get(as.character(FUN), mode = "function", envir = envir) 中的错误：模式 'function' 的对象 'guide_legend'

Question

I'm building a package and I have encountered a problem when building a vignette. I get an error message when I try to knit the vignette, but when running function from console there are no errors.

Example of input data

a <- c(1,1,2,2,3,3,4,4,5,5)
b <- c("T", "T", "N","N", "N","N", "T", "T", "N", "N")
c <- c("car","bike","car","bike","car","bike","car","bike","car","bike")
d <- c(10,12,15,19,25,29,38,47,66,90)
data <- data.frame(a,b,c,d)
colnames(data) <- c("xaxis", "rug", "type", "val")

Inside a package .r file this is a function I would like to call:

#' Test
#'
#' @param data test data
#'
#' @return
#' @importFrom ggplot2 ggplot geom_point geom_rug
#' @importFrom ggnewscale new_scale_color
#' @export
#'
test <- function(data){
  ggplot(data, aes(x = xaxis, y = val, color = type, group = type)) + geom_point() +
    new_scale_color() + geom_rug(aes(x = xaxis, color =rug), sides = "b")
}

I checked and installed the package and tried to call this test function in an RMarkdown document (vignette). This is the error i get:

Error in get(as.character(FUN), mode = "function", envir = envir) : object 'guide_legend' of mode 'function' was not found Calls: ... bump_aes_scales -> lapply -> FUN -> match.fun -> get Execution halted

However, when I call this function in console I get the expected plot with no error messages:

Expected_output

Furthermore, when I delete new_scale_color() from the test function I get no errors and vignette is knitted perfectly. Does anybody have an idea what seems to be the problem here and how to solve it?

Answer 1

The problem seems to be in the ggnewscale package, at this line, where it tries to create a call to ggplot::guide_legend . This is done via match.fun , which then throws an error at this line . It may be because match.fun only looks back through 2 layers of parent environments for the named function. You could try adding a non-exported, local, alias function to your own package prior to the problematic function:

guide_legend <- function(...) ggplot2::guide_legend(...)