map_dfr(..., .id='colname') should return empty DF with one column when input is empty list

[meowcat]

Brief description of the problem

library(tidyyverse)
map_dfr(list(), ~.x, .id = "test")

This returns a 0x0 tibble. However, I rely downstream on a column "test" existing (this is the only column that's guaranteed to exist). I would expect the result to be a 0x1 tibble with a column called "test", since the result of map_dfr with lists of different-named element + .id is a dataframe with the union of all columns plus the ID column:

list(list(a=1, b=2), list(a=1, c=3)) %>% map_dfr(~.x)
list(list(a=1, b=2), list(a=1, c=3)) %>% map_dfr(~.x, .id = "col")

Therefore, with zero columns, this should be the union of all columns (none) plus the ID column.

[hadley]

Root cause is probably a vctrs issue. Reprex:

purrr::map_dfr(list(), ~.x, .id = "test")
#> # A tibble: 0 × 0
vctrs::vec_rbind(.names_to = "test")
#> data frame with 0 columns and 0 rows

^{Created on 2022-08-23 by the reprex package (v2.0.1)}

[hadley]

Now tracking in vctrs.