In crowding/nse: Tools for Lazy and Non-Standard Evaluation

In R, arguments have tags, as opposed to just lists. You need to be able to substitute into the tags of a call.

Let's say that we wanted to have a statement "foreach" that worked like "for" but accumulated and returned the results in a list. So an example usage would look like:

sum <- 0
parts <- foreach(i, 1:3, {
  sum <- sum + i
  sqrt(i)
})

Following the principle of least surprise, let's say we want foreach to work similarly to for and mimic R's quirks regarding for: That each iteration of the body is and the placeholder variable i is defined in the environment enclosing the for statement.

vec <- i
out <- vector(length(i), list)
for (iter in seq(len(values))) {
  i <- vec[[iter]]
  out[[iter]] <- {
    sim(a)
  }
}

So far so good. What happens when we want to nest two foreach loops?

out <- function(x)

x <- function(i)

crowding/nse documentation built on Jan. 5, 2024, 12:14 a.m.