Entity Normalization Preserves Types of Copy & Additional Variables (#25

)

* make sure to copy types of variables over when normalizing an entity

* added wrong the wrong branch, now is the right file

* added test

* included copy variables in test

* removed copied code from other pr

* removed whitespace

featuretools/entityset/entityset.py

@@ -710,6 +710,8 @@ def normalize_entity(self, base_entity_id, new_entity_id, index,
 
         transfer_types = {}
         transfer_types[new_index] = type(base_entity[index])
+        for v in additional_variables + copy_variables:
+            transfer_types[v] = type(base_entity[v])
 
         # create and add new entity
         new_entity_df = self.get_dataframe(base_entity_id)

featuretools/tests/entityset_tests/test_pandas_es.py

@@ -609,6 +609,25 @@ def test_normalize_entity(self, entityset):
         assert 'device_name' not in entityset['sessions'].df.columns
         assert 'device_type' in entityset['device_types'].df.columns
 
+    def test_normalize_entity_copies_variable_types(self, entityset):
+        entityset['log'].convert_variable_type('value', variable_types.Ordinal, convert_data=False)
+        assert entityset['log'].variable_types['value'] == variable_types.Ordinal
+        assert entityset['log'].variable_types['priority_level'] == variable_types.Ordinal
+        entityset.normalize_entity('log', 'values_2', 'value_2',
+                                   additional_variables=['priority_level'],
+                                   copy_variables=['value'],
+                                   make_time_index=False)
+
+        assert len(entityset.get_forward_relationships('log')) == 3
+        assert entityset.get_forward_relationships('log')[2].parent_entity.id == 'values_2'
+        assert 'priority_level' in entityset['values_2'].df.columns
+        assert 'value' in entityset['values_2'].df.columns
+        assert 'priority_level' not in entityset['log'].df.columns
+        assert 'value' in entityset['log'].df.columns
+        assert 'value_2' in entityset['values_2'].df.columns
+        assert entityset['values_2'].variable_types['priority_level'] == variable_types.Ordinal
+        assert entityset['values_2'].variable_types['value'] == variable_types.Ordinal
+
     def test_make_time_index_keeps_original_sorting(self):
         trips = {
             'trip_id': [999 - i for i in xrange(1000)],