Esempio n. 1
0
    def test_hive_to_mysql(self):
        test_hive_results = 'test_hive_results'

        mock_hive_hook = MockHiveServer2Hook()
        mock_hive_hook.get_records = MagicMock(return_value=test_hive_results)

        mock_mysql_hook = MockMySqlHook()
        mock_mysql_hook.run = MagicMock()
        mock_mysql_hook.insert_rows = MagicMock()

        with patch('airflow.operators.hive_to_mysql.HiveServer2Hook',
                   return_value=mock_hive_hook):
            with patch('airflow.operators.hive_to_mysql.MySqlHook',
                       return_value=mock_mysql_hook):

                op = HiveToMySqlTransfer(
                    mysql_conn_id='airflow_db',
                    task_id='hive_to_mysql_check',
                    sql="""
                        SELECT name
                        FROM airflow.static_babynames
                        LIMIT 100
                        """,
                    mysql_table='test_static_babynames',
                    mysql_preoperator=[
                        'DROP TABLE IF EXISTS test_static_babynames;',
                        'CREATE TABLE test_static_babynames (name VARCHAR(500))',
                    ],
                    dag=self.dag)
                op.clear(start_date=DEFAULT_DATE, end_date=DEFAULT_DATE)
                op.run(start_date=DEFAULT_DATE,
                       end_date=DEFAULT_DATE,
                       ignore_ti_state=True)

        raw_select_name_query = mock_hive_hook.get_records.call_args_list[0][
            0][0]
        actual_select_name_query = re.sub(r'\s{2,}', ' ',
                                          raw_select_name_query).strip()
        expected_select_name_query = 'SELECT name FROM airflow.static_babynames LIMIT 100'
        self.assertEqual(expected_select_name_query, actual_select_name_query)

        actual_hive_conf = mock_hive_hook.get_records.call_args_list[0][1][
            'hive_conf']
        expected_hive_conf = {
            'airflow.ctx.dag_owner': 'airflow',
            'airflow.ctx.dag_id': 'test_dag_id',
            'airflow.ctx.task_id': 'hive_to_mysql_check',
            'airflow.ctx.execution_date': '2015-01-01T00:00:00+00:00'
        }
        self.assertEqual(expected_hive_conf, actual_hive_conf)

        expected_mysql_preoperator = [
            'DROP TABLE IF EXISTS test_static_babynames;',
            'CREATE TABLE test_static_babynames (name VARCHAR(500))'
        ]
        mock_mysql_hook.run.assert_called_with(expected_mysql_preoperator)

        mock_mysql_hook.insert_rows.assert_called_with(
            table='test_static_babynames', rows=test_hive_results)
Esempio n. 2
0
 def test_hive_to_mysql(self):
     op = HiveToMySqlTransfer(
         mysql_conn_id='airflow_db',
         task_id='hive_to_mysql_check',
         create=True,
         sql="""
             SELECT name
             FROM airflow.static_babynames
             LIMIT 100
             """,
         mysql_table='test_static_babynames',
         mysql_preoperator=[
             'DROP TABLE IF EXISTS test_static_babynames;',
             'CREATE TABLE test_static_babynames (name VARCHAR(500))',
         ],
         dag=self.dag)
     op.clear(start_date=DEFAULT_DATE, end_date=DEFAULT_DATE)
     op.run(start_date=DEFAULT_DATE, end_date=DEFAULT_DATE,
            ignore_ti_state=True)