Add `sort` argument to `fct_lump_` functions?

[DesiQuintans]

The fct_lump_ functions are all about deciding which levels to keep or lump together based on their frequency, so I think that they should also have an option (e.g. sort = c("no", "asc", "desc")) to return those levels in an order that is based on their frequency.

Edit: Or perhaps, add an example to the man page for these functions showing that you can use fct_infreq() and its friends to do the ordering. However, doing this requires a fct_relevel() at the end to ensure "Other" is the last value.

Thank you for a great package! <3

library(forcats)

set.seed(12345)

repeated_states <- rep.int(x = state.name, times = runif(n = length(state.name), min = 1, max = 300))

sort(table(repeated_states), decreasing = TRUE)
#> repeated_states
#>        Georgia      Minnesota       Maryland          Texas   Pennsylvania 
#>            296            289            285            278            271 
#>       Arkansas         Alaska         Oregon     New Mexico   South Dakota 
#>            265            262            260            238            234 
#>           Utah        Arizona       Illinois        Florida        Alabama 
#>            232            228            220            218            216 
#>    Mississippi       Nebraska   North Dakota       Missouri        Wyoming 
#>            212            209            204            193            188 
#>   Rhode Island         Nevada       Delaware     New Jersey         Kansas 
#>            185            163            153            145            139 
#>     California  Massachusetts      Tennessee      Louisiana           Iowa 
#>            137            136            129            121            117 
#>       Kentucky        Montana           Ohio       Oklahoma    Connecticut 
#>            117            117            111            109             98 
#>       Michigan       Virginia        Vermont  New Hampshire North Carolina 
#>             98             97             78             68             57 
#>          Maine       Colorado          Idaho South Carolina     Washington 
#>             54             50             46             41             18 
#>      Wisconsin  West Virginia         Hawaii       New York        Indiana 
#>             17             13             11              2              1

as_fct <- fct_lump_n(repeated_states, 10)

levels(as_fct)
#>  [1] "Alaska"       "Arkansas"     "Georgia"      "Maryland"     "Minnesota"   
#>  [6] "New Mexico"   "Oregon"       "Pennsylvania" "South Dakota" "Texas"       
#> [11] "Other"

as_ordered_fct <- fct_lump_n(repeated_states, 10) |> fct_infreq() |> fct_relevel("Other", after = Inf)

levels(as_ordered_fct)
#>  [1] "Georgia"      "Minnesota"    "Maryland"     "Texas"        "Pennsylvania"
#>  [6] "Arkansas"     "Alaska"       "Oregon"       "New Mexico"   "South Dakota"
#> [11] "Other"

^{Created on 2024-12-17 with reprex v2.1.1}