Support project.dataset in schema argument

Author: vinceatbluelabs

pybigquery/sqlalchemy_bigquery.py

@@ -382,7 +382,8 @@ def _table_reference(self, provided_schema_name, provided_table_name,
                 else:
                     dataset_id_from_schema = provided_schema_name_split[0]
             elif len(provided_schema_name_split) == 2:
-                pass
+                project_id_from_schema = provided_schema_name_split[0]
+                dataset_id_from_schema = provided_schema_name_split[1]
             # TODO: Get a test for an else statement here
         project_id = project_id_from_schema or project_id_from_table or client_project
         dataset_id = dataset_id_from_schema or dataset_id_from_table or self.dataset_id

test/test_sqlalchemy_bigquery.py

@@ -489,6 +489,7 @@ def test_get_columns(inspector, inspector_using_test_dataset):
                              (None, 'dataset.table', 'project'),
                              (None, 'project.dataset.table', 'other_project'),
                              ('project', 'dataset.table', 'other_project'),
+                             ('project.dataset', 'table', 'other_project'),
                          ])
 def test_table_reference(dialect, provided_schema_name,
                          provided_table_name, client_project):