Running the same dplyr commands over a bunch of group_by variables

Question

I want to run the same commands (basically some general filter(), mutate() summarise() etc.) over a bunch of variables with the group_by() command. Is there a way to bundle this process instead of having to use a block of code for each individual variable? I've tried using a for loop, which unfortunately didn't work for me. My goal is to only have to write one block of code for all group_by variables.

library(dplyr)
         
df <- mtcars         
         
Test1 <- df %>%
  group_by(mpg) %>%
  filter(hp > 100) %>%
  summarise(N = n())

Test2 <- df %>%
  group_by(cyl) %>%
  filter(hp > 100) %>%
  summarise(N = n())

Test3 <- df %>%
  group_by(disp) %>%
  filter(hp > 100) %>%
  summarise(N = n())

Thank you in advance!

Answer 1

You can create the vector of variables you want to group by and iterate over that using purrr::map() and use dplyr::count() to succinctly group_by & summarize(). One key thing here is to surround the first map() statement with {} curly braces to allow you to pipe the . into a something other than the first argument. The other key thing here is that if you want to filter by hp > 100, do that upstream to avoid repeating it. Finally I pivot_longer() so that the variable name in each output can be combined into a single data.frame with the same column names.

Also note that for this example I chose different variables that work better as a grouping keys because they are basically discrete variables.

library(tidyverse)
         
varz <- c("cyl", "am", "gear", "carb")

mtcars %>%
  filter(hp > 100) %>% 
  {map(varz, \(x){count(., !!sym(x))})} %>% 
  map(~pivot_longer(.x, -n, names_to = "var")) %>% 
  bind_rows() %>% 
  select(var, value, n)
#> # A tibble: 14 x 3
#>    var   value     n
#>    <chr> <dbl> <int>
#>  1 cyl       4     2
#>  2 cyl       6     7
#>  3 cyl       8    14
#>  4 am        0    16
#>  5 am        1     7
#>  6 gear      3    14
#>  7 gear      4     5
#>  8 gear      5     4
#>  9 carb      1     2
#> 10 carb      2     6
#> 11 carb      3     3
#> 12 carb      4    10
#> 13 carb      6     1
#> 14 carb      8     1

^{Created on 2022-02-18 by the reprex package (v2.0.1)}

Answer 2

Since you already use a tidyverse approach you could just pivot the table:

library(tidyverse)

df <- mtcars

df %>%
  filter(hp > 100) %>%
  select(mpg, cyl, hp) %>%
  pivot_longer(cols = c(mpg, cyl, hp),
               names_to = "var",
               values_to = "value") %>%
  group_by(var) %>%
  summarise(N = n())
#> # A tibble: 3 × 2
#>   var       N
#>   <chr> <int>
#> 1 cyl      23
#> 2 hp       23
#> 3 mpg      23

Would that solve your problem? This way you can put any functions after the pivoted dataset and apply them to all variables as you wish. You can pivot it back if you prefer to have them as variables again. The main advantage of this approach is that you never split your dataset. However, there are downsides to this as well.

Another option could be using purrr.

Answer 3

One option would be to store the variable names in a vector or list, then loop over this list using e.g. lapply where I make use of the .data pro-noun:

library(dplyr)

df <- mtcars         

var_list <- c("mpg", "cyl", "disp")

lapply(var_list, function(x) {
  df %>%
    group_by(.data[[x]]) %>%
    filter(hp > 100) %>%
    summarise(N = n())  
})
#> [[1]]
#> # A tibble: 18 × 2
#>      mpg     N
#>    <dbl> <int>
#>  1  10.4     2
#>  2  13.3     1
#>  3  14.3     1
#>  4  14.7     1
#>  5  15       1
#>  6  15.2     2
#>  7  15.5     1
#>  8  15.8     1
#>  9  16.4     1
#> 10  17.3     1
#> 11  17.8     1
#> 12  18.1     1
#> 13  18.7     1
#> 14  19.2     2
#> 15  19.7     1
#> 16  21       2
#> 17  21.4     2
#> 18  30.4     1
#> 
#> [[2]]
#> # A tibble: 3 × 2
#>     cyl     N
#>   <dbl> <int>
#> 1     4     2
#> 2     6     7
#> 3     8    14
#> 
#> [[3]]
#> # A tibble: 18 × 2
#>     disp     N
#>    <dbl> <int>
#>  1  95.1     1
#>  2 121       1
#>  3 145       1
#>  4 160       2
#>  5 168.      2
#>  6 225       1
#>  7 258       1
#>  8 276.      3
#>  9 301       1
#> 10 304       1
#> 11 318       1
#> 12 350       1
#> 13 351       1
#> 14 360       2
#> 15 400       1
#> 16 440       1
#> 17 460       1
#> 18 472       1

EDIT To store the results in a dataframe I would first name the vector of variables, inside the function use the same name for the grouping column. i.e. I simply used "value" then use bind_rows to bind the results together where as an identifier I added the variable name:

var_list <- c("mpg", "cyl", "disp")
names(var_list) <- var_list
lapply(var_list, function(x) {
  df %>%
    group_by(value = .data[[x]]) %>%
    filter(hp > 100) %>%
    summarise(N = n())  
}) %>% 
  bind_rows(.id = "var")
#> # A tibble: 39 x 3
#>    var   value     N
#>    <chr> <dbl> <int>
#>  1 mpg    10.4     2
#>  2 mpg    13.3     1
#>  3 mpg    14.3     1
#>  4 mpg    14.7     1
#>  5 mpg    15       1
#>  6 mpg    15.2     2
#>  7 mpg    15.5     1
#>  8 mpg    15.8     1
#>  9 mpg    16.4     1
#> 10 mpg    17.3     1
#> # ... with 29 more rows