split_cols_by() does wrong ordering if labels_var is a factor and not a character

[danielinteractive]

In the below example we have a labels variable (which does not have the same alphabetic ordering as the original arm variable). This gives the wrong labeling of the columns when it is a factor. Only when it is character it works as expected. I would guess that fixing this should be simply putting as.character() in the right place?

library(rtables)
library(dplyr)

ex_adsl <- ex_adsl %>%
  mutate(
    ARMLAB = factor(case_when( # need to omit factor to solve problem below
      ARM == "A: Drug X" ~ "Drug X",
      ARM == "B: Placebo" ~ "Placebo",
      ARM == "C: Combination" ~ "Combination")
  ))

lyt_orig <- basic_table() %>%
  split_cols_by("ARM") %>%
  analyze(c("AGE", "BMRKR2"))
tbl_orig <- build_table(lyt_orig, ex_adsl)

lyt_lab <- basic_table() %>%
  split_cols_by("ARM", labels_var = "ARMLAB") %>%
  analyze(c("AGE", "BMRKR2"))
tbl_lab <- build_table(lyt_lab, ex_adsl)

tbl_orig
tbl_lab # wrong labeling here

Session info:

R version 3.6.3 (2020-02-29)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Debian GNU/Linux 9 (stretch)

Matrix products: default
BLAS/LAPACK: /usr/lib/libopenblasp-r0.2.19.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8       
 [4] LC_COLLATE=en_US.UTF-8     LC_MONETARY=en_US.UTF-8    LC_MESSAGES=C             
 [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                  LC_ADDRESS=C              
[10] LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] dplyr_1.0.0   rtables_0.3.6 magrittr_1.5 

loaded via a namespace (and not attached):
 [1] fansi_0.4.1       assertthat_0.2.1  packrat_0.5.0     digest_0.6.25     crayon_1.3.4     
 [6] R6_2.4.1          lifecycle_0.2.0   pillar_1.4.3      cli_2.2.0         rlang_0.4.9      
[11] rstudioapi_0.11   vctrs_0.3.1       generics_0.0.2    ellipsis_0.3.0    tools_3.6.3      
[16] glue_1.4.0        purrr_0.3.4       compiler_3.6.3    pkgconfig_2.0.3   htmltools_0.5.1.1
[21] tidyselect_1.1.0  tibble_3.0.1 

[danielinteractive]

@anajens fyi

[gmbecker]

Typo in the commit message but this is fixed in 39c655e and pushed as a narrow bugfix to devel. The regression test will be added to the branch I'm currently working on for now (mostly cause I forgot to add it to the commit that got pushed directly to devel).