[Bug] ordered factors in tibbles can't be imputed

[mayer79]

The following code raises an error:

library(missRanger)
library(ggplot2)

imp <- missRanger(diamonds |> generateNA(seed = 1))

# Error
in `[<-`:
! Assigned data `pred` must be compatible with existing data.
ℹ Error occurred for column `cut`.
Caused by error in `vec_assign()`:
! Can't convert <factor<ff2b9>> to <ordered<ff2b9>>.
Run `rlang::last_trace()` to see where the error occurred.

The problem does not occur when:

• diamonds is converted to a normal data.frame
• Predictive mean matching is used via pmm.k > 0
• The ordered factor is turned into an unordered factor

Tibbles seem to be picky with ordered factors in subset-assignment.