def setUp(self):
db.merge_conn(
Connection(conn_id='cassandra_test',
conn_type='cassandra',
host='host-1,host-2',
port='9042',
schema='test_keyspace',
extra='{"load_balancing_policy":"TokenAwarePolicy"}'))
db.merge_conn(
Connection(conn_id='cassandra_default_with_schema',
conn_type='cassandra',
host='cassandra',
port='9042',
schema='s'))
hook = CassandraHook("cassandra_default")
session = hook.get_conn()
cqls = [
"DROP SCHEMA IF EXISTS s",
"""
CREATE SCHEMA s WITH REPLICATION =
{ 'class' : 'SimpleStrategy', 'replication_factor' : 1 }
""",
]
for cql in cqls:
session.execute(cql)
session.shutdown()
hook.shutdown_cluster()
def setUp(self):
configuration.load_test_config()
db.merge_conn(
models.Connection(
conn_id='cassandra_test', conn_type='cassandra',
host='host-1,host-2', port='9042', schema='test_keyspace',
extra='{"load_balancing_policy":"TokenAwarePolicy"}'))
db.merge_conn(
models.Connection(
conn_id='cassandra_default_with_schema', conn_type='cassandra',
host='cassandra', port='9042', schema='s'))
hook = CassandraHook("cassandra_default")
session = hook.get_conn()
cqls = [
"DROP SCHEMA IF EXISTS s",
"""
CREATE SCHEMA s WITH REPLICATION =
{ 'class' : 'SimpleStrategy', 'replication_factor' : 1 }
""",
]
for cql in cqls:
session.execute(cql)
session.shutdown()
hook.shutdown_cluster()
def test_record_exists(self):
hook = CassandraHook()
session = hook.get_conn()
cqls = [
"DROP SCHEMA IF EXISTS s",
"""
CREATE SCHEMA s WITH REPLICATION =
{ 'class' : 'SimpleStrategy', 'replication_factor' : 1 }
""",
"DROP TABLE IF EXISTS s.t",
"CREATE TABLE s.t (pk1 text, pk2 text, c text, PRIMARY KEY (pk1, pk2))",
"INSERT INTO s.t (pk1, pk2, c) VALUES ('foo', 'bar', 'baz')",
]
for cql in cqls:
session.execute(cql)
self.assertTrue(
hook.record_exists("s.t", {
"pk1": "foo",
"pk2": "bar"
}))
self.assertFalse(
hook.record_exists("s.t", {
"pk1": "foo",
"pk2": "baz"
}))
def _query_cassandra(self):
"""
Queries cassandra and returns a cursor to the results.
"""
self.hook = CassandraHook(cassandra_conn_id=self.cassandra_conn_id)
session = self.hook.get_conn()
cursor = session.execute(self.cql)
return cursor
def _query_cassandra(self):
"""
Queries cassandra and returns a cursor to the results.
"""
hook = CassandraHook(cassandra_conn_id=self.cassandra_conn_id)
session = hook.get_conn()
cursor = session.execute(self.cql)
return cursor
def test_get_conn(self):
with mock.patch.object(Cluster, "connect") as mock_connect, \
mock.patch("socket.getaddrinfo", return_value=[]) as mock_getaddrinfo:
mock_connect.return_value = 'session'
hook = CassandraHook(cassandra_conn_id='cassandra_test')
hook.get_conn()
mock_getaddrinfo.assert_called()
mock_connect.assert_called_once_with('test_keyspace')
cluster = hook.get_cluster()
self.assertEqual(cluster.contact_points, ['host-1', 'host-2'])
self.assertEqual(cluster.port, 9042)
self.assertTrue(isinstance(cluster.load_balancing_policy, TokenAwarePolicy))
def test_get_conn(self):
with mock.patch.object(Cluster, "connect") as mock_connect, \
mock.patch("socket.getaddrinfo", return_value=[]) as mock_getaddrinfo:
mock_connect.return_value = 'session'
hook = CassandraHook(cassandra_conn_id='cassandra_test')
hook.get_conn()
mock_getaddrinfo.assert_called()
mock_connect.assert_called_once_with('test_keyspace')
cluster = hook.get_cluster()
self.assertEqual(cluster.contact_points, ['host-1', 'host-2'])
self.assertEqual(cluster.port, 9042)
self.assertTrue(isinstance(cluster.load_balancing_policy, TokenAwarePolicy))
