mutate() loses groups

[hadley]

Moved from tidyverse/dplyr#919, by kykuo

library(dplyr)
library(dtplyr)

iris %>%
  group_by(Species) %>% 
  mutate(x = cumsum(Sepal.Length)) %>% 
  groups

data.table::data.table(iris) %>% 
  group_by(Species) %>%
  mutate(x = cumsum(Sepal.Length)) %>%
  groups

[HuangNa2]

Seems like mutate reduces the group by 1 everytime its called

library(dplyr)
library(dtplyr)

iris %>%
  group_by(Species,Petal.Width,Sepal.Length) %>% 
  mutate(x = cumsum(Sepal.Length)) %>% 
  groups

Results in :

[[1]]
Species

[[2]]
Petal.Width

From a bit of testing it appears that it removes the last grouped element each time mutate is called.