I have a tibble in which one column is a list of tibbles. Some tibbles in this list have more than one row. Here is a minimal example:
library(tibble)
tmp <- tribble(
~x, ~y,
1, tibble(a=1, b=2),
2, tibble(a=4:5, b=6:7))
which produces
> tmp
# A tibble: 2 x 2
x y
<dbl> <list>
1 1 <tibble [1 x 2]>
2 2 <tibble [2 x 2]>
I want to create a tibble that has the same number of rows as tmp and
that expands the list-column into as many columns as necessary. In the
example, I want
x a1 a2 b1 b2
<int> <dbl> <dbl> <dbl> <dbl>
1 1 1 NA 2 NA
2 2 4 5 6 7
I had thought to do this with some combination of unnest() and pivot_wider(), or perhaps flatten(), but I cannot quite get it to work. What is the best way to do this?
In this example, the tibbles in the list-column never have more than two rows. But in my real list-columns, some tibbles have one row, some have two, and some have more than two. So I am looking for a solution that works when the tibbles in the list-column have any arbitrary number of rows.
Here's one solution using unnest_wider
library(tidyr)
unnest_wider(tmp, y) %>%
unnest_wider(a, names_repair = ~gsub('...', 'a', .)) %>%
unnest_wider(b, names_repair = ~gsub('...', 'b', .))
New names:
* `` -> ...1
...
New names:
* `` -> ...1
* `` -> ...2
# A tibble: 2 x 5
x a1 a2 b1 b2
<dbl> <dbl> <int> <dbl> <int>
1 1 1 NA 2 NA
2 2 4 5 6 7