Fix DataFrame.values with no columns but index (#14134)

Fixes the following

```python
In [32]: cudf.DataFrame(index=range(10)).values
Out[32]: array([], shape=(0, 0), dtype=float64)
```

Authors:
  - Matthew Roeschke (https://github.com/mroeschke)

Approvers:
  - GALI PREM SAGAR (https://github.com/galipremsagar)

URL: #14134

python/cudf/cudf/core/frame.py

@@ -437,7 +437,7 @@ def get_column_values_na(col):
         ncol = self._num_columns
         if ncol == 0:
             return make_empty_matrix(
-                shape=(0, 0), dtype=np.dtype("float64"), order="F"
+                shape=(len(self), ncol), dtype=np.dtype("float64"), order="F"
             )
 
         if dtype is None:

python/cudf/cudf/tests/test_dataframe.py

@@ -10374,3 +10374,9 @@ def test_dataframe_init_from_nested_dict():
     pdf = pd.DataFrame(regular_dict)
     gdf = cudf.DataFrame(regular_dict)
     assert_eq(pdf, gdf)
+
+
+def test_data_frame_values_no_cols_but_index():
+    result = cudf.DataFrame(index=range(5)).values
+    expected = pd.DataFrame(index=range(5)).values
+    assert_eq(result, expected)