def _assert_get_lb_policy(self, policy_name, policy_args, expected_policy_type,
expected_child_policy_type=None,
should_throw=False):
thrown = False
try:
policy = CassandraHook.get_lb_policy(policy_name, policy_args)
self.assertTrue(isinstance(policy, expected_policy_type))
if expected_child_policy_type:
self.assertTrue(isinstance(policy._child_policy,
expected_child_policy_type))
except Exception:
thrown = True
self.assertEqual(should_throw, thrown)
def _assert_get_lb_policy(self, policy_name, policy_args, expected_policy_type,
expected_child_policy_type=None,
should_throw=False):
thrown = False
try:
policy = CassandraHook.get_lb_policy(policy_name, policy_args)
self.assertTrue(isinstance(policy, expected_policy_type))
if expected_child_policy_type:
self.assertTrue(isinstance(policy._child_policy,
expected_child_policy_type))
except Exception:
thrown = True
self.assertEqual(should_throw, thrown)
def test_table_exists_with_keyspace_from_session(self):
hook = CassandraHook("cassandra_default_with_schema")
session = hook.get_conn()
cqls = [
"DROP TABLE IF EXISTS t",
"CREATE TABLE t (pk1 text PRIMARY KEY)",
]
for cql in cqls:
session.execute(cql)
self.assertTrue(hook.table_exists("t"))
self.assertFalse(hook.table_exists("u"))
session.shutdown()
hook.shutdown_cluster()
def test_record_exists_with_keyspace_from_session(self):
hook = CassandraHook("cassandra_default_with_schema")
session = hook.get_conn()
cqls = [
"DROP TABLE IF EXISTS t",
"CREATE TABLE t (pk1 text, pk2 text, c text, PRIMARY KEY (pk1, pk2))",
"INSERT INTO t (pk1, pk2, c) VALUES ('foo', 'bar', 'baz')",
]
for cql in cqls:
session.execute(cql)
self.assertTrue(hook.record_exists("t", {"pk1": "foo", "pk2": "bar"}))
self.assertFalse(hook.record_exists("t", {"pk1": "foo", "pk2": "baz"}))
session.shutdown()
hook.shutdown_cluster()
def test_table_exists_with_keyspace_from_session(self):
hook = CassandraHook("cassandra_default_with_schema")
session = hook.get_conn()
cqls = [
"DROP TABLE IF EXISTS t",
"CREATE TABLE t (pk1 text PRIMARY KEY)",
]
for cql in cqls:
session.execute(cql)
self.assertTrue(hook.table_exists("t"))
self.assertFalse(hook.table_exists("u"))
session.shutdown()
hook.shutdown_cluster()
def test_record_exists_with_keyspace_from_session(self):
hook = CassandraHook("cassandra_default_with_schema")
session = hook.get_conn()
cqls = [
"DROP TABLE IF EXISTS t",
"CREATE TABLE t (pk1 text, pk2 text, c text, PRIMARY KEY (pk1, pk2))",
"INSERT INTO t (pk1, pk2, c) VALUES ('foo', 'bar', 'baz')",
]
for cql in cqls:
session.execute(cql)
self.assertTrue(hook.record_exists("t", {"pk1": "foo", "pk2": "bar"}))
self.assertFalse(hook.record_exists("t", {"pk1": "foo", "pk2": "baz"}))
session.shutdown()
hook.shutdown_cluster()
def poke(self, context):
self.log.info('Sensor check existence of table: %s', self.table)
hook = CassandraHook(self.cassandra_conn_id)
return hook.table_exists(self.table)
def poke(self, context):
self.log.info('Sensor check existence of record: %s', self.keys)
hook = CassandraHook(self.cassandra_conn_id)
return hook.record_exists(self.table, self.keys)
def get_hook(self):
if self.conn_type == 'mysql':
from airflow.hooks.mysql_hook import MySqlHook
return MySqlHook(mysql_conn_id=self.conn_id)
elif self.conn_type == 'google_cloud_platform':
from airflow.gcp.hooks.bigquery import BigQueryHook
return BigQueryHook(bigquery_conn_id=self.conn_id)
elif self.conn_type == 'postgres':
from airflow.hooks.postgres_hook import PostgresHook
