Convert list of list into a tibble

Question

I have read this question, and I cannot make sense of it in my case, where I have ggplots in my tibble.

Say I have:

library(tidyverse)
f <- function(n) {
  p <- tibble(x = rnorm(30, n)) %>%
    ggplot(aes(x = x)) + geom_density()
  return(list(n = n, p = p))
}
m <- map(seq(10), f)

I would like to turn m into a tibble, with ten rows and two columns, named n and p. I also would like my code to work with any number of columns, any type of column, if possible.

as_tibble(transpose(m)) gives me

# A tibble: 10 x 2
   n         p     
   <list>    <list>
 1 <int [1]> <gg>  
 2 <int [1]> <gg>  
 3 <int [1]> <gg>  
 ...

i.e. each cell is a list, with one element. I would like to transform each list to a scalar.

What I have tried:

• I have used map_int for the column n, but what to do with column p?
• unnest also work for column n, but not with p.
• Same with unlist on the column p (even with recursive = FALSE).

Answer 1

Here is one possibility. You modify the function such that it only returns the plot. Then you create the tibble with only column n and create column p using map.

f2 <- function(n) {
  tibble(x = rnorm(30, n)) %>%
    ggplot(aes(x = x)) + geom_density()
}

tibble(n = 1:10,
       p = map(n, f2))

Then column n is an integer. p is a list but this is the best you can get since columns of tibbles cannot contain objects of type gg or ggplot but only lists containing these objects.

As far as I know the only possible column types are: int, dbl, date, dttm, factor, logical, character and list.