Filter based on number of distinct values per grou

Question

I have the following df:

names    sex
adam      M
jill      F
stewart   M
jordan    M
alica     F
jordan    F

How do I filter the rows so that I only get the names that are both M and F, in this case, jordan.

Answer 1

If all your data is like this, you can simply find rows with duplicate values:

dat[duplicated(dat$names),]

Example:

> dat <- data.frame(names = c("adam", "jill", "stewart", "jordan", "alicia", "jordan"),
+                   sex = c("M", "F", "M", "M", "F", F)
+ )
> dat
    names   sex
1    adam     M
2    jill     F
3 stewart     M
4  jordan     M
5  alicia     F
6  jordan FALSE
> dat[duplicated(dat$names),]
   names   sex
6 jordan FALSE

or if you want a vector of names:

> as.character(dat[duplicated(dat$names),]$names)
[1] "jordan"

Answer 2

We can group by 'names' and filter the 'sex' having unique number of elements greater than 1

library(dplyr)
df %>% 
   group_by(names) %>%
   filter(n_distinct(sex) > 1)

---

Or another option is to group by 'names' and filter the groups having both the 'M' and 'F'

df %>%
  group_by(names) %>%
  filter(all(c("M", "F") %in% sex))