Improve

dask · Aug 16, 2024 · bcc1c8d · bcc1c8d
1 parent 5d7b71e
commit bcc1c8d
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/dask_expr/_collection.py b/dask_expr/_collection.py
@@ -4237,7 +4237,7 @@ def value_counts(
             if isinstance(self.dtype, CategoricalDtype):
                 # unobserved or huge categories will lead to oom errors
                 if self.cat.known:
-                    split_out = max(len(self.dtype.categories) // 100_000, 1)
+                    split_out = 1 + len(self.dtype.categories) // 100_000
                 else:
                     split_out = True
             else:

diff --git a/dask_expr/tests/test_reductions.py b/dask_expr/tests/test_reductions.py
@@ -550,5 +550,5 @@ def test_cat_value_counts_large_unknown_categories():
 
     df = from_pandas(pdf, npartitions=50)
     result = df.x.value_counts()
-    assert result.npartitions == 2  # known but large
+    assert result.npartitions == 3  # known but large
     assert_eq(result, expected, check_index=False, check_dtype=False)