n_distinct gives unexpected result for identical lists

[karldw]

I noticed that n_distinct(mylist) gives a different value than length(unique(mylist)) when mylist is a list of lists. This might be a bug?

I think this is a duplicate of #3699 and therefore also #2355, but those discussions are closed and marked as fixed.

library(dplyr)
(mylist <- c(list(1L), list(1L)))
#> [[1]]
#> [1] 1
#> 
#> [[2]]
#> [1] 1

identical(mylist[1], mylist[2])
#> [1] TRUE
length(unique(mylist))
#> [1] 1
n_distinct(mylist)
#> [1] 2

# The behavior is the same if we're using a list column in a tibble.
tibble(l = mylist) %>% mutate(n = n_distinct(l))
#> # A tibble: 2 x 2
#>   l             n
#>   <list>    <int>
#> 1 <int [1]>     2
#> 2 <int [1]>     2

Using dplyr v0.8.5, rlang 0.4.5, and vctrs 0.2.4.

[romainfrancois]

This appears to be fixed in the development version, soon to be released as 1.0.0

library(dplyr, warn.conflicts = FALSE)
(mylist <- c(list(1L), list(1L)))
#> [[1]]
#> [1] 1
#> 
#> [[2]]
#> [1] 1
identical(mylist[1], mylist[2])
#> [1] TRUE
length(unique(mylist))
#> [1] 1
n_distinct(mylist)
#> [1] 1

# The behavior is the same if we're using a list column in a tibble.
tibble(l = mylist) %>% 
  mutate(n = n_distinct(l))
#> # A tibble: 2 x 2
#>   l             n
#>   <list>    <int>
#> 1 <int [1]>     1
#> 2 <int [1]>     1

^{Created on 2020-04-16 by the reprex package (v0.3.0)}