Why does <<- assignment work in my function, but <- doesn't?

Question

I am trying to understand the difference between <- and <<- in practice. I wrote the following function in R that relies on a couple of other small function that I wrote:

fun.exec <- function(x=dat){
  id1 <- prompt1()
  id2 <- prompt2()
  el.type <- data.switch(di=id1)
  dat.sifted <- data.sift(x, nc=id2)
  plots.list <- evol.tiles(ds=dat.sifted, dt=el.type, nc=id2)
  p <- evol.plot(l=plots.list, dt=el.type)
}

Functions prompt1 and prompt2 take an input from a user, el.type() assigns string name to the data (for use in describing different plots automatically), data.sift() extract relevant data from a big data frame object, evol.tiles() generates various ggplots to be organized in a grid, and evol.plot() puts the plots in a grid.

As can be seen, both data.sift() and evol.tiles() functions use the id2 user's input. When I execute this function as is, I get an error:

Error in evol.tiles(ds = dat.sifted, dt = el.type, nc = id2) : object 
'id2' not found 

If I replace id2 <- prompt2() with id2 <<- prompt2(), the code works as expected.

What I don't understand is why, as is, the code does not break on the data.sift() function, which also calls for id2. I read help for assignments, a couple of related posts on StackOverflow, and the Scope section from An Introduction to R but I am still not sure what the problem is. It's almost as if after being used in data.sift() the variable was no longer available in the environment and I don't understand that is.

Any help will be greatly appreciated.

UPDATE: Here is the code for prompts:

prompt1 <- function(){
  cat('What do you want to create plots for? Your options are:
        1: data type A,
        2: data type B,
        3: data type C')
  readline(prompt="Enter an integer: ")
}

prompt2 <- function(){
  cat('How many nodes do you want to visualize?')
  n <- readline(prompt="Enter an integer: ")
  cat('\nProvide coordinates of each node to visualize separated by commas.')
  l <- vector("list", n)
  for (i in 1:n){
    el <- readline(prompt=paste('Enter coordinnates for node',i,': '))
    l[[i]] <- el
  }
  return(l)
}

for data.sift():

data.sift <- function(x, nc){
  nl <- lapply(nc, function(l){as.integer(unlist(strsplit(l,",")))})
  ds <- vector("list", length(nl))
  for (i in 1:length(nl)){
    ds[[i]] <- x[(x$x == nl[[i]][1] & x$y == nl[[i]][2] & x$z == nl[[i]][3]),]
  }
  return(ds)
}

and for evol.tiles():

evol.tiles <- function(ds, dt, nc){
  require(ggplot2)
  my.cols <- rainbow(length(ds))
  my.names <- as.character(nc)
  names(my.cols) <- my.names

  my.list <- list()
  for (i in 1:6){
    for (ii in 1:length(id2)){
      p <- ggplot(NULL, aes_(x = as.name(names(ds[[ii]][4]))))
      p <- p + geom_line(data = ds[[ii]], 
                         aes_(y = as.name(names(ds[[ii]][i])), 
                              colour = as.character(nc[[ii]])))
    }
    p <- p  + scale_colour_manual("Node",
                          breaks = as.character(nc),
                          values = my.cols)
    my.list[[i-dr[1]+1]] <- p
  }
  return(my.list)
}

Answer 1

As posted in the comments, I think I found the issue - while working on a minimal working example I discovered that in my function evol.tiles() I was calling the id2 variable instead of nc (in the inner loop). I guess when I used <<- for prompt2() I was assigned it globally and then it could be found when called from within evol.tiles() but with <- it was not available to evol.tiles().

Even so, I still don't quite understand why was that happening. I would think that that the function should look in the parenting environment for the missing arguments and since the id2 was defined within fun.exec(), the time.evol() should be able to find the right value.

Here is a simple example showing how I would expect the code to behave, even with an incorrectly named variable as it was in my case:

test <- function(){x*x}
test()
Error in test() : object 'x' not found

If I run the test() function on its own, I get the same error as with my function, whic is what I would expect. However, if I assign value to x, i.e.,

x <- 2
test()
[1] 4 

the function works just fine. Could someone tell me why wasn't my function behaving in the same way?

UPDATE: @Aaron: The example you gave, i.e., rm(x); test2 <- function() { x <- 2; test() }; test2(), executes for me just fine:

So I still don't understand what the issue is. Any thoughts?