Filter column by group based on another column (R dplyr)-CodePudding

I want to group a column and filter based on occurrence of certain values in another column. In my case, I have an ID variable and want to keep only instances where Time = 1 and = 2. This is a larger dataset so I'm not able to simply keep duplicates.

I tried using group_by and filter, but it has not produced the desired table.

df <- tibble(s1 = c(1001, 1001, 1002, 1002, 1003, 1004), s2 = c(1, 2, 1, 2, 2, 1))
df_filter <- df %>% 
  group_by(s1) %>% 
  filter(all(c(1, 2)) %in% s2)

Input:

ID	Time
1001	1
1001	2
1002	1
1002	2
1003	2
1004	1

Desired Output:

s1	s2
1001	1
1001	2
1002	1
1002	2

CodePudding user response：

You can use n() from dplyr library. It will give you the count of rows. Then you can compare it with 1.

Here's how you can do it:

df_filter <- df %>% 
  group_by(s1) %>%
  filter(n() > 1)

And if you want only 1 or 2 in s2, then you can extend your filter like this:

df_filter <- df %>% 
  group_by(s1) %>%
  filter(s2 == 1 | s2 == 2) %>%
  filter(n() > 1)