return PostgresHook(postgres_conn_id=self.conn_id)
elif self.conn_type == 'pig_cli':
from airflow.hooks.pig_hook import PigCliHook
return PigCliHook(pig_cli_conn_id=self.conn_id)
elif self.conn_type == 'hive_cli':
from airflow.hooks.hive_hooks import HiveCliHook
return HiveCliHook(hive_cli_conn_id=self.conn_id)
elif self.conn_type == 'presto':
from airflow.hooks.presto_hook import PrestoHook
return PrestoHook(presto_conn_id=self.conn_id)
elif self.conn_type == 'hiveserver2':
from airflow.hooks.hive_hooks import HiveServer2Hook
return HiveServer2Hook(hiveserver2_conn_id=self.conn_id)
elif self.conn_type == 'sqlite':
from airflow.hooks.sqlite_hook import SqliteHook
return SqliteHook(sqlite_conn_id=self.conn_id)
elif self.conn_type == 'jdbc':
from airflow.hooks.jdbc_hook import JdbcHook
return JdbcHook(jdbc_conn_id=self.conn_id)
elif self.conn_type == 'mssql':
from airflow.hooks.mssql_hook import MsSqlHook
return MsSqlHook(mssql_conn_id=self.conn_id)
elif self.conn_type == 'oracle':
from airflow.hooks.oracle_hook import OracleHook
return OracleHook(oracle_conn_id=self.conn_id)
elif self.conn_type == 'vertica':
from airflow.contrib.hooks.vertica_hook import VerticaHook
return VerticaHook(vertica_conn_id=self.conn_id)
elif self.conn_type == 'cloudant':
from airflow.contrib.hooks.cloudant_hook import CloudantHook
return CloudantHook(cloudant_conn_id=self.conn_id)
elif self.conn_type == 'jira':
from airflow.contrib.hooks.jira_hook import JiraHook
return JiraHook(jira_conn_id=self.conn_id)
elif self.conn_type == 'redis':
from airflow.contrib.hooks.redis_hook import RedisHook
return RedisHook(redis_conn_id=self.conn_id)
elif self.conn_type == 'wasb':
from airflow.contrib.hooks.wasb_hook import WasbHook
return WasbHook(wasb_conn_id=self.conn_id)
elif self.conn_type == 'docker':
from airflow.hooks.docker_hook import DockerHook
return DockerHook(docker_conn_id=self.conn_id)
elif self.conn_type == 'azure_data_lake':
from airflow.contrib.hooks.azure_data_lake_hook import AzureDataLakeHook
return AzureDataLakeHook(azure_data_lake_conn_id=self.conn_id)
elif self.conn_type == 'azure_cosmos':
from airflow.contrib.hooks.azure_cosmos_hook import AzureCosmosDBHook
return AzureCosmosDBHook(azure_cosmos_conn_id=self.conn_id)
elif self.conn_type == 'cassandra':
from airflow.contrib.hooks.cassandra_hook import CassandraHook
return CassandraHook(cassandra_conn_id=self.conn_id)
elif self.conn_type == 'mongo':
from airflow.contrib.hooks.mongo_hook import MongoHook
return MongoHook(conn_id=self.conn_id)
elif self.conn_type == 'gcpcloudsql':
from airflow.gcp.hooks.cloud_sql import CloudSqlDatabaseHook
return CloudSqlDatabaseHook(gcp_cloudsql_conn_id=self.conn_id)
elif self.conn_type == 'grpc':
from airflow.contrib.hooks.grpc_hook import GrpcHook
return GrpcHook(grpc_conn_id=self.conn_id)
raise AirflowException("Unknown hook type {}".format(self.conn_type))
class CassandraToGoogleCloudStorageOperator(BaseOperator):
"""
Copy data from Cassandra to Google cloud storage in JSON format
Note: Arrays of arrays are not supported.
"""
template_fields = ('cql', 'bucket', 'filename', 'schema_filename',)
template_ext = ('.cql',)
ui_color = '#a0e08c'
@apply_defaults
def __init__(self,
cql,
bucket,
filename,
schema_filename=None,
approx_max_file_size_bytes=1900000000,
cassandra_conn_id='cassandra_default',
google_cloud_storage_conn_id='google_cloud_default',
delegate_to=None,
*args,
**kwargs):
"""
:param cql: The CQL to execute on the Cassandra table.
