hey can anyone explain `.groups = 'drop'`

[Rishant]

.groups = 'drop' , its use / meaning

[mara]

Check out the article here (specifically the summarise() section of Verbs)

Grouped data

Excerpting from the above article (though, again, it's worth just reading the relevant section in full):

Since version 1.0.0 the groups may also be kept (.groups = "keep") or dropped (.groups = "drop").

by_sex_gender %>% 
  summarise(n = n(), .groups = "keep") %>% 
  group_vars()
#> [1] "sex"    "gender"

by_sex_gender %>% 
  summarise(n = n(), .groups = "drop") %>% 
  group_vars()
#> character(0)

When the output no longer have grouping variables, it becomes ungrouped (i.e. a regular tibble).

[Rishant]

Thank you so much. Appreciate it alot.

[Andrzej]

Hi,
A few links with interesting comments, in the first you can find and I quote:
""bc Hadley wants it that way".
https://stackoverflow.com/questions/48489219/why-does-summarize-drop-a-group

and more detailed explanation below:

Difference between .groups argument and ungroup() in dplyr?

and here as well:

Tidyverse groups function for summarize?

best wishes,
Andrzej

[mikecrobp]

I think I read, when dplyr 1.0 was released, that Hadley thought the default should have been drop in the first place. But they made the wrong call before.

[Rishant]

does it have any practical meaning(have u faced any problems without it) , or is it just high level stuff and don't need to care?

[mikecrobp]

I have issues with it when I try to use the summarised df and try to select a subset of columns. I then get a message saying that the grouped fields have been added back rather than just the subset.