Feature request: at function in summarise, mutate, ...

[mgirlich]

I often encounter the situation where summarise_at nearly does the job, but a different function is needed for one column. E.g.

library(dplyr)

mtcars %>% 
  group_by(cyl) %>% 
  summarise(
    mpg = mean(mpg),
    disp = mean(disp),
    hp = mean(hp),
    drat = mean(drat),
    wt = max(wt),
  )

Instead of the above code, it would be very nice to be able to write something like

mtcars %>% 
  group_by(cyl) %>% 
  summarise(
    at(c("mpg", "disp", "hp", "drat"), mean),
    wt = max(wt)
  )

Using the (not exported) function dplyr:::manip_at there basically already is a function nearly suited for the task

at <- function(.tbl, .vars, .funs, ...) {
  # .vars <- check_dot_cols(.vars, .cols)
  dplyr:::manip_at(
    .tbl, .vars, .funs, enquo(.funs), caller_env(), 
    .include_group_vars = TRUE, ...
  )
}


mtcars %>% 
  group_by(cyl) %>% 
  {summarise(
    .,
    !!!at(., c("mpg", "disp", "hp", "drat"), mean),
    wt = max(wt)
  )}

I would very much appreciate such a functionality integrated into dplyr.

[moodymudskipper]

It would be nice if this could work :

mtcars %>% 
  group_by(cyl) %>% 
  summarize(!!!map(lst(mpg, disp, hp, drat),mean),
            wt = max(wt))

This "almost" works :

mtcars %>% 
  group_by(cyl) %>% 
  summarize(!!!with(.,map(lst(mpg, disp, hp, drat),mean)),
            wt = max(wt))
# # A tibble: 3 x 6
#     cyl      mpg     disp       hp     drat    wt
#   <dbl>    <dbl>    <dbl>    <dbl>    <dbl> <dbl>
# 1     4 20.09062 230.7219 146.6875 3.596563 3.190
# 2     6 20.09062 230.7219 146.6875 3.596563 3.460
# 3     8 20.09062 230.7219 146.6875 3.596563 5.424  

And this works:

mtcars %>% 
  group_by(cyl) %>% 
  summarize(list(map_dfc(lst(mpg, disp, hp, drat),mean)),
            wt = max(wt)) %>%
  unnest

# A tibble: 3 x 6
#     cyl    wt   mpg  disp    hp  drat
#   <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1     4  3.19  26.7  105.  82.6  4.07
# 2     6  3.46  19.7  183. 122.   3.59
# 3     8  5.42  15.1  353. 209.   3.23

[hadley]

Our current plan is to support something like this via #2326

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/