:type cql: str
:param bucket: The bucket to upload to.
:type bucket: str
:param filename: The filename to use as the object name when uploading
to Google cloud storage. A {} should be specified in the filename
to allow the operator to inject file numbers in cases where the
file is split due to size.
:type filename: str
:param schema_filename: If set, the filename to use as the object name
when uploading a .json file containing the BigQuery schema fields
for the table that was dumped from MySQL.
:type schema_filename: str
:param approx_max_file_size_bytes: This operator supports the ability
to split large table dumps into multiple files (see notes in the
filenamed param docs above). Google cloud storage allows for files
to be a maximum of 4GB. This param allows developers to specify the
file size of the splits.
:type approx_max_file_size_bytes: long
:param cassandra_conn_id: Reference to a specific Cassandra hook.
:type cassandra_conn_id: str
:param google_cloud_storage_conn_id: Reference to a specific Google
cloud storage hook.
:type google_cloud_storage_conn_id: str
:param delegate_to: The account to impersonate, if any. For this to
work, the service account making the request must have domain-wide
delegation enabled.
:type delegate_to: str
"""
super(CassandraToGoogleCloudStorageOperator, self).__init__(*args, **kwargs)
self.cql = cql
self.bucket = bucket
self.filename = filename
self.schema_filename = schema_filename
self.approx_max_file_size_bytes = approx_max_file_size_bytes
self.cassandra_conn_id = cassandra_conn_id
self.google_cloud_storage_conn_id = google_cloud_storage_conn_id
self.delegate_to = delegate_to
self.hook = None
# Default Cassandra to BigQuery type mapping
CQL_TYPE_MAP = {
'BytesType': 'BYTES',
'DecimalType': 'FLOAT',
'UUIDType': 'BYTES',
'BooleanType': 'BOOL',
'ByteType': 'INTEGER',
'AsciiType': 'STRING',
'FloatType': 'FLOAT',
'DoubleType': 'FLOAT',
'LongType': 'INTEGER',
'Int32Type': 'INTEGER',
'IntegerType': 'INTEGER',
'InetAddressType': 'STRING',
'CounterColumnType': 'INTEGER',
'DateType': 'TIMESTAMP',
'SimpleDateType': 'DATE',
'TimestampType': 'TIMESTAMP',
'TimeUUIDType': 'BYTES',
'ShortType': 'INTEGER',
'TimeType': 'TIME',
'DurationType': 'INTEGER',
'UTF8Type': 'STRING',
'VarcharType': 'STRING',
}
def execute(self, context):
cursor = self._query_cassandra()
files_to_upload = self._write_local_data_files(cursor)
# If a schema is set, create a BQ schema JSON file.
if self.schema_filename:
files_to_upload.update(self._write_local_schema_file(cursor))
# Flush all files before uploading
for file_handle in files_to_upload.values():
file_handle.flush()
self._upload_to_gcs(files_to_upload)
# Close all temp file handles.
for file_handle in files_to_upload.values():
file_handle.close()
# Close all sessions and connection associated with this Cassandra cluster
self.hook.shutdown_cluster()
def _query_cassandra(self):
"""
Queries cassandra and returns a cursor to the results.
"""
self.hook = CassandraHook(cassandra_conn_id=self.cassandra_conn_id)
session = self.hook.get_conn()
cursor = session.execute(self.cql)
return cursor
def _write_local_data_files(self, cursor):
"""
Takes a cursor, and writes results to a local file.
:return: A dictionary where keys are filenames to be used as object
names in GCS, and values are file handles to local files that
contain the data for the GCS objects.
