fix: two CellService bugs — multi-hierarchy dict overwrite and missing kwargs

TM1py/Services/CellService.py

@@ -772,10 +772,17 @@ def clear_with_dataframe(self, cube: str, df: "pd.DataFrame", dimension_mapping:
                     ).filter_by_level(0)
 
                 elif isinstance(hierarchies, Iterable):
+                    # Build a union across all hierarchies so every hierarchy
+                    # contributes to the MDX crossjoin instead of the last one
+                    # overwriting all previous entries (issue #1068).
+                    combined: Optional["MdxHierarchySet"] = None
                     for hierarchy in hierarchies:
-                        mdx_selections[dimension] = MdxHierarchySet.tm1_subset_all(
+                        hset = MdxHierarchySet.tm1_subset_all(
                             dimension=dimension, hierarchy=hierarchy
                         ).filter_by_level(0)
+                        combined = hset if combined is None else combined.union(hset)
+                    if combined is not None:
+                        mdx_selections[dimension] = combined
 
                 else:
                     raise ValueError(f"Unexpected value type for key '{dimension}' in dimension_mapping")
@@ -3123,19 +3130,20 @@ def execute_view_dataframe_pivot(
 
     @require_pandas
     def execute_mdx_dataframe_pivot(
-        self, mdx: str, dropna: bool = False, fill_value: bool = None, sandbox_name: str = None
+        self, mdx: str, dropna: bool = False, fill_value: bool = None, sandbox_name: str = None, **kwargs
     ) -> "pd.DataFrame":
         """Execute MDX Query to get a pandas pivot data frame in the shape as specified in the Query
 
         :param mdx:
         :param dropna:
         :param fill_value:
         :param sandbox_name: str
+        :param kwargs: Additional keyword arguments forwarded to extract_cellset_dataframe_pivot
         :return:
         """
         cellset_id = self.create_cellset(mdx=mdx, sandbox_name=sandbox_name)
         return self.extract_cellset_dataframe_pivot(
-            cellset_id=cellset_id, dropna=dropna, fill_value=fill_value, sandbox_name=sandbox_name
+            cellset_id=cellset_id, dropna=dropna, fill_value=fill_value, sandbox_name=sandbox_name, **kwargs
         )
 
     def execute_mdx_cellcount(self, mdx: str, sandbox_name: str = None, **kwargs) -> int:

