How to make an index column based on already existing column having a specific pattern in r?

I have a column name set, in a dataframe df which looks like

df <- data.frame(set <- c("","","","","","set","","","","","set","","","","","set"))

now I want a column set_sequence based on pattern from column set which should look like:

df <- data.frame(set <- c("","","","","","set","","","","","set","","","","","set"),
                 set_seq <- c("","","","","","1","","","","","2","","","","","3"))

Can anyone help me how can I do that in r I tried cumsum function from data.table but didn't help. Also I have empty cells not NA cells

Solution

I'm not sure about exact condition, but just for data you provided,

library(tidyverse)
dat <-  data.frame(set = c("","","","","","set","","","","","set","","","","","set"))
dat %>%
  group_by(set) %>%
  mutate(set_seq = ifelse(set == "",  "", as.character(1:n())))

   set   set_seq
   <chr> <chr>  
 1 ""    ""     
 2 ""    ""     
 3 ""    ""     
 4 ""    ""     
 5 ""    ""     
 6 "set" "1"    
 7 ""    ""     
 8 ""    ""     
 9 ""    ""     
10 ""    ""     
11 "set" "2"    
12 ""    ""     
13 ""    ""     
14 ""    ""     
15 ""    ""     
16 "set" "3"

if you see this result as data.frame,

   set set_seq
1             
2             
3             
4             
5             
6  set       1
7             
8             
9             
10            
11 set       2
12            
13            
14            
15            
16 set       3