"""
file_no = 0
tmp_file_handle = NamedTemporaryFile(delete=True)
tmp_file_handles = {self.filename.format(file_no): tmp_file_handle}
for row in cursor:
row_dict = self.generate_data_dict(row._fields, row)
s = json.dumps(row_dict)
if PY3:
s = s.encode('utf-8')
tmp_file_handle.write(s)
# Append newline to make dumps BigQuery compatible.
tmp_file_handle.write(b'\n')
if tmp_file_handle.tell() >= self.approx_max_file_size_bytes:
file_no += 1
tmp_file_handle = NamedTemporaryFile(delete=True)
tmp_file_handles[self.filename.format(file_no)] = tmp_file_handle
return tmp_file_handles
def _write_local_schema_file(self, cursor):
"""
Takes a cursor, and writes the BigQuery schema for the results to a
local file system.
:return: A dictionary where key is a filename to be used as an object
name in GCS, and values are file handles to local files that
contains the BigQuery schema fields in .json format.
"""
schema = []
tmp_schema_file_handle = NamedTemporaryFile(delete=True)
for name, type in zip(cursor.column_names, cursor.column_types):
schema.append(self.generate_schema_dict(name, type))
json_serialized_schema = json.dumps(schema)
if PY3:
json_serialized_schema = json_serialized_schema.encode('utf-8')
tmp_schema_file_handle.write(json_serialized_schema)
return {self.schema_filename: tmp_schema_file_handle}
def _upload_to_gcs(self, files_to_upload):
hook = GoogleCloudStorageHook(
google_cloud_storage_conn_id=self.google_cloud_storage_conn_id,
delegate_to=self.delegate_to)
for object, tmp_file_handle in files_to_upload.items():
hook.upload(self.bucket, object, tmp_file_handle.name, 'application/json')
@classmethod
def generate_data_dict(cls, names, values):
row_dict = {}
for name, value in zip(names, values):
row_dict.update({name: cls.convert_value(name, value)})
return row_dict
@classmethod
def convert_value(cls, name, value):
if not value:
return value
elif isinstance(value, (text_type, int, float, bool, dict)):
return value
elif isinstance(value, binary_type):
return b64encode(value).decode('ascii')
elif isinstance(value, UUID):
return b64encode(value.bytes).decode('ascii')
elif isinstance(value, (datetime, Date)):
return str(value)
elif isinstance(value, Decimal):
return float(value)
elif isinstance(value, Time):
return str(value).split('.')[0]
elif isinstance(value, (list, SortedSet)):
return cls.convert_array_types(name, value)
elif hasattr(value, '_fields'):
return cls.convert_user_type(name, value)
elif isinstance(value, tuple):
return cls.convert_tuple_type(name, value)
elif isinstance(value, OrderedMapSerializedKey):
return cls.convert_map_type(name, value)
else:
raise AirflowException('unexpected value: ' + str(value))
@classmethod
def convert_array_types(cls, name, value):
return [cls.convert_value(name, nested_value) for nested_value in value]
@classmethod
def convert_user_type(cls, name, value):
"""
Converts a user type to RECORD that contains n fields, where n is the
number of attributes. Each element in the user type class will be converted to its
corresponding data type in BQ.
"""
names = value._fields
values = [cls.convert_value(name, getattr(value, name)) for name in names]
return cls.generate_data_dict(names, values)
@classmethod
def convert_tuple_type(cls, name, value):
"""
Converts a tuple to RECORD that contains n fields, each will be converted
to its corresponding data type in bq and will be named 'field_<index>', where
index is determined by the order of the tuple elements defined in cassandra.
"""
names = ['field_' + str(i) for i in range(len(value))]
values = [cls.convert_value(name, value) for name, value in zip(names, value)]
return cls.generate_data_dict(names, values)
@classmethod
def convert_map_type(cls, name, value):
"""
Converts a map to a repeated RECORD that contains two fields: 'key' and 'value',
each will be converted to its corresponding data type in BQ.
