Converting nested tibble to character vector in R

Question

I have a nested dataframe tb which looks like that:

>tb
# A tibble: 26 x 3
   league         fdj_data          five38_data      
   <chr>          <list>            <list>           
 1 Ch.D1 Danemark <tibble [14 x 1]> <tibble [14 x 1]>
 2 Ch.D1 Ecosse   <tibble [10 x 1]> <tibble [14 x 1]>
 3 Ligue 2        <tibble [20 x 1]> <tibble [19 x 1]>
 4 Serie B        <tibble [18 x 1]> <tibble [21 x 1]>
 5 Liga Segunda   <tibble [20 x 1]> <tibble [20 x 1]>
 6 Ch.D1 Pays-Bas <tibble [18 x 1]> <tibble [21 x 1]>
 7 Ch.D1 Grèce    <tibble [12 x 1]> <tibble [16 x 1]>
 8 Ch.D1 Suède    <tibble [16 x 1]> <tibble [19 x 1]>
 9 Ch.D1 Turquie  <tibble [18 x 1]> <tibble [21 x 1]>
10 Ch.D1 Russie   <tibble [14 x 1]> <tibble [19 x 1]>
# ... with 16 more rows

I also have a function (let's call it func) which takes 2 character vectors (possibly of different lengths) and output another character vector (same length length as the first arg)

I want to add another column which contains, for each row, func(fdj_data, five38_data)

I've tried

tb %>% 
  mutate(new_var = func(fdj_data, five38_data))

and

tb %>% 
  mutate(fdj_data = as.character(fdj_data),
         five38_data = as.character(five38_data)) %>%
  mutate(new_var = func(fdj_data, five38_data))

But both don't work. I also tried with purrr::map() but I wasn't more successful

Have you got any idea ?

Likely a candidate for `mutate(newvar = purrr::map2(fdj_data, five38_data, func))` — r2evans, Aug 31 '18 at 21:14

score 0 · Answer 1 · answered Sep 15 '18 at 19:36

You should not have to experience any problem if your func is properly vectorized. Please see example below:

library(tibble)
library(dplyr)
set.seed(123)
func <- function(fdj_data, five38_data) {
  sample(LETTERS, length(fdj_data), replace = TRUE)
}


tb <- tibble(league = paste0(sample(letters, 26, replace = TRUE), 1:26))
tb$fdj_data <- lapply(sample(10:22, 26, replace = TRUE), function(x)tibble(rnorm(x)))
tb$five38_data <- lapply(sample(10:22, 26, replace = TRUE), function(x)tibble(rnorm(x)))
tb

tb %>% 
  mutate(new_var = func(fdj_data, five38_data))

Output:

# A tibble: 26 x 4
   league fdj_data          five38_data       new_var
   <chr>  <list>            <list>            <chr>  
 1 h1     <tibble [17 x 1]> <tibble [14 x 1]> C      
 2 u2     <tibble [17 x 1]> <tibble [19 x 1]> U      
 3 k3     <tibble [13 x 1]> <tibble [10 x 1]> F      
 4 w4     <tibble [11 x 1]> <tibble [20 x 1]> R      
 5 y5     <tibble [22 x 1]> <tibble [16 x 1]> L      
 6 b6     <tibble [21 x 1]> <tibble [19 x 1]> F      
 7 n7     <tibble [18 x 1]> <tibble [18 x 1]> Z      
 8 x8     <tibble [20 x 1]> <tibble [17 x 1]> L      
 9 o9     <tibble [10 x 1]> <tibble [19 x 1]> A      
10 l10    <tibble [16 x 1]> <tibble [18 x 1]> Y      
# ... with 16 more rows

Converting nested tibble to character vector in R

1 Answers1