Wrap function passing NULL to lower-level haven::read_dta function in R - r

Question

I am trying to build a function wrapping over haven::read_dta() similar to the wrap_function() defined in the code below.
My wrap_function() has a default variables = NULL, which should be able to pass NULL to haven::read_dta()'s col_select argument if no values are specified. However, passing the NULL from variables to col_select throws an error (i.e. 'Error: Can't find any columns matching col_select in data.').
Can someone help me understand why this happens and how could I build a wrap_function capable of passing a NULL default value to the lower-level function?
Thanks!
library(reprex)
library(haven)
df_ <- data.frame(a = 1:5,
b = letters[1:5])
haven::write_dta(df_,
path = "file.dta")
# works well:
haven::read_dta(file = "file.dta",
col_select = NULL)
#> # A tibble: 5 x 2
#> a b
#> <dbl> <chr>
#> 1 1 a
#> 2 2 b
#> 3 3 c
#> 4 4 d
#> 5 5 e
# does not work:
wrap_function <- function(file, variables = NULL){
haven::read_dta(file = file,
col_select = variables)
}
wrap_function("file.dta")
#> Note: Using an external vector in selections is ambiguous.
#> ℹ Use `all_of(variables)` instead of `variables` to silence this message.
#> ℹ See <https://tidyselect.r-lib.org/reference/faq-external-vector.html>.
#> This message is displayed once per session.
#> Error: Can't find any columns matching `col_select` in data.
sessioninfo::session_info()
#> ─ Session info ───────────────────────────────────────────────────────────────
#> setting value
#> version R version 4.0.3 (2020-10-10)
#> os CentOS Linux 8
#> system x86_64, linux-gnu
#> ui X11
#> language (EN)
#> collate en_US.UTF-8
#> ctype en_US.UTF-8
#> date 2021-05-14
#>
#> ─ Packages ───────────────────────────────────────────────────────────────────
#> package * version date lib source
#> cli 2.4.0 2021-04-05 [1] CRAN (R 4.0.3)
#> crayon 1.4.1.9000 2021-04-16 [1] Github (r-lib/crayon#965d1dc)
#> digest 0.6.27 2020-10-24 [1] CRAN (R 4.0.3)
#> ellipsis 0.3.1 2020-05-15 [1] CRAN (R 4.0.2)
#> evaluate 0.14 2019-05-28 [1] CRAN (R 4.0.2)
#> fansi 0.4.2 2021-01-15 [1] CRAN (R 4.0.3)
#> forcats 0.5.1 2021-01-27 [1] CRAN (R 4.0.3)
#> fs 1.5.0 2020-07-31 [1] CRAN (R 4.0.2)
#> glue 1.4.2 2020-08-27 [1] CRAN (R 4.0.2)
#> haven * 2.3.1 2020-06-01 [1] CRAN (R 4.0.2)

Answer 1

TLDR: You just need to embrace the argument by wrapping it in double curly brackets{{ }}, previously called "curly-curly". This passes the variable properly. See the programming with dplyr vignette for more info.
wrap_function <- function(file, variables = NULL){
haven::read_dta(file = file,
col_select = {{ variables }})
}
wrap_function("file.dta")
#> # A tibble: 5 x 2
#> a b
#> <dbl> <chr>
#> 1 1 a
#> 2 2 b
#> 3 3 c
#> 4 4 d
#> 5 5 e
Unfortunately it's a little hard to understand that this is necessary without looking at the code. If you look up the haven repository, you can see that read_dta uses the double-curly around col_select as well. This is a pretty good indication that you need to use it in your wrapper function.
If you look further, it is using them to pass the argument to a function skip_cols, which uses it inside tidyselect::vars_select. The reason this is needed is so that you can delay evaluation of the argument until the point that you actually need it. In other words, it lets you call the function like this:
wrap_function("file.dta", variables = a)
instead of forcing you to do something like
wrap_function("file.dta", variables = "a")
and saves you a lot of typed quotes, especially with a lot of columns. You see this pattern in dplyr and other tidyverse functions a lot, especially any time an argument refers to a dataframe column rather than a variable.
In other words, you want to not have the code check exactly what a is until you reach skip_cols, which knows that a refers to a column inside the file you're reading. If you don't use the curly braces, it will think that a is some object in your working environment.