"""
converted_map = []
for k, v in zip(value.keys(), value.values()):
converted_map.append({
'key': cls.convert_value('key', k),
'value': cls.convert_value('value', v)
})
return converted_map
@classmethod
def generate_schema_dict(cls, name, type):
field_schema = dict()
field_schema.update({'name': name})
field_schema.update({'type': cls.get_bq_type(type)})
field_schema.update({'mode': cls.get_bq_mode(type)})
fields = cls.get_bq_fields(name, type)
if fields:
field_schema.update({'fields': fields})
return field_schema
@classmethod
def get_bq_fields(cls, name, type):
fields = []
if not cls.is_simple_type(type):
names, types = [], []
if cls.is_array_type(type) and cls.is_record_type(type.subtypes[0]):
names = type.subtypes[0].fieldnames
types = type.subtypes[0].subtypes
elif cls.is_record_type(type):
names = type.fieldnames
types = type.subtypes
if types and not names and type.cassname == 'TupleType':
names = ['field_' + str(i) for i in range(len(types))]
elif types and not names and type.cassname == 'MapType':
names = ['key', 'value']
for name, type in zip(names, types):
field = cls.generate_schema_dict(name, type)
fields.append(field)
return fields
@classmethod
def is_simple_type(cls, type):
return type.cassname in CassandraToGoogleCloudStorageOperator.CQL_TYPE_MAP
@classmethod
def is_array_type(cls, type):
return type.cassname in ['ListType', 'SetType']
@classmethod
def is_record_type(cls, type):
return type.cassname in ['UserType', 'TupleType', 'MapType']
@classmethod
def get_bq_type(cls, type):
if cls.is_simple_type(type):
return CassandraToGoogleCloudStorageOperator.CQL_TYPE_MAP[type.cassname]
elif cls.is_record_type(type):
return 'RECORD'
elif cls.is_array_type(type):
return cls.get_bq_type(type.subtypes[0])
else:
raise AirflowException('Not a supported type: ' + type.cassname)
@classmethod
def get_bq_mode(cls, type):
if cls.is_array_type(type) or type.cassname == 'MapType':
return 'REPEATED'
elif cls.is_record_type(type) or cls.is_simple_type(type):
return 'NULLABLE'
else:
raise AirflowException('Not a supported type: ' + type.cassname)
class CassandraToGoogleCloudStorageOperator(BaseOperator):
"""
Copy data from Cassandra to Google cloud storage in JSON format
Note: Arrays of arrays are not supported.
:param cql: The CQL to execute on the Cassandra table.
:type cql: str
:param bucket: The bucket to upload to.
:type bucket: str
:param filename: The filename to use as the object name when uploading
to Google cloud storage. A {} should be specified in the filename
to allow the operator to inject file numbers in cases where the
file is split due to size.
:type filename: str
:param schema_filename: If set, the filename to use as the object name
when uploading a .json file containing the BigQuery schema fields
for the table that was dumped from MySQL.
:type schema_filename: str
:param approx_max_file_size_bytes: This operator supports the ability
to split large table dumps into multiple files (see notes in the
filename param docs above). This param allows developers to specify the
file size of the splits. Check https://cloud.google.com/storage/quotas
to see the maximum allowed file size for a single object.
:type approx_max_file_size_bytes: long
:param cassandra_conn_id: Reference to a specific Cassandra hook.
:type cassandra_conn_id: str
:param gzip: Option to compress file for upload
:type gzip: bool
:param google_cloud_storage_conn_id: Reference to a specific Google
cloud storage hook.
:type google_cloud_storage_conn_id: str
:param delegate_to: The account to impersonate, if any. For this to
work, the service account making the request must have domain-wide
delegation enabled.
:type delegate_to: str
"""
template_fields = (
'cql',
'bucket',
'filename',
'schema_filename',
)
template_ext = ('.cql', )
ui_color = '#a0e08c'
@apply_defaults
def __init__(self,
cql,
bucket,
filename,
schema_filename=None,
approx_max_file_size_bytes=1900000000,
gzip=False,
cassandra_conn_id='cassandra_default',
google_cloud_storage_conn_id='google_cloud_default',
delegate_to=None,
*args,
**kwargs):
super(CassandraToGoogleCloudStorageOperator,
self).__init__(*args, **kwargs)
self.cql = cql
self.bucket = bucket
self.filename = filename
self.schema_filename = schema_filename
self.approx_max_file_size_bytes = approx_max_file_size_bytes
self.cassandra_conn_id = cassandra_conn_id
self.google_cloud_storage_conn_id = google_cloud_storage_conn_id
self.delegate_to = delegate_to
self.gzip = gzip
self.hook = None
# Default Cassandra to BigQuery type mapping
CQL_TYPE_MAP = {
'BytesType': 'BYTES',
'DecimalType': 'FLOAT',
'UUIDType': 'BYTES',
'BooleanType': 'BOOL',
'ByteType': 'INTEGER',
'AsciiType': 'STRING',
'FloatType': 'FLOAT',
'DoubleType': 'FLOAT',
'LongType': 'INTEGER',
'Int32Type': 'INTEGER',
'IntegerType': 'INTEGER',
'InetAddressType': 'STRING',
'CounterColumnType': 'INTEGER',
'DateType': 'TIMESTAMP',
'SimpleDateType': 'DATE',
'TimestampType': 'TIMESTAMP',
'TimeUUIDType': 'BYTES',
'ShortType': 'INTEGER',
'TimeType': 'TIME',
'DurationType': 'INTEGER',
'UTF8Type': 'STRING',
'VarcharType': 'STRING',
}
def execute(self, context):
cursor = self._query_cassandra()
files_to_upload = self._write_local_data_files(cursor)
