summarise + across returns different output than summarise_if

[cactusoxbird]

I'm hoping someone can help me understand why summarise(across()) and summarise_if() do not yield the same results when provided with the same group_by() specification. Here's a simplified example of what I'm running into:

library(tidyverse)
library(lubridate)

fake_data <- tribble(
  ~location, ~sub_location, ~start_date, ~end_date, ~measurement,
  "Area 1", "A", as_date("2021-04-01"), as_date("2021-04-25"), 70,
  "Area 1", "A", as_date("2021-04-01"), as_date("2021-04-25"), 75,
  "Area 1", "B", as_date("2021-04-01"), as_date("2021-04-25"), 85,
  "Area 1", "B", as_date("2021-04-01"), as_date("2021-04-25"), 67,
  "Area 2", "A", as_date("2021-04-01"), as_date("2021-04-25"), 47,
  "Area 2", "B", as_date("2021-04-01"), as_date("2021-04-25"), 90)

fake_data %>%
  group_by(location, sub_location) %>%
  summarise_if(.predicate = is.numeric,
               .funs = ~ mean(.x, na.rm = T))
#> # A tibble: 4 x 3
#> # Groups:   location [2]
#>   location sub_location measurement
#>   <chr>    <chr>              <dbl>
#> 1 Area 1   A                   72.5
#> 2 Area 1   B                   76  
#> 3 Area 2   A                   47  
#> 4 Area 2   B                   90

fake_data %>%
  group_by(location, sub_location) %>%
  summarise(across(.cols = where(is.numeric),
                   .funs = ~ mean(.x, na.rm = T))) 
#> `summarise()` has grouped output by 'location', 'sub_location'. You can override using the `.groups` argument.
#> # A tibble: 6 x 3
#> # Groups:   location, sub_location [4]
#>   location sub_location measurement
#>   <chr>    <chr>              <dbl>
#> 1 Area 1   A                     70
#> 2 Area 1   A                     75
#> 3 Area 1   B                     85
#> 4 Area 1   B                     67
#> 5 Area 2   A                     47
#> 6 Area 2   B                     90

^{Created on 2021-04-17 by the reprex package (v1.0.0)}

[andresrcs]

You are using the wrong argument name for across(), it is .fns not .funs, although, I think it should fail more gracefully.

library(tidyverse)
library(lubridate)
#> 
#> Attaching package: 'lubridate'
#> The following objects are masked from 'package:base':
#> 
#>     date, intersect, setdiff, union

fake_data <- tribble(
    ~location, ~sub_location, ~start_date, ~end_date, ~measurement,
    "Area 1", "A", as_date("2021-04-01"), as_date("2021-04-25"), 70,
    "Area 1", "A", as_date("2021-04-01"), as_date("2021-04-25"), 75,
    "Area 1", "B", as_date("2021-04-01"), as_date("2021-04-25"), 85,
    "Area 1", "B", as_date("2021-04-01"), as_date("2021-04-25"), 67,
    "Area 2", "A", as_date("2021-04-01"), as_date("2021-04-25"), 47,
    "Area 2", "B", as_date("2021-04-01"), as_date("2021-04-25"), 90)

fake_data %>%
    group_by(location, sub_location) %>%
    summarise(across(.cols = where(is.numeric),
                     .fns = ~ mean(.x, na.rm = T)))
#> `summarise()` has grouped output by 'location'. You can override using the `.groups` argument.
#> # A tibble: 4 x 3
#> # Groups:   location [2]
#>   location sub_location measurement
#>   <chr>    <chr>              <dbl>
#> 1 Area 1   A                   72.5
#> 2 Area 1   B                   76  
#> 3 Area 2   A                   47  
#> 4 Area 2   B                   90

^{Created on 2021-04-18 by the reprex package (v2.0.0)}

[cactusoxbird]

Oh wow, thanks for your help!