extracting grouping variables for fgroup_by using collapse package

Question

I want to create a custom function for the collapse package where I supply unquoted grouping variables like so:

library(collapse)
library(tidyverse)

fgroup_by_no_entries <- function(df, ...) {
  df %>% 
    fgroup_by(...) %>%
    fmutate(no_of_entries = GRPN()) %>%
    fungroup()
}

mtcars %>% 
  fgroup_by_no_entries(., c(cyl, gear))
# incorrect solution

# or
fgroup_by_no_entries <- function(df, grouping_vars) {
  grouping_vars <- ensyms(grouping_vars)  
  # or
  # grouping_vars <- as.list(substitute(grouping_vars))[-1]  
  
  df %>%
    fgroup_by(!!!grouping_vars) %>%  
    fmutate(no_of_entries = GRPN()) %>%
    fungroup()
}
mtcars %>% 
  fgroup_by_no_entries(., c(cyl, gear))
# doesnt work

If I was using dplyr, I could use across like so:

group_by_no_entries <- function(df, grouping_vars) {
  df %>%
    group_by(across({{grouping_vars}})) %>% 
    mutate(no_of_entries = n()) %>%
    ungroup()
}
mtcars %>% 
  group_by_no_entries(., c(cyl, gear))

but it seems across within collaspe can only be used within fmutate and fsummarise, see here.

How do I capture grouping names within a collapse function?

thanks

Answer 1

What I think is happening is that fgroup_by (or GRP) does not handle (yet) non-standard evaluation. Similarly, group_by would fail without the curly braces {{. I am not sure why you get this result (when using c).

The problem is that {{ is a tidyverse-specific non-standard evaluation function, that collapse has not yet implemented. I am not sure how collapse deals with non-standard evaluation, but the simplest way I could find was to use collapse::.c instead of c, using your first function.

.c is a "small helper function" that allows non-standard (that is, not quoted) concatenation, such that .c(a, b) == c("a", "b").

mtcars %>% 
  fgroup_by_no_entries(., .c(cyl, gear))

all.equal(
    fgroup_by_no_entries(mtcars, .c(cyl, gear)) |> 
      getElement("no_of_entries"),
    group_by_no_entries(mtcars, .c(cyl, gear)) |> 
      getElement("no_of_entries")
)
#[1] TRUE

Answer 2

One way to do this is to capture the function's arguments and convert calls of type c(x, y) to c("x", "y"). This is assuming that you don't want to use dplyr::group_by() which is compatible with collapse::fmutate() which is probably the easiest thing to do but not quite as fast.

library(collapse)
  
fgroup_by_no_entries  <- function(df, ...) {

  args <- substitute(alist(df, ...))
  args[-1] <- lapply(args[-1], \(x) {
    if (is.call(x) && identical(x[[1]], quote(c))) {
      x[-1] <- as.character(x[-1])
    }
    x
  })
  do.call(fgroup_by, eval(args)) |> 
    fmutate(no_of_entries = GRPN()) |> 
    fungroup()
  }

mtcars |>
  fgroup_by_no_entries(c(gear, carb)) |>
  head(4)

                mpg cyl disp  hp drat    wt  qsec vs am gear carb no_of_entries
Mazda RX4      21.0   6  160 110 3.90 2.620 16.46  0  1    4    4             4
Mazda RX4 Wag  21.0   6  160 110 3.90 2.875 17.02  0  1    4    4             4
Datsun 710     22.8   4  108  93 3.85 2.320 18.61  1  1    4    1             4
Hornet 4 Drive 21.4   6  258 110 3.08 3.215 19.44  1  0    3    1             3

This also works with the other selecting options offered by fgroup_by() e.g.

mtcars |> 
  fgroup_by_no_entries(gear:carb) |>
  head(4)

                mpg cyl disp  hp drat    wt  qsec vs am gear carb no_of_entries
Mazda RX4      21.0   6  160 110 3.90 2.620 16.46  0  1    4    4             4
Mazda RX4 Wag  21.0   6  160 110 3.90 2.875 17.02  0  1    4    4             4
Datsun 710     22.8   4  108  93 3.85 2.320 18.61  1  1    4    1             4
Hornet 4 Drive 21.4   6  258 110 3.08 3.215 19.44  1  0    3    1             3

mtcars |>
  fmutate(across(gear:carb, as.factor)) |> 
  fgroup_by_no_entries(is.factor) |>
  head(4)

                mpg cyl disp  hp drat    wt  qsec vs am gear carb no_of_entries
Mazda RX4      21.0   6  160 110 3.90 2.620 16.46  0  1    4    4             4
Mazda RX4 Wag  21.0   6  160 110 3.90 2.875 17.02  0  1    4    4             4
Datsun 710     22.8   4  108  93 3.85 2.320 18.61  1  1    4    1             4
Hornet 4 Drive 21.4   6  258 110 3.08 3.215 19.44  1  0    3    1             3