improve factor handling

[rhijmans]

Hi, In sample_strat you have

    samples <- samples %>%
      dplyr::left_join(samples, sraster_cats, by = c("strata" = "value"))

Where

    sraster_cats <- cats(sraster) %>% 
      as.data.frame()

Which assumes that the name of the first column of sraster_cats is always "value". That is not correct. I think you should either use the actual column name, or, perhaps easiest, change it to "value" like this.

    sraster_cats <- cats(sraster) %>% 
      as.data.frame()
   colnames(sraster_cats)[1] <- "value"

I found this because "sgsR" triggered a reverse dependency error with "terra". I was able to get rid of that error for your example; but your assumption about the column name is still wrong.

[tgoodbody]

Hi @rhijmans

Thanks for letting me know about this and sorry for taking up your time.

Ive used your code suggestion and tested outputs with no errors.

Appreciate your time

T

  if(!is.null(terra::cats(sraster)[[1]])){
    message("'sraster' has factor values. Converting to allow mapping.")
    
    #--- change suggested by R Hijmans ---#
    sraster_cats <- cats(sraster) %>% 
      as.data.frame()
    colnames(sraster_cats)[1] <- "value"
    
  }