More examples in `across()` documentation

[AmeliaMN]

When I use across() I almost always want to do the thing to every single column. For whatever reason, that tidyselect action is hard for me to remember/find. Is it all()? Is it everything()? Etc. It would be great to have an example in the across() documentation that does the thing to every column. Additionally, it would be awesome to have an example that does the thing conditionally depending on the variable type. For example, rounding all the numeric variables.

[marcuslehr]

I second this. Honestly I find across() super confusing and frequently revert to the _if/_at/_all functions to do what I'm trying to do. I've tried to learn across() several times and I either can't figure out how to make it work or it's very painful. I'd frankly be onboard with reverting to the legacy functions, but I'm sure that won't happen so some better examples in the doc would certainly help.

[philibe]

For across() on all columns I use something like that:

library(dplyr)
library(tibble)
library(stringr)
mydatas <- (mtcars 
            %>%tibble::rownames_to_column() 
            %>% rename (
              mpg_aaa=mpg,
              disp_bbb=disp,
              model=rowname
            )            
            %>% mutate(
              brand=stringr::str_split_i(model  ,"\\s", 1)
            )
            %>% select(
              brand,
              everything()
            )
)
mydatas

mydatas_regroup <- (mydatas
                    %>% group_by(brand)
                    %>% summarise(
                      across(everything(),
                             ~ (function (var,val) {
                               if (is.numeric(val) & (stringr::str_detect (var,"_aaa|_bbb")) ) {
                                 round( sum(val, na.rm = TRUE),0)
                               } else if (is.numeric(val) ) {
                                 round( sum(val, na.rm = TRUE),2)
                               } else {
                                 paste0(
                                   toString(
                                     paste0(
                                       
                                       "[",
                                       sort(unique(val)),
                                       "]",
                                       collapse = ",")
                                   )
                                 )
                               }   
                             })(cur_column(), .)
                      )
                    )
)

# A tibble: 22 × 13
   brand    model                  mpg_aaa   cyl disp_bbb    hp  drat    wt  qsec    vs    am  gear  carb
   <chr>    <chr>                    <dbl> <dbl>    <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
 1 AMC      [AMC Javelin]               15     8      304   150  3.15  3.44  17.3     0     0     3     2
 2 Cadillac [Cadillac Fleetwood]        10     8      472   205  2.93  5.25  18.0     0     0     3     4
 3 Camaro   [Camaro Z28]                13     8      350   245  3.73  3.84  15.4     0     0     3     4
 4 Chrysler [Chrysler Imperial]         15     8      440   230  3.23  5.34  17.4     0     0     3     4
 5 Datsun   [Datsun 710]                23     4      108    93  3.85  2.32  18.6     1     1     4     1
 6 Dodge    [Dodge Challenger]          16     8      318   150  2.76  3.52  16.9     0     0     3     2
 7 Duster   [Duster 360]                14     8      360   245  3.21  3.57  15.8     0     0     3     4
 8 Ferrari  [Ferrari Dino]              20     6      145   175  3.62  2.77  15.5     0     1     5     6
 9 Fiat     [Fiat 128],[Fiat X1-9]      60     8      158   132  8.16  4.14  38.4     2     2     8     2
10 Ford     [Ford Pantera L]            16     8      351   264  4.22  3.17  14.5     0     1     5     4
# ℹ 12 more rows
# ℹ Use `print(n = ...)` to see more rows

PS: I'm a simple user

[DavisVaughan]

We do have a where(is.double) example near the top of the ?across examples, but we are missing one with everything().

I think we should add an example that uses everything() at the very top of the across() examples, because that is a very common case.