Fix vec_order_info() when there are zero columns (#1866)

* Push a group size in the 0 column early exit case

* NEWS bullet

Author: DavisVaughan

NEWS.md

@@ -1,5 +1,7 @@
 # vctrs (development version)
 
+* Fixed an issue with `vec_rank()` and 0-column data frames (#1863).
+
 # vctrs 0.6.3
 
 * Fixed an issue where certain ALTREP row names were being materialized when
@@ -56,7 +58,7 @@
   `multiple = "warning"`, which have been removed from the documentation and
   silently soft-deprecated. Official deprecation for those options will start in
   a future release (#1791).
-  
+
 * `vec_locate_matches()` has changed its default `needles_arg` and
   `haystack_arg` values from `""` to `"needles"` and `"haystack"`, respectively.
   This generally generates more informative error messages (#1792).
@@ -70,7 +72,7 @@
 
 * The `numeric_version` type from base R is now better supported in equality,
   comparison, and order based operations (tidyverse/dplyr#6680).
-  
+
 * R >=3.5.0 is now explicitly required. This is in line with the tidyverse
   policy of supporting the [5 most recent versions of
   R](https://www.tidyverse.org/blog/2019/04/r-version-support/).
@@ -79,7 +81,7 @@
 
 * New `vec_expand_grid()`, which is a lower level helper that is similar to
   `tidyr::expand_grid()` (#1325).
-  
+
 * New `vec_set_intersect()`, `vec_set_difference()`, `vec_set_union()`, and
   `vec_set_symmetric_difference()` which compute set operations like
   `intersect()`, `setdiff()`, and `union()`, but the vctrs variants don't strip
@@ -146,7 +148,7 @@
   like specifying `repair = "unique", quiet = TRUE`. When the `"*_quiet"`
   options are used, any setting of `quiet` is silently overridden (@jennybc,
   #1629).
-  
+
   `"unique_quiet"` and `"universal_quiet"` are also newly accepted for the name
   repair argument of several other functions that do not expose a `quiet`
   argument: `data_frame()`, `df_list()`, `vec_c()`, `list_unchop()`,
@@ -464,11 +466,11 @@
   to implement, but if your class has a static prototype, you might consider
   implementing a custom `vec_ptype()` method that returns a constant to
   improve performance in some cases (such as common type imputation).
-  
+
 * New `vec_detect_complete()`, inspired by `stats::complete.cases()`. For most
   vectors, this is identical to `!vec_equal_na()`. For data frames and
   matrices, this detects rows that only contain non-missing values.
-  
+
 * `vec_order()` can now order complex vectors (#1330).
 
 * Removed dependency on digest in favor of `rlang::hash()`.
@@ -477,7 +479,7 @@
   when used as a data frame column (#1318).
 
 * `register_s3()` is now licensed with the "unlicense" which makes it very
-  clear that it's fine to copy and paste into your own package 
+  clear that it's fine to copy and paste into your own package
   (@maxheld83, #1254).
 
 # vctrs 0.3.6

src/order.c

@@ -3971,6 +3971,9 @@ void df_order_internal(SEXP x,
   // Special case no columns
   if (n_cols == 0) {
     init_order(p_order);
+    if (size != 0) {
+      groups_size_maybe_push(size, p_group_infos);
+    }
     return;
   }
 

tests/testthat/test-order.R

@@ -1275,6 +1275,29 @@ test_that("Indistinct NA and NaN are reported in the same group", {
   expect_identical(info[[3]], 2L)
 })
 
+# ------------------------------------------------------------------------------
+# `vec_order_info(<data.frame>)`
+
+test_that("Zero column data frames with >0 rows work (#1863)", {
+  # All rows are treated as being from the same group
+  x <- data_frame(.size = 5)
+  info <- vec_order_info(x)
+
+  expect_identical(info[[1]], 1:5) # Order
+  expect_identical(info[[2]], 5L)  # Group sizes
+  expect_identical(info[[3]], 5L)  # Max group size
+})
+
+test_that("Zero column data frames with exactly 0 rows work (#1863)", {
+  # This is a particularly special case, since we don't actually push a group size
+  x <- data_frame(.size = 0L)
+  info <- vec_order_info(x)
+
+  expect_identical(info[[1]], integer())
+  expect_identical(info[[2]], integer())
+  expect_identical(info[[3]], 0L)
+})
+
 # ------------------------------------------------------------------------------
 # vec_sort
 

tests/testthat/test-rank.R

@@ -63,6 +63,22 @@ test_that("works with data frames", {
   expect_identical(vec_rank(df, ties = "sequential"), c(2L, 3L, 1L, 4L, 5L))
 })
 
+test_that("works with data frames with 0 columns and >0 rows (#1863)", {
+  # All rows are treated as being from the same group
+  df <- data_frame(.size = 5)
+
+  expect_identical(vec_rank(df, ties = "min"), c(1L, 1L, 1L, 1L, 1L))
+  expect_identical(vec_rank(df, ties = "sequential"), c(1L, 2L, 3L, 4L, 5L))
+  expect_identical(vec_rank(df, ties = "sequential", direction = "desc"), c(1L, 2L, 3L, 4L, 5L))
+})
+
+test_that("works with data frames with 0 columns and 0 rows (#1863)", {
+  df <- data_frame(.size = 0)
+
+  expect_identical(vec_rank(df, ties = "min"), integer())
+  expect_identical(vec_rank(df, ties = "sequential"), integer())
+})
+
 test_that("can control the direction per column", {
   df <- data_frame(
     x = c(1, 2, 1, 2, 2),