How to group and ungroup in R?

Question

I have a dataframe like as shown below

test_df <- data.frame("SN" = c("ABC123","ABC123","ABC123","MNO098","MNO098","MNO098"), 
                      "code" = c("ABC1111","DEF222","GHI133","","MNO1123","MNO567"), 
                      "d_time" = c("2220-08-27","2220-05-27","2220-02-27","2220-11-27","2220-02-27",""))

I am trying to do 2 things

1) create 2 new columns (p_id,v_id) by stripping alphabets from columns SN and code and retain only 9 digits

2) create a lag column (p_vid) based on v_id for each person sorted based on his/her d_time

t_df <- test_df %>% group_by(SN)
t_df %>% arrange((d_time), .by_group = TRUE) ->> sorted_df #sorted based on d_time
transform_ids = function(DF){  # this function is to create person and visit_occurrence ids
  DF %>% 
    mutate(p_id = as.integer(str_remove_all(.$SN,"[a-z]|[A-Z]") %>%   #retaining only the numeric part
                                    str_sub(1,9))) %>%
    mutate(v_id = as.integer(str_remove_all(.$code,"[a-z]|[A-Z]") %>% 
                                              str_sub(1,9))) %>%
    group_by(p_id) %>%
    mutate(pre_vid = lag(v_id)) %>%
    ungroup
}
transform_ids(sorted_df)

But when I do this I encounter the below error

Error in View : Column p_id must be length 3 (the group size) or one, not 6 Error: Column p_id must be length 3 (the group size) or one, not 6 In addition: Warning message: In view(transform_ids(t_df)) : Show Traceback Rerun with Debug Error: Column p_id must be length 3 (the group size) or one, not 6

I expect my output to be like as shown below. Basically I am trying to link each v_id of a person to his previous visit which is p_vid

Answer 1

To generate the p_id and v_id columns, just use sub:

t_df$p_id <- gsub("[A-Z]+", "", t_df$SN)
t_df$v_id <- gsub("[A-Z]+", "", t_df$code)

For the p_vid column, use lag() from the dplyr package:

t_df %>%
group_by(p_id) %>%
mutate(p_vid = lag(v_id, order_by=d_time, default=0))

The output from the above actually gives you a tibble. If you want a data frame, just use:

t_df <- as.data.frame(t_df)

Output:

  SN     code    d_time     p_id  v_id  p_vid
  <fct>  <fct>   <fct>      <chr> <chr> <chr>
1 ABC123 ABC1111 2220-08-27 123   1111  222  
2 ABC123 DEF222  2220-05-27 123   222   133  
3 ABC123 GHI133  2220-02-27 123   133   0    
4 MNO098 ""      2220-11-27 098   ""    1123 
5 MNO098 MNO1123 2220-02-27 098   1123  567  
6 MNO098 MNO567  ""         098   567   0