bug using tidyverse tibbles

[JorritMontijn]

Hi! Thank you for building a great package. I'm not sure if anyone is still bugfixing issues, but even if not, then I hope the following will be useful for other users.

I am unsure what the exact source of this bug is, but I suspect it's a combination of an updated R version and using tibbles as data frames. When I run disparate_impact_remover(), I get the following error:

Error in `x[order(x, na.last = na.last, decreasing = decreasing)]`:
! Can't subset columns past the end.
ℹ Locations 383, 410, 435, …, 3673, and 4798 don't exist.
ℹ There is only 1 column.
Run `rlang::last_trace()` to see where the error occurred.
Warning message:
In xtfrm.data.frame(x) : cannot xtfrm data frames

I've fixed this by turning the tibble's column into an array using the following subfunction:

get_group_data = function(data,protected,subgroup,feature) {
  #get subdata
  Y <- data[protected == subgroup, feature]
  
  #turn tibble/list into array
  return(Y[[feature]])
}

and replacing all lines where the to-be-repaired column was extracted from
Y <- data[protected == subgroup, feature]
into
Y <- get_group_data(data,protected,subgroup,feature)

I'm sure this bug is also present in other functions, but so far I've only used the disparate_impact_remover(). The fix is fairly simple, so I hope you have the time to update the package.

Thank you!

[jakwisn]

@JorritMontijn Thanks for bringing it up. Would you be willing to create a PR with a solution?