fix: Fix perfect groupby (#19461)

pola-rs · Oct 26, 2024 · ce001f1 · ce001f1
1 parent d616866
commit ce001f1
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 0 deletions.
diff --git a/crates/polars-core/src/frame/group_by/perfect.rs b/crates/polars-core/src/frame/group_by/perfect.rs
@@ -61,6 +61,7 @@ where
                     per_thread_offsets.push(std::cmp::min(cache_aligned_offset, len));
                 }
             }
+            *per_thread_offsets.last_mut().unwrap() = len;
 
             let groups_ptr = unsafe { SyncPtr::new(groups.as_mut_ptr()) };
             let first_ptr = unsafe { SyncPtr::new(first.as_mut_ptr()) };
@@ -75,6 +76,10 @@ where
                     let end = per_thread_offsets[thread_no + 1];
                     let end = T::Native::from_usize(end).unwrap();
 
+                    if start == end {
+                        return;
+                    }
+
                     let push_to_group = |cat, row_nr| unsafe {
                         debug_assert!(cat < len);
                         let buf = &mut *groups.add(cat);

diff --git a/py-polars/tests/unit/datatypes/test_categorical.py b/py-polars/tests/unit/datatypes/test_categorical.py
@@ -864,3 +864,15 @@ def test_nested_categorical_concat(
 
     with pytest.raises(pl.exceptions.StringCacheMismatchError):
         pl.concat([a, b])
+
+
+def test_perfect_group_by_19452() -> None:
+    n = 40
+    df2 = pl.DataFrame(
+        {
+            "a": pl.int_range(n, eager=True).cast(pl.String).cast(pl.Categorical),
+            "b": pl.int_range(n, eager=True),
+        }
+    )
+
+    assert df2.with_columns(a=(pl.col("b")).over(pl.col("a")))["a"].is_sorted()