I wanted to create a group variable with specific conditions based on the values of GT. If GT is ‘I’, it should generate a different group number. If GT is ‘G’, the following rules should apply:

1. For the first four consecutive ‘G’s, assign the same group number, and for the next four consecutive ‘G’s, assign the next group number.
2. If there are five consecutive ‘G’s, assign the same group number to the first three ‘G’s and a different group number to the remaining two ‘G’s.

My Attempt:

library(tidyverse)
df1 <- 
  data.frame(
   GT = c(rep("G", 9), rep("I", 2), rep("G", 2), rep("I", 1))
  )

df2 <- 
  df1 %>%
  mutate(
      grp1 = case_when(
         GT == "I" ~ row_number()
      , .default = consecutive_id(GT)
      )
         ) %>%
  group_by(GT, grp1) %>%
  mutate(count = n()) %>%
  ungroup() %>% 
  mutate(
    grp2 = case_when(
        GT == "G" & count %/% 4 > 0  ~ row_number() %/% 5 + row_number(1)
      , .default = grp1
    )
  , grp3 = case_when(
        GT == "G"~ (cumsum(GT == "G") - 1) %/% 4 + 1
      , .default = grp1 
        )
      )

df2
#> # A tibble: 14 × 5
#>    GT     grp1 count  grp2  grp3
#>    <chr> <int> <int> <dbl> <dbl>
#>  1 G         1     9     1     1
#>  2 G         1     9     1     1
#>  3 G         1     9     1     1
#>  4 G         1     9     1     1
#>  5 G         1     9     2     2
#>  6 G         1     9     2     2
#>  7 G         1     9     2     2
#>  8 G         1     9     2     2
#>  9 G         1     9     2     3
#> 10 I        10     1    10    10
#> 11 I        11     1    11    11
#> 12 G         3     2     3     3
#> 13 G         3     2     3     3
#> 14 I        14     1    14    14

The group variable grp2 is very close to requirements. However, I want df2[8:9, 4] should have different group number than df2[5:7, 4]. Any hint please!