Renaming in `group_by()` can lead to grouping by nonexistent columns

[markfairbanks]

pacman::p_load(tidytable)

df <- tidytable(x = 1:3, y = c("a", "a", "b"))

df %>%
  group_by(new_y = y)
#> # A tidytable: 3 × 2
#> # Groups:      new_y
#>       x y    
#>   <int> <chr>
#> 1     1 a    
#> 2     2 a    
#> 3     3 b