# If a schema is set, create a BQ schema JSON file.
if self.schema_filename:
files_to_upload.update(self._write_local_schema_file(cursor))
# Flush all files before uploading
for file_handle in files_to_upload.values():
file_handle.flush()
self._upload_to_gcs(files_to_upload)
# Close all temp file handles.
for file_handle in files_to_upload.values():
file_handle.close()
# Close all sessions and connection associated with this Cassandra cluster
self.hook.shutdown_cluster()
def _query_cassandra(self):
"""
Queries cassandra and returns a cursor to the results.
"""
self.hook = CassandraHook(cassandra_conn_id=self.cassandra_conn_id)
session = self.hook.get_conn()
cursor = session.execute(self.cql)
return cursor
def _write_local_data_files(self, cursor):
"""
Takes a cursor, and writes results to a local file.
:return: A dictionary where keys are filenames to be used as object
names in GCS, and values are file handles to local files that
contain the data for the GCS objects.
"""
file_no = 0
tmp_file_handle = NamedTemporaryFile(delete=True)
tmp_file_handles = {self.filename.format(file_no): tmp_file_handle}
for row in cursor:
row_dict = self.generate_data_dict(row._fields, row)
s = json.dumps(row_dict)
if PY3:
s = s.encode('utf-8')
tmp_file_handle.write(s)
# Append newline to make dumps BigQuery compatible.
tmp_file_handle.write(b'\n')
if tmp_file_handle.tell() >= self.approx_max_file_size_bytes:
file_no += 1
tmp_file_handle = NamedTemporaryFile(delete=True)
tmp_file_handles[self.filename.format(
file_no)] = tmp_file_handle
return tmp_file_handles
def _write_local_schema_file(self, cursor):
"""
Takes a cursor, and writes the BigQuery schema for the results to a
local file system.
:return: A dictionary where key is a filename to be used as an object
name in GCS, and values are file handles to local files that
contains the BigQuery schema fields in .json format.
"""
schema = []
tmp_schema_file_handle = NamedTemporaryFile(delete=True)
for name, type in zip(cursor.column_names, cursor.column_types):
schema.append(self.generate_schema_dict(name, type))
json_serialized_schema = json.dumps(schema)
if PY3:
json_serialized_schema = json_serialized_schema.encode('utf-8')
tmp_schema_file_handle.write(json_serialized_schema)
return {self.schema_filename: tmp_schema_file_handle}
def _upload_to_gcs(self, files_to_upload):
hook = GoogleCloudStorageHook(
google_cloud_storage_conn_id=self.google_cloud_storage_conn_id,
delegate_to=self.delegate_to)
for object, tmp_file_handle in files_to_upload.items():
hook.upload(self.bucket, object, tmp_file_handle.name,
'application/json', self.gzip)
@classmethod
def generate_data_dict(cls, names, values):
row_dict = {}
for name, value in zip(names, values):
row_dict.update({name: cls.convert_value(name, value)})
return row_dict
@classmethod
def convert_value(cls, name, value):
if not value:
return value
elif isinstance(value, (text_type, int, float, bool, dict)):
return value
elif isinstance(value, binary_type):
return b64encode(value).decode('ascii')
elif isinstance(value, UUID):
return b64encode(value.bytes).decode('ascii')
elif isinstance(value, (datetime, Date)):
return str(value)
elif isinstance(value, Decimal):
return float(value)
elif isinstance(value, Time):
return str(value).split('.')[0]
elif isinstance(value, (list, SortedSet)):
return cls.convert_array_types(name, value)
elif hasattr(value, '_fields'):
return cls.convert_user_type(name, value)
elif isinstance(value, tuple):
return cls.convert_tuple_type(name, value)
elif isinstance(value, OrderedMapSerializedKey):
return cls.convert_map_type(name, value)
else:
raise AirflowException('unexpected value: ' + str(value))
@classmethod
def convert_array_types(cls, name, value):
return [
cls.convert_value(name, nested_value) for nested_value in value
]
@classmethod
def convert_user_type(cls, name, value):
"""
Converts a user type to RECORD that contains n fields, where n is the
number of attributes. Each element in the user type class will be converted to its
corresponding data type in BQ.
