left_join on a nested list - r

Question

I have a nested df x and an unnested df y.
How can I join these two together so that the final output is a a single row with the id and val columns from x and a new column for each of the respective num values in order of appearance, num_1, num_2 ...?
library(tidyverse)
x <- tibble(id = list(letters[1:6]), val = 13)
x
#> # A tibble: 1 x 2
#> id val
#> <list> <dbl>
#> 1 <chr [6]> 13
y <- tibble(id = letters[1:6], num = rnorm(6))
y
#> # A tibble: 6 x 2
#> id num
#> <chr> <dbl>
#> 1 a 0.532
#> 2 b -0.106
#> 3 c -0.105
#> 4 d 0.973
#> 5 e -0.825
#> 6 f -0.951
map2(x, y, left_join, by = 'id')
Error in UseMethod("left_join"): no applicable method for 'left_join' applied to an object of class "list"
Created on 2020-08-14 by the reprex package (v0.3.0)
Edit: I'm looking for something loosely like this while still maintaining the ID column.
x %>%
unnest(id) %>%
left_join(y) %>%
mutate(n = row_number()) %>%
pivot_wider(id_cols = -id,
values_from = num,
names_from = n)
#> Joining, by = "id"
#> # A tibble: 1 x 7
#> val `1` `2` `3` `4` `5` `6`
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 13 1.28 -0.387 -0.438 -0.0826 0.164 -1.24

Answer 1

Continuing with your work, you can try the following.
library(dplyr)
library(tidyr)
x %>%
unnest(id) %>%
left_join(y, by = "id") %>%
mutate(name = row_number(), id = list(id)) %>%
pivot_wider(values_from = num, names_glue = "num_{name}")
# # A tibble: 1 x 8
# id val num_1 num_2 num_3 num_4 num_5 num_6
# <list> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1 <chr [6]> 13 1.28 -0.387 -0.438 -0.0826 0.164 -1.24
or
x %>%
mutate(num = map(id, ~ tibble::deframe(y) %>% .[match(names(.), .x)] %>% unname)) %>%
unnest_wider(num, names_sep = "_")
# # A tibble: 1 x 8
# id val num_1 num_2 num_3 num_4 num_5 num_6
# <list> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
# 1 <chr [6]> 13 1.28 -0.387 -0.438 -0.0826 0.164 -1.24
Description of the second solution
deframe() in tibble transforms a two-column data.frame to a named vector, the first column is converted to vector names and the second one is converted to vector values. deframe(y) %>% .[match(names(.), .x)] is equivalent to deframe(y)[match(names(deframe(y)), .x)]. The deframe(y) part appears twice, so I move it to the front of a pipe and use the . symbol to represent it behind the pipe. This line is to match the position of id columns of both data and reorder num column of y.

Answer 2

Based on your y you're not going to have multiple columns but adjusting the example a little, is this what you were aiming for?
x <- tibble(id = list(letters[1:6]), val = 13)
y <- tibble(id = rep(letters[1:6],2), num = rnorm(12),
name = paste0("num_", rep(1:2, each = 6)))
map_dfr(x$id[[1]], ~tibble(id = .x, val = x$val)) %>%
left_join(
pivot_wider(y, names_from = name, values_from = num)
)
#> Joining, by = "id"
#> # A tibble: 6 x 4
#> id val num_1 num_2
#> <chr> <dbl> <dbl> <dbl>
#> 1 a 13 0.609 1.97
#> 2 b 13 0.956 -1.84
#> 3 c 13 0.425 0.297
#> 4 d 13 0.0379 -0.784
#> 5 e 13 -0.532 -0.769
#> 6 f 13 0.538 -1.10