Tests/test_cellservice_unit.py

@@ -0,0 +1,142 @@
+"""Unit tests for CellService logic that can run without a live TM1 server.
+
+Covers:
+  - clear_with_dataframe: multiple hierarchies in dimension_mapping (issue #1068)
+  - execute_mdx_dataframe_pivot: **kwargs forwarding (issue #1113)
+"""
+
+import unittest
+from unittest.mock import MagicMock, patch, call
+
+import pandas as pd
+
+from TM1py.Services.CellService import CellService
+
+
+def _make_cell_service():
+    """Return a CellService with the REST connection mocked out."""
+    rest = MagicMock()
+    rest.is_data_admin = True
+    rest.is_ops_admin = True
+    rest.is_admin = True
+    rest.version = "11.8.000"
+    svc = CellService.__new__(CellService)
+    svc._rest = rest
+    svc._tm1_rest = rest
+    return svc
+
+
+class TestClearWithDataframeMultipleHierarchies(unittest.TestCase):
+    """Regression test for #1068.
+
+    When dimension_mapping maps a single dimension name to a list of hierarchies,
+    the generated MDX must include *all* hierarchies, not just the last one.
+    """
+
+    def _run(self, dimension_mapping, cube="MyCube", dim_names=None):
+        """Run clear_with_dataframe with mocked dependencies and return the MDX string."""
+        svc = _make_cell_service()
+
+        if dim_names is None:
+            dim_names = list(dimension_mapping.keys())
+
+        # Mock cube_service.get_dimension_names to return the cube's dimensions
+        mock_cube_svc = MagicMock()
+        mock_cube_svc.get_dimension_names.return_value = dim_names
+        svc.get_cube_service = MagicMock(return_value=mock_cube_svc)
+
+        # Capture the MDX passed to clear_with_mdx instead of running it
+        captured = {}
+        svc.clear_with_mdx = lambda cube, mdx, **kw: captured.update({"mdx": mdx})
+
+        df = pd.DataFrame({d: ["elem1"] for d in dim_names if d not in dimension_mapping})
+
+        svc.clear_with_dataframe(cube=cube, df=df, dimension_mapping=dimension_mapping)
+        return captured.get("mdx", "")
+
+    def test_single_hierarchy_still_works(self):
+        """Sanity-check: a str mapping produces valid MDX."""
+        mdx = self._run(
+            dim_names=["Year", "Month", "MyDim"],
+            dimension_mapping={"MyDim": "MyHierarchy"},
+        )
+        self.assertIn("myhierarchy", mdx.lower())
+
+    def test_two_hierarchies_both_appear(self):
+        """Both hierarchies must appear in the MDX (not just the last one)."""
+        mdx = self._run(
+            dim_names=["Year", "Month", "MyDim"],
+            dimension_mapping={"MyDim": ["Hierarchy1", "Hierarchy2"]},
+        )
+        mdx_lower = mdx.lower()
+        self.assertIn("hierarchy1", mdx_lower, "First hierarchy missing from MDX")
+        self.assertIn("hierarchy2", mdx_lower, "Second hierarchy missing from MDX")
+        # Both should appear in a UNION, not just the last one overwriting
+        self.assertIn("union", mdx_lower, "Expected UNION of hierarchies in MDX")
+
+    def test_three_hierarchies_all_appear(self):
+        """All three hierarchies must appear."""
+        mdx = self._run(
+            dim_names=["Year", "Month", "MyDim"],
+            dimension_mapping={"MyDim": ["H1", "H2", "H3"]},
+        )
+        mdx_lower = mdx.lower()
+        for h in ("h1", "h2", "h3"):
+            self.assertIn(h, mdx_lower, f"{h} missing from MDX")
+
+
+class TestExecuteMdxDataframePivotKwargs(unittest.TestCase):
+    """Regression test for #1113.
+
+    execute_mdx_dataframe_pivot must forward **kwargs to extract_cellset_dataframe_pivot.
+    """
+
+    def test_kwargs_forwarded(self):
+        svc = _make_cell_service()
+
+        svc.create_cellset = MagicMock(return_value="cellset-id-123")
+        svc.extract_cellset_dataframe_pivot = MagicMock(return_value=pd.DataFrame())
+
+        svc.execute_mdx_dataframe_pivot(
+            mdx="SELECT ... FROM [Cube]",
+            dropna=True,
+            fill_value=0,
+            sandbox_name="MySandbox",
+            use_compact_json=True,
+        )
+
+        svc.extract_cellset_dataframe_pivot.assert_called_once_with(
+            cellset_id="cellset-id-123",
+            dropna=True,
+            fill_value=0,
+            sandbox_name="MySandbox",
+            use_compact_json=True,
+        )
+
+    def test_extra_kwargs_forwarded(self):
+        """Arbitrary extra kwargs like cell_properties are forwarded."""
+        svc = _make_cell_service()
+        svc.create_cellset = MagicMock(return_value="cid")
+        svc.extract_cellset_dataframe_pivot = MagicMock(return_value=pd.DataFrame())
+
+        svc.execute_mdx_dataframe_pivot(
+            mdx="SELECT FROM [Cube]",
+            cell_properties=["FormattedValue"],
+        )
+
+        _, kwargs = svc.extract_cellset_dataframe_pivot.call_args
+        self.assertIn("cell_properties", kwargs)
+        self.assertEqual(kwargs["cell_properties"], ["FormattedValue"])
+
+    def test_no_kwargs_still_works(self):
+        """Calling without kwargs must not raise."""
+        svc = _make_cell_service()
+        svc.create_cellset = MagicMock(return_value="cid")
+        svc.extract_cellset_dataframe_pivot = MagicMock(return_value=pd.DataFrame())
+
+        result = svc.execute_mdx_dataframe_pivot(mdx="SELECT FROM [Cube]")
+        self.assertIsInstance(result, pd.DataFrame)
+
+
+if __name__ == "__main__":
+    unittest.main()