Add Dataframe and Index nunique

python/cudf/cudf/core/dataframe.py

@@ -6145,6 +6145,37 @@ def __dataframe__(
             self, nan_as_null=nan_as_null, allow_copy=allow_copy
         )
 
+    def nunique(self, axis=0, dropna=True):
+        """
+        Count number of distinct elements in specified axis.
+        Return Series with number of distinct elements. Can ignore NaN values.
+        
+        Parameters
+        ----------
+        axis : {0 or 'index', 1 or 'columns'}, default 0
+            The axis to use. 0 or 'index' for row-wise, 1 or 'columns' for column-wise.
+        dropna : bool, default True
+            Don't include NaN in the counts.
+
+        Returns
+        -------
+        Series
+
+        Examples
+        --------
+        >>> import cudf
+        >>> df = cudf.DataFrame({'A': [4, 5, 6], 'B': [4, 1, 1]})
+        >>> df.nunique()
+        A    3
+        B    2
+        dtype: int64
+        """
+        if axis != 0:
+            raise NotImplementedError(
+                "axis parameter is not supported yet."
+            )
+
+        return cudf.Series(super().nunique(method="sort", dropna=dropna))
 
 def from_dataframe(df, allow_copy=False):
     return df_protocol.from_dataframe(df, allow_copy=allow_copy)

python/cudf/cudf/core/frame.py

@@ -6420,6 +6420,27 @@ def ge(self, other, axis="columns", level=None, fill_value=None):
             other=other, fn="ge", fill_value=fill_value, can_reindex=True
         )
 
+    def nunique(self, method: builtins.str = "sort", dropna: bool = True):
+        """
+        Returns a per column mapping with counts of unique values for
+        each column.
+
+        Parameters
+        ----------
+        method : builtins.str, default "sort"
+            Method used by cpp_distinct_count
+        dropna : bool, default True
+            Don't include NaN in the counts.
+
+        Returns
+        -------
+        dict
+            Name and unique value counts of each column in frame.
+        """
+        return {
+            name: col.distinct_count(method=method, dropna=dropna) 
+                for name, col in self._data.items()
+        }
 
 def _get_replacement_values_for_columns(
     to_replace: Any, value: Any, columns_dtype_map: Dict[Any, Any]

python/cudf/cudf/core/series.py

@@ -2947,7 +2947,7 @@ def nunique(self, method="sort", dropna=True):
             raise NotImplementedError(msg)
         if self.null_count == len(self):
             return 0
-        return self._column.distinct_count(method, dropna)
+        return super().nunique(method, dropna)
 
     def value_counts(
         self,

python/cudf/cudf/core/single_column_frame.py

@@ -336,3 +336,21 @@ def _make_operands_for_binop(
                 return NotImplemented
 
         return {result_name: (self._column, other, reflect, fill_value)}
+
+    def nunique(self, method: builtins.str = "sort", dropna: bool = True):
+        """
+        Returns count of unique values for the column.
+
+        Parameters
+        ----------
+        method : builtins.str, default "sort"
+            Method used by cpp_distinct_count
+        dropna : bool, default True
+            Don't include NaN in the counts.
+
+        Returns
+        -------
+        int
+            Number of unique values in the column.
+        """
+        return sum(super().nunique(method=method, dropna=dropna).values())

python/cudf/cudf/tests/test_dataframe.py

@@ -9077,3 +9077,27 @@ def test_dataframe_assign_cp_np_array():
     gdf[[f"f_{i}" for i in range(n)]] = cp_ndarray
 
     assert_eq(pdf, gdf)
+
+@pytest.mark.parametrize(
+    "data", [{"a": [1, 2, 3], "b": [1, 1, 0]}],
+)
+def test_dataframe_nunique(data):
+    gdf = cudf.DataFrame(data)
+    pdf = gdf.to_pandas()
+
+    actual = gdf.nunique()
+    expected = pdf.nunique()
+
+    assert_eq(expected, actual)
+
+@pytest.mark.parametrize(
+    "data", [{ "key": [0, 1, 1, 0, 0, 1], "val": [1, 8, 3, 9, -3, 8]}],
+)
+def test_dataframe_nunique_index(data):
+    gdf = cudf.DataFrame(data)
+    pdf = gdf.to_pandas()
+
+    actual = gdf.index.nunique()
+    expected = pdf.index.nunique()
+
+    assert_eq(expected, actual)

python/cudf/cudf/tests/test_series.py

@@ -1517,3 +1517,27 @@ def test_series_transpose(data):
     assert_eq(pd_transposed, cudf_transposed)
     assert_eq(pd_property, cudf_property)
     assert_eq(cudf_transposed, csr)
+
+@pytest.mark.parametrize(
+    "data", [1, 3, 5, 7, 7],
+)
+def test_series_nunique(data):
+    cd_s = cudf.Series(data)
+    pd_s = cd_s.to_pandas()
+
+    actual = cd_s.nunique()
+    expected = pd_s.nunique()
+
+    assert_eq(expected, actual)
+
+@pytest.mark.parametrize(
+    "data", [1, 3, 5, 7, 7],
+)
+def test_series_nunique_index(data):
+    cd_s = cudf.Series(data)
+    pd_s = cd_s.to_pandas()
+
+    actual = cd_s.index.nunique()
+    expected = pd_s.index.nunique()
+
+    assert_eq(expected, actual)