Bug fix for GetRecordResponse not printable and error message for ingest_dataframe

Author: mollyheamazon

sagemaker-core/src/sagemaker/core/utils/utils.py

@@ -298,6 +298,22 @@ def __new__(cls):
             cls._instance = super().__new__(cls)
         return cls._instance
 
+    def __repr__(self):
+        """Return clean representation for debugging."""
+        return "Unassigned()"
+
+    def __str__(self):
+        """Return empty string for clean printing."""
+        return ""
+
+    def __iter__(self):
+        """Return empty iterator to prevent iteration errors."""
+        return iter([])
+
+    def __bool__(self):
+        """Return False for truthiness checks."""
+        return False
+
 
 class SingletonMeta(type):
     """

sagemaker-core/tests/unit/generated/test_utils.py

@@ -371,3 +371,50 @@ def test_serialize_method_nested_shape():
             "StringValue": "string",
         },
     }
+
+
+class TestUnassignedBehavior:
+    """Test Unassigned class methods for proper behavior.
+    
+    Bug fix: GetRecordResponse is not printable and cannot be parsed via iterator.
+    Error: TypeError: 'Unassigned' object is not iterable
+    """
+
+    def test_unassigned_repr(self):
+        """Test that Unassigned has clean repr."""
+        u = Unassigned()
+        assert repr(u) == "Unassigned()"
+
+    def test_unassigned_str(self):
+        """Test that Unassigned converts to empty string."""
+        u = Unassigned()
+        assert str(u) == ""
+
+    def test_unassigned_bool(self):
+        """Test that Unassigned is falsy."""
+        u = Unassigned()
+        assert not u
+        assert bool(u) is False
+
+    def test_unassigned_iter(self):
+        """Test that Unassigned is iterable and returns empty list."""
+        u = Unassigned()
+        result = list(u)
+        assert result == []
+
+    def test_unassigned_singleton(self):
+        """Test that Unassigned is a singleton."""
+        u1 = Unassigned()
+        u2 = Unassigned()
+        assert u1 is u2
+
+    def test_unassigned_in_conditional(self):
+        """Test that Unassigned works correctly in conditionals."""
+        u = Unassigned()
+        
+        # Should evaluate to False
+        if u:
+            pytest.fail("Unassigned should be falsy")
+        
+        # Should work with not
+        assert not u

sagemaker-mlops/src/sagemaker/mlops/feature_store/ingestion_manager_pandas.py

@@ -85,7 +85,17 @@ def run(
             wait (bool): whether to wait for the ingestion to finish or not.
             timeout (Union[int, float]): ``concurrent.futures.TimeoutError`` will be raised
                 if timeout is reached.
+        
+        Raises:
+            ValueError: If wait=False with max_workers=1 and max_processes=1.
         """
+        # Validate async ingestion requirements
+        if not wait and self.max_workers == 1 and self.max_processes == 1:
+            raise ValueError(
+                "Async ingestion (wait=False) requires max_processes > 1 or max_workers > 1. "
+                "Single-threaded ingestion only supports synchronous mode (wait=True)."
+            )
+        
         if self.max_workers == 1 and self.max_processes == 1:
             self._run_single_process_single_thread(data_frame=data_frame, target_stores=target_stores)
         else:

sagemaker-mlops/tests/unit/sagemaker/mlops/feature_store/test_ingestion_manager_pandas.py

@@ -254,3 +254,83 @@ def test_ingest_row_skips_none_values(self, feature_definitions):
         # Only id should be in record, name is None
         assert len(record) == 1
         assert record[0].feature_name == "id"
+
+
+class TestAsyncIngestionValidation:
+    """Test async ingestion validation with max_processes=1.
+    
+    Bug fix: Error message unclear when trying to use async ingestion with 1 process.
+    """
+
+    def test_async_with_single_process_single_worker_raises_clear_error(self):
+        """Test that wait=False with max_processes=1 and max_workers=1 raises clear error."""
+        manager = IngestionManagerPandas(
+            feature_group_name="test-fg",
+            feature_definitions={"id": {"FeatureType": "String", "CollectionType": None}},
+            max_workers=1,
+            max_processes=1,
+        )
+        
+        df = pd.DataFrame({"id": ["1", "2", "3"]})
+        
+        with pytest.raises(ValueError) as exc_info:
+            manager.run(data_frame=df, wait=False)
+        
+        error_message = str(exc_info.value)
+        assert "Async ingestion (wait=False)" in error_message
+        assert "max_processes > 1 or max_workers > 1" in error_message
+        assert "wait=True" in error_message
+
+    @patch("sagemaker.mlops.feature_store.ingestion_manager_pandas.CoreFeatureGroup")
+    def test_sync_with_single_process_single_worker_works(self, mock_fg_class):
+        """Test that wait=True with max_processes=1 and max_workers=1 works."""
+        mock_fg = Mock()
+        mock_fg_class.return_value = mock_fg
+        
+        manager = IngestionManagerPandas(
+            feature_group_name="test-fg",
+            feature_definitions={"id": {"FeatureType": "String", "CollectionType": None}},
+            max_workers=1,
+            max_processes=1,
+        )
+        
+        df = pd.DataFrame({"id": ["1", "2", "3"]})
+        
+        # Should not raise validation error
+        manager.run(data_frame=df, wait=True)
+
+    def test_async_with_multiple_workers_no_validation_error(self):
+        """Test that wait=False with max_workers > 1 doesn't raise validation error."""
+        manager = IngestionManagerPandas(
+            feature_group_name="test-fg",
+            feature_definitions={"id": {"FeatureType": "String", "CollectionType": None}},
+            max_workers=2,
+            max_processes=1,
+        )
+        
+        df = pd.DataFrame({"id": ["1", "2", "3"]})
+        
+        # Should not raise validation error (will fail on actual ingestion, but that's expected)
+        with pytest.raises(Exception) as exc_info:
+            manager.run(data_frame=df, wait=False)
+        
+        # Make sure it's not the validation error
+        assert "Async ingestion (wait=False)" not in str(exc_info.value)
+
+    def test_async_with_multiple_processes_no_validation_error(self):
+        """Test that wait=False with max_processes > 1 doesn't raise validation error."""
+        manager = IngestionManagerPandas(
+            feature_group_name="test-fg",
+            feature_definitions={"id": {"FeatureType": "String", "CollectionType": None}},
+            max_workers=1,
+            max_processes=2,
+        )
+        
+        df = pd.DataFrame({"id": ["1", "2", "3"]})
+        
+        # Should not raise validation error (will fail on actual ingestion, but that's expected)
+        with pytest.raises(Exception) as exc_info:
+            manager.run(data_frame=df, wait=False)
+        
+        # Make sure it's not the validation error
+        assert "Async ingestion (wait=False)" not in str(exc_info.value)