def test_write_table_sync_append(self, data_dir, credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
csv_file_path = data_dir + 'simple_csv/in/tables/table.csv'
my_writer.write_table_sync(csv_file_path,
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
my_writer.write_table_sync(csv_file_path,
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration(),
incremental=True)
query = 'SELECT * FROM %s.%s ORDER BY 1 ASC, 2 ASC' % (os.environ.get(
'BIGQUERY_DATASET'), os.environ.get('BIGQUERY_TABLE'))
client = self.get_client('service_account_manage')
query_job = client.query(query)
row_data = list(query_job)
assert len(row_data) == 4
assert row_data[0].col1 == 'val1'
assert row_data[0].col2 == 1
assert row_data[1].col1 == 'val1'
assert row_data[1].col2 == 1
assert row_data[2].col1 == 'val2'
assert row_data[2].col2 == 2
assert row_data[3].col1 == 'val2'
assert row_data[3].col2 == 2
def test_create_dataset_invalid_name(self, data_dir, credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
try:
my_writer.write_table_sync(
data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET') + ' INVALID',
fixtures.get_table_configuration())
pytest.fail('Must raise exception.')
except exceptions.UserException as err:
assert 'Cannot create dataset' in str(err)
def test_write_table_connection_error(self, data_dir, credentials_type):
client = self.get_client(credentials_type=credentials_type)
error_msg = "Some connection error!"
client.load_table_from_file = MagicMock(
side_effect=requests.exceptions.ConnectionError(error_msg))
my_writer = writer.Writer(client)
try:
my_writer.write_table(data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
except exceptions.UserException as err:
msg = 'Loading data into table {}.{} failed: {}'\
.format(os.environ.get('BIGQUERY_DATASET'), os.environ.get('BIGQUERY_TABLE'), error_msg)
assert msg in str(err)
def test_write_table_sync_error_too_many_values(self, data_dir,
credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
my_writer.write_table_sync(data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
try:
my_writer.write_table_sync(
data_dir + 'simple_csv_with_extra_column/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration_with_extra_column(), True)
pytest.fail('Must raise exception.')
except exceptions.UserException as err:
assert 'Too many values in row' in str(err)
def test_write_table_sync_error_invalid_datatype(self, data_dir,
credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
my_writer.write_table_sync(data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
try:
my_writer.write_table_sync(
data_dir + 'simple_csv_invalid_data_types/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration_with_invalid_data_type(),
True)
pytest.fail('Must raise exception.')
except exceptions.UserException as err:
assert 'Could not parse \'val1\' as INT64 for field col2' in str(
err)
def test_invalid_project(self, data_dir, credentials_type):
if (credentials_type == 'service_account'):
credentials = self.get_service_account_user_credentials()
else:
credentials = self.get_oauth_credentials()
bigquery_client = bigquery.Client('invalid-project', credentials)
my_writer = writer.Writer(bigquery_client)
try:
my_writer.write_table_sync(
data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
pytest.fail('Must raise exception.')
except exceptions.UserException as err:
assert 'Project invalid-project was not found.' in str(err)
def test_write_table_sync_with_invalid_column_order(
self, data_dir, credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
my_writer.write_table_sync(data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
try:
my_writer.write_table_sync(
'%ssimple_csv_invalid_column_order/in/tables/table.csv' %
data_dir, os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration_with_invalid_column_order(),
True)
pytest.fail('Must raise exception.')
except exceptions.UserException as err:
assert 'Column order mismatch. Actual configuration: '\
'col2, col1. Expected BigQuery: col1, col2.'\
in str(err)
def test_write_table_sync_newlines(self, data_dir, credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
my_writer.write_table_sync(data_dir + 'newlines/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
query = 'SELECT * FROM %s.%s ORDER BY 1 ASC' % (os.environ.get(
'BIGQUERY_DATASET'), os.environ.get('BIGQUERY_TABLE'))
client = self.get_client('service_account_manage')
query_job = client.query(query)
row_data = list(query_job)
assert len(row_data) == 2
assert row_data[0].col1 == 'val1\non new line'
assert row_data[0].col2 == 1
assert row_data[1].col1 == 'val2'
assert row_data[1].col2 == 2
def test_invalid_token(self, data_dir):
invalid_credentials = Credentials(
'access-token',
token_uri='https://accounts.google.com/o/oauth2/token',
client_id=os.environ.get('OAUTH_CLIENT_ID'),
client_secret=os.environ.get('OAUTH_CLIENT_SECRET'))
bigquery_client = bigquery.Client(self.get_project(),
invalid_credentials)
my_writer = writer.Writer(bigquery_client)
try:
my_writer.write_table(data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
pytest.fail('Must raise exception.')
except RefreshError as err:
assert str(err) == 'The credentials do not contain the necessary' \
' fields need to refresh the access token.' \
' You must specify refresh_token, token_uri,' \
' client_id, and client_secret.'
def test_write_table_schema(self, data_dir, credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
my_writer.write_table_sync(data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
client = self.get_client('service_account_manage')
dataset = client.dataset(os.environ.get('BIGQUERY_DATASET'))
table_reference = dataset.table(os.environ.get('BIGQUERY_TABLE'))
table = client.get_table(table_reference)
assert table.schema[0].field_type == 'STRING'
assert table.schema[0].fields == ()
assert table.schema[0].mode == 'NULLABLE'
assert table.schema[0].name == 'col1'
assert table.schema[1].field_type == 'INTEGER'
assert table.schema[1].fields == ()
assert table.schema[1].mode == 'NULLABLE'
assert table.schema[1].name == 'col2'
def test_write_table_async(self, data_dir, credentials_type):
my_writer = writer.Writer(
self.get_client(credentials_type=credentials_type))
job = my_writer.write_table(
data_dir + 'simple_csv/in/tables/table.csv',
os.environ.get('BIGQUERY_DATASET'),
fixtures.get_table_configuration())
assert job.state == 'RUNNING'
client = self.get_client('service_account_manage')
dataset_reference = client.dataset(os.environ.get('BIGQUERY_DATASET'))
try:
dataset = client.get_dataset(dataset_reference)
except exceptions.NotFound:
pytest.fail('Must not raise an exception.')
table_reference = dataset.table(os.environ.get('BIGQUERY_TABLE'))
try:
client.get_table(table_reference)
except exceptions.NotFound:
pytest.fail('Must not raise an exception.')
