dev version of tabyl() fails on grouped data_frame

[sfirke]

it ends up running on the grouping variable, not on the intended variable:

> mtcars %>% group_by(am, cyl) %>% summarise(mean_mpg = mean(mpg)) %>% tabyl(cyl)
Adding missing grouping variables: `am`
  am n percent
1  0 3     0.5
2  1 3     0.5

[sfirke]

Look to understand why this happens and perhaps rewrite the code, but if something elegant is not possible then have tabyl() ungroup the input (and print a warning?).

[sfirke]

This is due to dplyr::select() not letting you unselect a grouping variable - so the fix is indeed to to ungroup(). Let's not print a warning, it's intuitive enough behavior even if you were expecting something grouped.