Sort 3 way tabyl factor levels (#287)

closes #250, three-way tabyls where 3rd variable is a factor are now sorted by factor level

DESCRIPTION

@@ -2,9 +2,9 @@ Package: janitor
 Title: Simple Tools for Examining and Cleaning Dirty Data
 Version: 1.1.1.9000
 Authors@R: c(person("Sam", "Firke", email = "[email protected]", role = c("aut", "cre")),
+    person("Bill", "Denney", email = "[email protected]", role = "ctb"),
     person("Chris", "Haid", email = "[email protected]", role = "ctb"),
     person("Ryan", "Knight", email = "[email protected]", role = "ctb"),
-    person("Bill", "Denney", email = "[email protected]", role = "ctb"),
     person("Malte", "Grosser", email = "[email protected]", role = "ctb"))
 Description: The main janitor functions can: perfectly format data.frame column
     names; provide quick counts of variable combinations (i.e., frequency

NEWS.md

@@ -18,12 +18,18 @@ Two new function `janitor::chisq.test()` and `janitor::fisher.test()` allow to a
 
 * `adorn_totals()` gains an argument `"name"` that allows the user to specify a value other than "Total" to appear as the name of the added row and/or column.  (#263)  Thanks to **@StephieLaPugh** for suggesting and **@daniel-barnett** for implementing.
 
+* `remove_empty()` now works with matrices (returning a matrix).  (#215)  Thanks to **@jsta** for reporting and **@billdenney** for patching.
+
+* If the third variable in a three-way tabyl is a factor, the resulting list is sorted in order of its levels (#250).  Empty factor levels in the 3rd variable are still omitted regardless of the value of `show_missing_levels`.
+
 ## Bug fixes
 
-* `remove_empty()` now works with matrices (returning a matrix).  (#215)  Thanks to **@jsta** for reporting and **@billdenney** for patching.
+
 * `excel_numeric_to_date()` no longer gives an overflow error for integer input (for dates since 1968).  (#241)  Thanks to **@hideaki** for reporting and **@billdenney** for patching.
+
 * `clean_names()` and `make_clean_names()` now support 'none' as a case option. (#269) Thanks to **@andrewbarros** for reporting and patching.  
 
+
 # janitor 1.1.1 (2018-07-30)
 
 ## Release summary
@@ -34,6 +40,7 @@ Patches a bug introduced in version 1.1.0 where `excel_numeric_to_date()` would
 
 * `excel_numeric_to_date()` again handles `NA` correctly, in version 1.1.0 the function would error if any values of the input vector were `NA`. (#220). Thanks **@emilelatour** for reporting and **@billdenney** for patching.
 
+
 # janitor 1.1.0 (2018-07-17)
 
 ## Release summary

R/tabyl.R

@@ -227,6 +227,11 @@ tabyl_2way <- function(dat, var1, var2, show_na = TRUE, show_missing_levels = TR
 # a list of two-way frequency tables, split into a list on a third variable
 tabyl_3way <- function(dat, var1, var2, var3, show_na = TRUE, show_missing_levels = TRUE) {
   dat <- dplyr::select(dat, !! var1, !! var2, !! var3)
+
+  # Keep factor levels for ordering the list at the end
+  if(is.factor(dat[[3]])){
+    third_levels_for_sorting <- levels(dat[[3]])
+  }
   dat[[3]] <- as.character(dat[[3]]) # don't want empty factor levels in the result list - they would be empty data.frames
 
   # grab class of 1st variable to restore it later
@@ -240,6 +245,9 @@ tabyl_3way <- function(dat, var1, var2, var3, show_na = TRUE, show_missing_level
   if (show_na && sum(is.na(dat[[3]])) > 0) {
     dat[[3]] <- factor(dat[[3]], levels = c(sort(unique(dat[[3]])), "NA_"))
     dat[[3]][is.na(dat[[3]])] <- "NA_"
+    if(exists("third_levels_for_sorting")){
+      third_levels_for_sorting <- c(third_levels_for_sorting, "NA_")
+    }
   }
 
   if (!show_missing_levels) { # this shows missing factor levels, to make the crosstabs consistent across each data.frame in the list based on values of var3
@@ -258,6 +266,11 @@ tabyl_3way <- function(dat, var1, var2, var3, show_na = TRUE, show_missing_level
     purrr::map(tabyl_2way, var1, var2, show_na = show_na, show_missing_levels = show_missing_levels) %>%
     purrr::map(reset_1st_col_status, col1_class, col1_levels) # reset class of var in 1st col to its input class, #168
 
+  # reorder when var 3 is a factor, per #250
+  if(exists("third_levels_for_sorting")){
+    result <- result[order(third_levels_for_sorting[third_levels_for_sorting %in% unique(dat[[3]])])] 
+  }
+
   result
 }