Add conditional counter in new column R

Question

I have the following data:

structure(list(id = c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L), 
form = c("test", "test", "missing", "test", "test", "missing", 
"test", "test", "test", "missing")), row.names = c(NA, 10L), class = "data.frame")

id	form
1	test
1	test
1	missing
1	test
2	test
2	missing
2	test
3	test
3	test
3	missing

I need to add a column ("form_completed") that starts counting from 1 by "id", but skips the counting for a certain value under "form" which is a string variable. The output should be:

structure(list(id = c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 3L, 3L, 3L), 
form = c("test", "test", "missing", "test", "test", "missing", 
"test", "test", "test", "missing"), form_completed = c(1L, 
2L, NA, 3L, 1L, NA, 2L, 1L, 2L, NA)), row.names = c(NA, 10L), class = "data.frame")

id	form	form_completed
1	test	1
1	test	2
1	missing	NA
1	test	3
2	test	1
2	missing	NA
2	test	2
3	test	1
3	test	2
3	missing	NA

It seems pretty straightfoward, and I have tried different things with mutate() and row_count() in R, but can't find it to work. Any help would be much appreciated!

Answer 1

dplyr

You can use case_when + cumsum:

library(dplyr)
df %>% 
  group_by(id) %>% 
  mutate(a = case_when(form == 'test' ~ cumsum(form == "test")))

      id form        a
   <int> <chr>   <int>
 1     1 test        1
 2     1 test        2
 3     1 missing    NA
 4     1 test        3
 5     2 test        1
 6     2 missing    NA
 7     2 test        2
 8     3 test        1
 9     3 test        2
10     3 missing    NA

Or with ifelse:

df %>% 
  group_by(id) %>% 
  mutate(a = ifelse(form == 'test', cumsum(form == "test"), NA))

---

base R

df$a[df$form == "test"] <- with(df[df$form == "test", ], ave(form, id, FUN = seq_along))

   id    form    a
1   1    test    1
2   1    test    2
3   1 missing <NA>
4   1    test    3
5   2    test    1
6   2 missing <NA>
7   2    test    2
8   3    test    1
9   3    test    2
10  3 missing <NA>