Replace globalvariables use with maintainable approach

[bailliem]

Work through the list of variables referenced in visR globalvariable to set those variables to NULL within the function call, where possible. At the moment we are using this to suppress note in R CMD check "no visible binding". Ideally, we resolve this at the function level.

See here for more details on the alternative approach STAT545-UBC/Discussion#451

[timtreis]

In the spirit of closing down issues:

Apparently the tidyr approved way is the usage of the .data pronoun as described here. Shall I integrate it?

[SHAESEN2]

Can you give an example of how this looks like for us?

[timtreis]

Didn't run any tests with it yet, but I assume we'd have to replace all implicit piping by explicit .data piping. Yucky

So instead of
adtte %>% dplyr::mutate(TRTDUR = TRTDUR*2)
we should then use
adtte %>% dplyr::mutate(TRTDUR = .data$TRTDUR*2)

If we don't want to take this step, I'd be in favour of closing this issue

[SHAESEN2]

If we want to do that I would be very explicit as '$' does partial matching.

adtte %>% dplyr::mutate(TRTDUR = .data[["TRTDUR"]]*2)

[timtreis]

But do we want to do that? Doesn't feel like it adds value, is very tedious work and will probably pop up again every time one of us writes a piping statement. At least I'm fairly sure I'd forget it.

Also, the issues says "maintainable" which the .data approach very much is not 👀

[SHAESEN2]

As proposed by Daniel, we should use .data to be very explicit inside all our functions which increases traceability but also ensure that the correct environment is passed throughout functions

am <- blah

mtcars %>% select(am) = mtcars %>% select(.data$am)
mtcars %>% select(.env(am))

[ddsjoberg]

+1 for using the .data and .env prefixes. The tidyverse functions are super easy to use with the unquoted variables names, but that simplicity comes with some ambiguity. Building on the example @SHAESEN2 listed above.

am <- "mpg"
mtcars |> dplyr::select(.data$am, .env$am) |> head()
#>                   am  mpg
#> Mazda RX4          1 21.0
#> Mazda RX4 Wag      1 21.0

Without the prefixes, the selecting is not 100% precise. This is a bit of a contrived example, but I do often find myself running into issues while writing functions with data frames having column names that match R object names outside the data frame.