Reputation: 6213
I need to intercept the value of an optional xlim
in a function so that I can change the units of it before plotting. The following function confirms that xlim
was passed, but I can't access the value.
foo <- function(x, y, ...) {
if ("xlim" %in% names(list(...))) {
print(xlim) # not found/can't use value!
}
# modify xlim and pass to plotting functions
return()
}
But foo(x = 1:5, y = 1:5, xlim = c(2,4))
gives:
Error in print(xlim) : object 'xlim' not found
What trick do I need use the value? Seems like it should just work, but I see from looking around on SO that the dots can be vexing. I've played a bit with exists
, deparse
etc but I don't really 'get' the proper use of those functions.
EDIT: so here is the final snippet which was the leanest way to access the value:
dots <- list(...)
if (any(names(dots) == "xlim")) {
xlim <- dots$xlim
print(xlim)
}
Upvotes: 4
Views: 77
Reputation: 99331
This is because xlim
is actually a list element, and is not (yet) an actual object in the function's environment. You could do
foo <- function(x, y, ...) {
m <- match.call(expand.dots = FALSE)$...
if(any(names(m) == "xlim")) m[["xlim"]]
else stop("no xlim value")
}
foo(x = 1:5, y = 1:5, xlim = c(2,4))
# c(2, 4)
foo(x = 1:5, y = 1:5, ylim = c(2,4))
# Error in foo(x = 1:5, y = 1:5, ylim = c(2, 4)) : no xlim value
You can see what match.call
is doing if we examine the function as
f <- function(x, y, ...) {
match.call(expand.dots = FALSE)$...
}
It is a list of all the entered dot arguments with their respective expressions, so there are many different ways to get the values, the above is just one way.
f(x = 1:5, y = 1:5, xlim = c(2,4))
# $xlim
# c(2, 4)
Alternatively, you could do
g <- function(x, y, ...) {
dots <- list(...)
any(names(dots) == "xlim")
}
g(x = 1:5, y = 1:5, xlim = c(2,4))
# [1] TRUE
Also keep in mind that match.call
keeps the argument as an unevaluated call
, while list(...)
evaluates the argument. This might be important for you passing the argument to other functions.
Upvotes: 2