fct_collapse applying collapsed factors in wrong order when group_other = TRUE

[gtm19]

I'm having an issue which is self explanatory from the title. Reprex below. As you can see, the classification produced by fct_collapse is incorrect.

As requested, a simpler set of examples is below.

Example 1 - Collapsing Character Variables

The fct_collapse function wrongly classifies raspberry as a vegetable, rather than broccoli.

library(forcats)
library(dplyr)
library(tibble)

df <-
  tibble(item = c("apple", "grape", "banana", "broccoli", "raspberry"))

df %>% 
  mutate(category = fct_collapse(item,
                                 fruit = c("apple", "grape", "banana", "raspberry"),
                                 vegetables = "broccoli", group_other = TRUE))
#> # A tibble: 5 x 2
#>   item      category  
#>   <chr>     <fct>     
#> 1 apple     fruit     
#> 2 grape     fruit     
#> 3 banana    fruit     
#> 4 broccoli  fruit     
#> 5 raspberry vegetables

Details

Example 2 - Collapsing _Factor_Variables

For clarity, and contrary to my previous explanation, this bug affects both character and factor variables (since fct_collapse converts the .f argument to a factor if it is a character variable):

df <-
  tibble(item = factor(c("apple", "grape", "banana", "broccoli", "raspberry")))

df %>% 
  mutate(category = fct_collapse(item,
                                 fruit = c("apple", "grape", "banana", "raspberry"),
                                 vegetables = "broccoli", group_other = TRUE))
#> # A tibble: 5 x 2
#>   item      category  
#>   <fct>     <fct>     
#> 1 apple     fruit     
#> 2 grape     fruit     
#> 3 banana    fruit     
#> 4 broccoli  fruit     
#> 5 raspberry vegetables

Example 3 - with group_other argument set to FALSE

Just to demonstrate that this bug does not occur when group_other is set to FALSE:

df <-
  tibble(item = factor(c("apple", "grape", "banana", "broccoli", "raspberry")))

df %>% 
  mutate(category = fct_collapse(item,
                                 fruit = c("apple", "grape", "banana", "raspberry"),
                                 vegetables = "broccoli", group_other = FALSE))
#> # A tibble: 5 x 2
#>   item      category  
#>   <fct>     <fct>     
#> 1 apple     fruit     
#> 2 grape     fruit     
#> 3 banana    fruit     
#> 4 broccoli  vegetables
#> 5 raspberry fruit

Session Info

sessionInfo()
#> R version 3.5.2 (2018-12-20)
#> Platform: x86_64-w64-mingw32/x64 (64-bit)
#> Running under: Windows 10 x64 (build 17134)
#> 
#> Matrix products: default
#> 
#> locale:
#> [1] LC_COLLATE=English_United Kingdom.1252 
#> [2] LC_CTYPE=English_United Kingdom.1252   
#> [3] LC_MONETARY=English_United Kingdom.1252
#> [4] LC_NUMERIC=C                           
#> [5] LC_TIME=English_United Kingdom.1252    
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] bindrcpp_0.2.2     tibble_2.0.1       dplyr_0.7.8       
#> [4] forcats_0.4.0.9000
#> 
#> loaded via a namespace (and not attached):
#>  [1] Rcpp_1.0.0       knitr_1.21       bindr_0.1.1      magrittr_1.5    
#>  [5] tidyselect_0.2.5 R6_2.3.0         rlang_0.3.1      fansi_0.4.0     
#>  [9] stringr_1.4.0    highr_0.7        tools_3.5.2      xfun_0.4        
#> [13] utf8_1.1.4       cli_1.0.1        htmltools_0.3.6  yaml_2.2.0      
#> [17] digest_0.6.18    assertthat_0.2.0 crayon_1.3.4     purrr_0.3.0     
#> [21] glue_1.3.0       evaluate_0.12    rmarkdown_1.11   stringi_1.2.4   
#> [25] compiler_3.5.2   pillar_1.3.1     pkgconfig_2.0.2

^{Created on 2019-02-17 by the reprex package (v0.2.1)}

[hadley]

Would you mind first making a simpler example?

[gtm19]

Would you mind first making a simpler example?

No problem @hadley. All done.

[hadley]

Thanks, that's perfect!

[hadley]

@AmeliaMN do you want to have a look at this?

[gtm19]

@AmeliaMN I have added some further comments on the original pull request which might help also.

[avishaitsur]

The reprex still results in incorrect levels.
Is there a solution?

[gtm19]

I believe fct_collapse was dropped from 0.4.0, but my pull request #176 contains a fix (should it be resurrected).