"""
names = value._fields
values = [
cls.convert_value(name, getattr(value, name)) for name in names
]
return cls.generate_data_dict(names, values)
@classmethod
def convert_tuple_type(cls, name, value):
"""
Converts a tuple to RECORD that contains n fields, each will be converted
to its corresponding data type in bq and will be named 'field_<index>', where
index is determined by the order of the tuple elements defined in cassandra.
"""
names = ['field_' + str(i) for i in range(len(value))]
values = [
cls.convert_value(name, value)
for name, value in zip(names, value)
]
return cls.generate_data_dict(names, values)
@classmethod
def convert_map_type(cls, name, value):
"""
Converts a map to a repeated RECORD that contains two fields: 'key' and 'value',
each will be converted to its corresponding data type in BQ.
"""
converted_map = []
for k, v in zip(value.keys(), value.values()):
converted_map.append({
'key': cls.convert_value('key', k),
'value': cls.convert_value('value', v)
})
return converted_map
@classmethod
def generate_schema_dict(cls, name, type):
field_schema = dict()
field_schema.update({'name': name})
field_schema.update({'type': cls.get_bq_type(type)})
field_schema.update({'mode': cls.get_bq_mode(type)})
fields = cls.get_bq_fields(name, type)
if fields:
field_schema.update({'fields': fields})
return field_schema
@classmethod
def get_bq_fields(cls, name, type):
fields = []
if not cls.is_simple_type(type):
names, types = [], []
if cls.is_array_type(type) and cls.is_record_type(
type.subtypes[0]):
names = type.subtypes[0].fieldnames
types = type.subtypes[0].subtypes
elif cls.is_record_type(type):
names = type.fieldnames
types = type.subtypes
if types and not names and type.cassname == 'TupleType':
names = ['field_' + str(i) for i in range(len(types))]
elif types and not names and type.cassname == 'MapType':
names = ['key', 'value']
for name, type in zip(names, types):
field = cls.generate_schema_dict(name, type)
fields.append(field)
return fields
@classmethod
def is_simple_type(cls, type):
return type.cassname in CassandraToGoogleCloudStorageOperator.CQL_TYPE_MAP
@classmethod
def is_array_type(cls, type):
return type.cassname in ['ListType', 'SetType']
@classmethod
def is_record_type(cls, type):
return type.cassname in ['UserType', 'TupleType', 'MapType']
@classmethod
def get_bq_type(cls, type):
if cls.is_simple_type(type):
return CassandraToGoogleCloudStorageOperator.CQL_TYPE_MAP[
type.cassname]
elif cls.is_record_type(type):
return 'RECORD'
elif cls.is_array_type(type):
return cls.get_bq_type(type.subtypes[0])
else:
raise AirflowException('Not a supported type: ' + type.cassname)
@classmethod
def get_bq_mode(cls, type):
if cls.is_array_type(type) or type.cassname == 'MapType':
return 'REPEATED'
elif cls.is_record_type(type) or cls.is_simple_type(type):
return 'NULLABLE'
else:
raise AirflowException('Not a supported type: ' + type.cassname)
def poke(self, context):
self.log.info('Sensor check existence of record: %s', self.keys)
hook = CassandraHook(self.cassandra_conn_id)
return hook.record_exists(self.table, self.keys)
def poke(self, context):
self.log.info('Sensor check existence of table: %s', self.table)
hook = CassandraHook(self.cassandra_conn_id)
return hook.table_exists(self.table)