Should as_tibble.tbl_df use `validate`?

[jennybc]

It is possible to mutilate a tibble and pass it through as_tibble(), even with validate = TRUE, without having the problems called out. Should validate be used even in this case?

library(tibble)

df <- tibble(
  x = 1:3,
  y = 4:6
)
names(df) <- ""
as_tibble(df)
#> # A tibble: 3 x 2
#>      ``  `NA`
#>   <int> <int>
#> 1     1     4
#> 2     2     5
#> 3     3     6
as_tibble(df, validate = TRUE)
#> # A tibble: 3 x 2
#>      ``  `NA`
#>   <int> <int>
#> 1     1     4
#> 2     2     5
#> 3     3     6
tibble:::check_tibble(df)
#> Error: Columns 1, 2 must be named

How this came up for me: I'm creating an S3 class that extends tibble. After some operations, my object should no longer bear that class, but should revert to being a plain tibble. Instead of only hand-editing the class, it seems like a good idea to also run it through as_tibble(). But then I found that as_tibble() just passes tibbles straight through, without any validation.

[krlmlr]

Thanks for bringing this up.

[github-actions]

This old thread has been automatically locked. If you think you have found something related to this, please open a new issue and link to this old issue if necessary.