def test_spark_process_runcmd(self, mock_popen): # Given mock_popen.return_value.stdout = six.StringIO( 'Spark-sql communicates using stdout') mock_popen.return_value.stderr = six.StringIO('stderr') mock_popen.return_value.wait.return_value = 0 # When hook = SparkSqlHook(conn_id='spark_default', sql='SELECT 1') with patch.object(hook.log, 'debug') as mock_debug: with patch.object(hook.log, 'info') as mock_info: hook.run_query() mock_debug.assert_called_with('Spark-Sql cmd: %s', [ 'spark-sql', '-e', 'SELECT 1', '--master', 'yarn', '--name', 'default-name', '--verbose', '--queue', 'default' ]) mock_info.assert_called_with( 'Spark-sql communicates using stdout') # Then self.assertEqual( mock_popen.mock_calls[0], call([ 'spark-sql', '-e', 'SELECT 1', '--master', 'yarn', '--name', 'default-name', '--verbose', '--queue', 'default' ], stderr=-2, stdout=-1))
def test_build_command(self): hook = SparkSqlHook(**self._config) # The subprocess requires an array but we build the cmd by joining on a space cmd = ' '.join(hook._prepare_command("")) # Check all the parameters assert "--executor-cores {}".format( self._config['executor_cores']) in cmd assert "--executor-memory {}".format( self._config['executor_memory']) in cmd assert "--keytab {}".format(self._config['keytab']) in cmd assert "--name {}".format(self._config['name']) in cmd assert "--num-executors {}".format( self._config['num_executors']) in cmd sql_path = get_after('-f', hook._prepare_command("")) assert self._config['sql'].strip() == sql_path # Check if all config settings are there for kv in self._config['conf'].split(","): k, v = kv.split('=') assert "--conf {0}={1}".format(k, v) in cmd if self._config['verbose']: assert "--verbose" in cmd
def execute(self, context): """ Call the SparkSqlHook to run the provided sql query """ self._hook = SparkSqlHook(sql=self._sql, conf=self._conf, conn_id=self._conn_id, executor_cores=self._executor_cores, executor_memory=self._executor_memory, keytab=self._keytab, name=self._name, num_executors=self._num_executors, master=self._master, yarn_queue=self._yarn_queue) self._hook.run_query()
def test_spark_process_runcmd_with_list(self, mock_popen): # Given mock_popen.return_value.stdout = six.StringIO('Spark-sql communicates using stdout') mock_popen.return_value.stderr = six.StringIO('stderr') mock_popen.return_value.wait.return_value = 0 # When hook = SparkSqlHook( conn_id='spark_default', sql='SELECT 1' ) hook.run_query(['--deploy-mode', 'cluster']) # Then self.assertEqual( mock_popen.mock_calls[0], call(['spark-sql', '-e', 'SELECT 1', '--master', 'yarn', '--name', 'default-name', '--verbose', '--queue', 'default', '--deploy-mode', 'cluster'], stderr=-2, stdout=-1) )