Suggestion: get_dupes results should be sorted by dupe_count descending

[sfirke]

Feature requests

Sort the results of get_dupes by dupe_count.

Remarks

Right now it sorts alphabetically by the grouping variable. But usually I then add %>% arrange(desc(dupe_count)) to start with the most-duplicated combination. I suppose this would potentially be a breaking change for some use cases. But I think it is ultimately more useful than the alphabetical sort.

Example

mtcars %>% get_dupes(cyl)

In my result, the group with cyl = 8 that has 14 records would appear first.

This would be easy to implement. Thoughts from others?

[billdenney]

It makes sense to me. When I use get_dupes(), the goal is either to confirm that there are no duplicates or to manually inspect them. Neither of my use cases is broken with your suggested change, and the manual inspection is improved.

My one addition would be to break a tie by sorting alphabetically second so that there is a consistent output order.

[jzadra]

Yes I think this is a good enhancement.