How can I dynamically group_by a dataframes variables in a function? - r

Question

I want a function where i can enter different numbers of column names and have them grouped. The first piece of code here works:
df <- data.frame(col_a = sample(1:10, 100, replace = T),
col_b = sample(letters, 100, replace = T),
col_c = sample(LETTERS, 100, replace = T))
my_fun = function(df, ...) {
df %>% group_by_(...) %>% summarise(n = n())
}
my_fun(df , 'col_a')
my_fun(df , 'col_a', 'col_b')
my_fun(df , 'col_a', 'col_b', 'col_c')
What I now want is to apply the complete function, so all possible values in each grouped variable are present. I've manually typed col_a and col_b into the complete() function below. I'd want to pass the possible values as a function argument though, as I'm not always going to be grouping by col_a & col_b.
my_fun = function(df, ...) {
df %>% group_by_(...) %>% summarise(count = n()) %>%
ungroup() %>%
complete(col_a = 1:10, col_b = letters, fill = list(count = 0))
}
my_fun(df , 'col_a', 'col_b')

Answer 1

You can capture the data as named list. group_by + summarise n() can be replaced with count.
library(tidyverse)
my_fun = function(df, ...) {
args <- list(...)
df %>%
count(across(all_of(names(args))), name = 'count') %>%
complete(!!!args, fill = list(count = 0))
}
This can be ran as -
my_fun(df , 'col_a' = 1:12)
# col_a count
# <int> <dbl>
# 1 1 9
# 2 2 15
# 3 3 4
# 4 4 11
# 5 5 7
# 6 6 12
# 7 7 12
# 8 8 10
# 9 9 5
#10 10 15
#11 11 0
#12 12 0
my_fun(df , 'col_a' = 1:10, 'col_b' = letters)
# col_a col_b count
# <int> <chr> <dbl>
# 1 1 a 1
# 2 1 b 0
# 3 1 c 0
# 4 1 d 0
# 5 1 e 0
# 6 1 f 1
# 7 1 g 0
# 8 1 h 0
# 9 1 i 0
#10 1 j 0
# … with 250 more rows