Avi, hi, thanks for your reply. I was vaguely aware of the '...'
construct, but hadn't ever worked out how to actually use it. Thank you
so much for teaching me that. I agree that I should use it in my
function.
Yeah, I just used the parentheses around some of my assignments to
avoid the trouble of typing the variable into R to verify the contents.
My work is all contained in a Rmarkdown file, that I finally export and
file with my notes from class to help me study for exams.
Thank you, again, for your generous help for me.
-Kevin
On Sat, 2025-03-15 at 12:00 +0100, r-help-requ...@r-project.org wrote:
> Kevin,
>
> I was amused by the use of the parentheses wrapping to get the REPL
> to show the effects of an assignment but would remove that in any
> final program if the output is not needed.
>
> I am not saying this is wrong, nor what I describe below, but just a
> discussion of how others might do it.
>
> I do somewhat wonder about the way you define the function below:
>
> compute.diff.means <- function(x) {
> return(mean(x[1:n_lt]) - mean(x[(n_lt+1):(n_lt+n_dk)]))
> }
>
> It is a function of x which you pass in but makes use of an external
> variable that it needs to find in the environment several times,
> n_lt, as well as n_dk. But these assignments happen just once in your
> code:
>
> (n_lt <- length(lt))
> (n_dk <- length(dk))
>
> Some people would write the function to include the variables as in:
>
> compute.diff.means <- function(x, n_lt, n_dk) {
> return(mean(x[1:n_lt]) - mean(x[(n_lt+1):(n_lt+n_dk)]))
> }
>
> The names of the variables can be the same or new ones.
>
> The reason you do this seems to be that you are using "apply" as
> shown below and may not know it can accommodate additional argument.
>
> (random.statistics <- apply(random.samples, 1, compute.diff.means))
>
> The above is an implicit loop that calls compute.diff.means()
> repeatedly over each row of your matrix. It passes the specific row
> as a vector as the first and only argument.
>
> If you ask "?apply" to document what the apply function does, you may
> note that like some other such functions, there is a "..." that
> actually means anything else you supply as extra arguments are passed
> along to the function. So, since your variables are not changing,
> then code like this:
>
> (random.statistics <- apply(random.samples, 1, compute.diff.means,
> n_lt, n_dk))
>
> Will call a function with a row vector and then the additional two
> arguments so each call will be to:
>
> compute.diff.means(ROW, n_lt, n_dk)
>
> Arguably, this approach may be no better but in some sense makes your
> function more portable and cleaner. If your code continued and did
> additional analyses like this, the function might be more easily re-
> usable.
>
>
______________________________________________
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide https://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
