Respect requested variable order in distinct() (#6346)

Fixes #6156
tidyverse · Jul 22, 2022 · 444c139 · 444c139
1 parent d7b9ae6
commit 444c139
Show file tree

Hide file tree

Showing 4 changed files with 19 additions and 9 deletions.
diff --git a/NEWS.md b/NEWS.md
@@ -1,5 +1,8 @@
 # dplyr (development version)
 
+* `distinct()` returns columns ordered the way you request, not the same
+  as the input data (#6156).
+
 * The `.keep`, `.before`, and `.after` arguments to `mutate()` 
   are no longer experimental.
 

diff --git a/R/distinct.R b/R/distinct.R
@@ -106,16 +106,18 @@ distinct_prepare <- function(.data,
     abort(bullets, call = error_call)
   }
 
-  # Always include grouping variables preserving input order
-  out_vars <- intersect(names(.data), c(distinct_vars, group_vars))
+  # Only keep unique vars
+  distinct_vars <- unique(distinct_vars)
+  # Missing grouping variables are added to the front
+  new_vars <- c(setdiff(group_vars, distinct_vars), distinct_vars)
 
   if (.keep_all) {
     keep <- seq_along(.data)
   } else {
-    keep <- out_vars
+    keep <- new_vars
   }
 
-  list(data = .data, vars = out_vars, keep = keep)
+  list(data = .data, vars = new_vars, keep = keep)
 }
 
 #' @export

diff --git a/tests/testthat/_snaps/distinct.md b/tests/testthat/_snaps/distinct.md
@@ -1,4 +1,4 @@
-# distinct gives a warning when selecting an unknown column (#3140)
+# distinct errors when selecting an unknown column (#3140)
 
     Code
       df <- tibble(g = c(1, 2), x = c(1, 2))

diff --git a/tests/testthat/test-distinct.R b/tests/testthat/test-distinct.R
@@ -102,9 +102,15 @@ test_that("distinct handles 0 columns edge case (#2954)", {
   expect_equal(nrow(distinct(tibble())), 0L)
 })
 
-test_that("distinct preserves order of the input variables (#3195)",{
+test_that("distinct respects order of the specified variables (#3195, #6156)",{
   d <- data.frame(x = 1:2, y = 3:4)
-  expect_equal(names(distinct(d, y, x)), c("x", "y"))
+  expect_named(distinct(d, y, x), c("y", "x"))
+})
+
+test_that("distinct adds grouping variables to front if missing",{
+  d <- data.frame(x = 1:2, y = 3:4)
+  expect_named(distinct(group_by(d, y), x), c("y", "x"))
+  expect_named(distinct(group_by(d, y), x, y), c("x", "y"))
 })
 
 test_that("distinct() understands both NA variants (#4516)", {
@@ -169,8 +175,7 @@ test_that("distinct() preserves attributes on bare data frames (#6318)", {
 
 # Errors ------------------------------------------------------------------
 
-
-test_that("distinct gives a warning when selecting an unknown column (#3140)", {
+test_that("distinct errors when selecting an unknown column (#3140)", {
   expect_snapshot({
     df <- tibble(g = c(1, 2), x = c(1, 2))