If you want to do something iteratively, use a `for` loop... — Frank, Apr 29 '16 at 16:48
@Frank sory I do not understand. the <<- has been used to make a state in the reproductible example and control that it has been evaluated multiple time. But I do not like it much, will change. — zakrapovic, Apr 29 '16 at 17:18
To clarify, I mean: `update_A = function(arg_A) A + arg_A; A = 0; arg_A = 5; n = 3; res = numeric(n); for (i in seq_along(res)) res[[i]] = (A <- update_A(arg_A)) ` — Frank, Apr 29 '16 at 17:21
Ok. `eval(substitute(replicate(times, list(...))))` seems fine to me. This is no big improvement over `replicate(list(whatever), n = 3)`... just saves you from manually writing `list`... — Frank, Apr 29 '16 at 18:18

zakrapovic · Accepted Answer · 2016-04-29T18:04:07.157

0

Here is my solution, any alternative will be enjoyed.

The things is to substitute the variable by it's value at the moment we call the function.

main_fun_solution <- function(arg_A, arg_B) {
    eval(substitute(main_fun(fun_A(arg_A),fun_B(arg_B), times = 3), list("arg_A" =  arg_A, "arg_B" =  arg_B)))
}

main_fun_solution(1,2)

NB: list("arg_A" = arg_A, "arg_B" = arg_B)` makes my heart bleed (the overall solution actually)

edited Apr 29 '16 at 18:04

answered Apr 28 '16 at 16:38

zakrapovic

413
3
17

Just came across [this](http://stackoverflow.com/questions/29084193/how-to-not-fall-into-rs-lazy-evaluation-trap?rq=1). I will read further, it looks like a good alternative. – zakrapovic Apr 28 '16 at 16:52

Evaluate elipsis (dots) multiple times, substitute arguments

Context

Problem

1 Answers1