Common elements among multiple groups

Asked Nov 19 '17 at 00:55

Active Nov 19 '17 at 00:55

Viewed 766 times

Suppose we have a data frame with a grouping column and a values column. E.g.,

test <- data.frame(Group = rep(c("A", "B", "C"), 3), Value = c(1, 10, 2, 4, 4, 4, 6, 6, 6))

What's the best way to get a vector of unique values which appear in all group, i.e., values common to all groups?

Here's a clumsy way (using dplyr):

test %>% 
 group_by(Value) %>% 
 mutate(In_groups = length(unique(Group))) %>% 
 ungroup() %>% 
 filter(In_groups == length(unique(Group))) %>% 
 .$Value %>% 
 unique()

It seems like there should be a better way to do this.

asked Nov 19 '17 at 00:55

Victor Kostyuk

`Reduce(intersect, split(test$Value, test$Group))` – Ronak Shah Nov 19 '17 at 01:00
@Ronak Shah: nice, but it doesn't preserve class (try using a date for Value, for example). – Victor Kostyuk Nov 19 '17 at 01:08
Somewhat simpler: `test %>% distinct %>% count(Value) %>% filter(n == n_distinct(test$Group))`... Or with data.table `setDT(test) %>% split(by="Group", keep.by=FALSE) %>% Reduce(f = fintersect)` – Frank Nov 19 '17 at 01:57
fwiw: you can substitute `n_distinct(Group))` for `length(unique(Group)))` – leerssej Nov 19 '17 at 02:18

Common elements among multiple groups

0 Answers0