Question

I created a simple data.frame:

data.frame(a = rep(LETTERS[1:4], each=4),
           b = c(sample(6,4),sample(6,4),sample(6,4),sample(6,4)))

   a b
1  A 6
2  A 4
3  A 2
4  A 3
5  B 5
6  B 1
7  B 3
8  B 6
9  C 2
10 C 3
11 C 5
12 C 1
13 D 4
14 D 5
15 D 1
16 D 3

How can I keep only those rows of the data.frame where the number in column b appears in all 4 letters of column a? So, for instance, the number 3 in column b appears for A, B, C and D in column a and should therefore be kept.

Answer 1

Here is a slightly different approach making use of recursive intersection.

set.seed(123)

df <- data.frame(a = rep(LETTERS[1:4], each=4),
                 b = c(sample(6,4),sample(6,4),sample(6,4),sample(6,4)))
with(df, df[b %in% Reduce(intersect, split(b, a)),])
   a b
3  A 6
4  A 3
5  B 6
7  B 3
10 C 3
11 C 6
14 D 3
16 D 6