Connections aim to ease the use of multiple sessions with cqlengine. Connections can be set on a model class, per query or using a context manager.
To use cqlengine, you need at least a default connection. If you initialize cqlengine’s connections with with
connection.setup, a connection will be created automatically. If you want to use another cluster/session, you need to register a new cqlengine connection. You register a connection with
from cassandra.cqlengine import connection connection.setup(['127.0.0.1') connection.register_connection('cluster2', ['127.0.0.2'])
register_connection() can take a list of hosts, as shown above, in which case it will create a connection with a new session. It can also take a session argument if you’ve already created a session:
from cassandra.cqlengine import connection from cassandra.cluster import Cluster session = Cluster(['127.0.0.1']).connect() connection.register_connection('cluster3', session=session)
You can change the default cqlengine connection on registration:
from cassandra.cqlengine import connection connection.register_connection('cluster2', ['127.0.0.2'] default=True)
or on the fly using
You can unregister a connection using
When using multiples connections, you also need to sync your models on all connections (and keyspaces) that you need operate on. Management commands have been improved to ease this part. Here is an example:
from cassandra.cqlengine import management keyspaces = ['ks1', 'ks2'] conns = ['cluster1', 'cluster2'] # registers your connections # ... # create all keyspaces on all connections for ks in keyspaces: management.create_simple_keyspace(ks, connections=conns) # define your Automobile model # ... # sync your models management.sync_table(Automobile, keyspaces=keyspaces, connections=conns)
cqlengine will select the default connection, unless your specify a connection using one of the following methods.
You can specify a default connection per model:
class Automobile(Model): __keyspace__ = 'test' __connection__ = 'cluster2' manufacturer = columns.Text(primary_key=True) year = columns.Integer(primary_key=True) model = columns.Text(primary_key=True) print len(Automobile.objects.all()) # executed on the connection 'cluster2'
You can use the
using() method to select a connection (or keyspace):
Automobile.objects.using(connection='cluster1').create(manufacturer='honda', year=2010, model='civic') q = Automobile.objects.filter(manufacturer='Tesla') autos = q.using(keyspace='ks2', connection='cluster2').all() for auto in autos: auto.using(connection='cluster1').save()
You can use the ContextQuery as well to select a connection:
with ContextQuery(Automobile, connection='cluster1') as A: A.objects.filter(manufacturer='honda').all() # executed on 'cluster1'
With a BatchQuery, you can select the connection with the context manager. Note that all operations in the batch need to use the same connection.
with BatchQuery(connection='cluster1') as b: Automobile.objects.batch(b).create(manufacturer='honda', year=2010, model='civic')