no visible binding for global variable for CRAN check

[kongdd]

when using zeallot, package can't pass CRAN check:

c(d_fit, info_peak) %<-% rough_fitting(INPUT)

# checking R code for possible problems ... NOTE
opt_season: no visible binding for global variable 'd_fit'
opt_season: no visible binding for global variable 'info_peak'

Undefined global functions or variables:
d_fit info_peak

Any suggestions?

[markfairbanks]

To pass CRAN checks you just need to use globalVariables() somewhere in your package to let CRAN checks know to ignore these variables.

your_function <- function(val) {
  c(d_fit, info_peak) %<-% rough_fitting(INPUT)
}

globalVariables(c("d_fit", "info_peak"))

You can specify all such variables throughout your package in a single globalVariables() call, or you can use it multiple times throughout the package.

[wleoncio]

Using globalVariables() is quite a controversial thing. The documentation for the function itself gives an example involving some local variables that codetools fails to find, but ends up stating that "it is much better practice to write code that can be checked thoroughly". In any case, I feel it's improper to declare variables as global when they are not.

Another solution would be to simply initialize the variables before the assignment:

d_fit <- info_peak <- NA  # or whatever makes more sense than NA
c(d_fit, info_peak) %<-% rough_fitting(INPUT)

[nteetor]

The dotty package includes a fix for this, https://github.com/kevinushey/dotty/blob/main/R/dotify.R

[nteetor]

The development version now includes a zeallous() function. Call this function in a package's .onLoad function to prevent the visible binding NOTEs produced by R CMD check.

.onLoad <- function(libname, pkgname) {
  zeallous()
}

^{Created on 2025-04-17 with reprex v2.1.1}

Thank you to @kevinushey and the dotty package for making me aware of codetools and this approach.