[R-pkg-devel] mget with Inherits Not Finding Variable in Caller

[Berry Boessenkool]

I had the same problem recently when developing a package and have some code here that might help you understand scoping a bit more.
It highlights the subtle difference between parent frame and environment.
I don't want to spam this list, so feel free to ignore ;)

Berry


a <- function(x) {aa <- "stuff"; b(x)}
b <- function(y) {
   print(ls(parent.frame()))            # parent.frame: aa, x      env in which the function was called  - dynamic scoping
   print(ls(parent.env(environment())) )# parent.env: a, b   enclosing env in which function was defined - lexical scoping
   message(aa) # # error: 'aa' not found         -      dynGet("aa") would work
}
a(7)
a <- function(x) {aa <- "stuff"; b(x)}
b <- function(y) {env <- parent.frame(); eval(substitute(aa), env)}
a(7)
b <- function(y) {eval.parent(substitute(aa))}
a(7)
b <- function(y) {eval.parent(aa)}
a(7) # aa not found: when passed to eval.parent, it is evaluated first (unless substituted)



________________________________
From: R-package-devel <r-package-devel-bounces using r-project.org> on behalf of Simon Urbanek <simon.urbanek using R-project.org>
Sent: Wednesday, December 1, 2021 00:25
To: Dario Strbenac <dstr7320 using uni.sydney.edu.au>
Cc: r-package-devel using r-project.org <r-package-devel using r-project.org>
Subject: Re: [R-pkg-devel] mget with Inherits Not Finding Variable in Caller


Dario,


> On Dec 1, 2021, at 12:00 PM, Dario Strbenac <dstr7320 using uni.sydney.edu.au> wrote:
>
> Good day,
>
> What I am misunderstanding about the inherits = TRUE option of mget? I expect the small example to work.
>
> f <- function(x, .iteration = i) g()
> g <- function() mget(".iteration", inherits = TRUE)
> f(10, 1)
> Error: value for �.iteration� not found
>

That has nothing to do with inherits and is expected - it's identical to

> f <- function(x, .iteration = i) g()
> g <- function() .iteration
> f(10, 1)
Error in g() : object '.iteration' not found

Please note that R is lexically scoped and you defined g in the global environment so it has no way of seeing inside f. This would work:

> f <- function(x, .iteration = i) {
+   g <- function() .iteration
+   g()
+ }
> f(10, 1)
[1] 1

since then the environment of f is the parent env of g.

If you want dynamic scoping (not what R uses!) you can use dynGet():

> f <- function(x, .iteration = i) g()
> g <- function() dynGet(".iteration")
> f(10,1)
[1] 1

but since that is non-standard the docs warn:

     �dynGet()� is somewhat experimental and to be used _inside_
     another function.  It looks for an object in the callers, i.e.,
     the �sys.frame()�s of the function.  Use with caution.

Cheers,
Simon




> --------------------------------------
> Dario Strbenac
> University of Sydney
> Camperdown NSW 2050
> Australia
> ______________________________________________
> R-package-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-package-devel
>

______________________________________________
R-package-devel using r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel

	[[alternative HTML version deleted]]