/usr/lib/python2.7/dist-packages/pycassa/batch.py is in python-pycassa 1.11.2.1-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 | """
The batch interface allows insert, update, and remove operations to be performed
in batches. This allows a convenient mechanism for streaming updates or doing a
large number of operations while reducing number of RPC roundtrips.
Batch mutator objects are synchronized and can be safely passed around threads.
.. code-block:: python
>>> b = cf.batch(queue_size=10)
>>> b.insert('key1', {'col1':'value11', 'col2':'value21'})
>>> b.insert('key2', {'col1':'value12', 'col2':'value22'}, ttl=15)
>>> b.remove('key1', ['col2'])
>>> b.remove('key2')
>>> b.send()
One can use the `queue_size` argument to control how many mutations will be
queued before an automatic :meth:`send` is performed. This allows simple streaming
of updates. If set to ``None``, automatic checkpoints are disabled. Default is 100.
Supercolumns are supported:
.. code-block:: python
>>> b = scf.batch()
>>> b.insert('key1', {'supercol1': {'colA':'value1a', 'colB':'value1b'}
... {'supercol2': {'colA':'value2a', 'colB':'value2b'}})
>>> b.remove('key1', ['colA'], 'supercol1')
>>> b.send()
You may also create a :class:`.Mutator` directly, allowing operations
on multiple column families:
.. code-block:: python
>>> b = Mutator(pool)
>>> b.insert(cf, 'key1', {'col1':'value1', 'col2':'value2'})
>>> b.insert(supercf, 'key1', {'subkey1': {'col1':'value1', 'col2':'value2'}})
>>> b.send()
.. note:: This interface does not implement atomic operations across column
families. All the limitations of the `batch_mutate` Thrift API call
applies. Remember, a mutation in Cassandra is always atomic per key per
column family only.
.. note:: If a single operation in a batch fails, the whole batch fails.
In addition mutators can be used as context managers, where an implicit
:meth:`send` will be called upon exit.
.. code-block:: python
>>> with cf.batch() as b:
... b.insert('key1', {'col1':'value11', 'col2':'value21'})
... b.insert('key2', {'col1':'value12', 'col2':'value22'})
Calls to :meth:`insert` and :meth:`remove` can also be chained:
.. code-block:: python
>>> cf.batch().remove('foo').remove('bar').send()
To use atomic batches (supported in Cassandra 1.2 and later), pass the atomic
option in when creating the batch:
.. code-block:: python
>>> cf.batch(atomic=True)
or when sending it:
.. code-block:: python
>>> b = cf.batch()
>>> b.insert('key1', {'col1':'val2'})
>>> b.insert('key2', {'col1':'val2'})
>>> b.send(atomic=True)
"""
import threading
from pycassa.cassandra.ttypes import (ConsistencyLevel, Deletion, Mutation, SlicePredicate)
__all__ = ['Mutator', 'CfMutator']
class Mutator(object):
"""
Batch update convenience mechanism.
Queues insert/update/remove operations and executes them when the queue
is full or `send` is called explicitly.
"""
def __init__(self, pool, queue_size=100, write_consistency_level=None, allow_retries=True, atomic=False):
"""
`pool` is the :class:`~pycassa.pool.ConnectionPool` that will be used
for operations.
After `queue_size` operations, :meth:`send()` will be executed
automatically. Use 0 to disable automatic sends.
"""
self._buffer = []
self._lock = threading.RLock()
self.pool = pool
self.limit = queue_size
self.allow_retries = allow_retries
self.atomic = atomic
if write_consistency_level is None:
self.write_consistency_level = ConsistencyLevel.ONE
else:
self.write_consistency_level = write_consistency_level
def __enter__(self):
return self
def __exit__(self, exc_type, exc_value, traceback):
self.send()
def _enqueue(self, key, column_family, mutations):
self._lock.acquire()
try:
mutation = (key, column_family.column_family, mutations)
self._buffer.append(mutation)
if self.limit and len(self._buffer) >= self.limit:
self.send()
finally:
self._lock.release()
return self
def send(self, write_consistency_level=None, atomic=None):
""" Sends all operations currently in the batch and clears the batch. """
if write_consistency_level is None:
write_consistency_level = self.write_consistency_level
if atomic is None:
atomic = self.atomic
mutations = {}
conn = None
self._lock.acquire()
try:
for key, column_family, cols in self._buffer:
mutations.setdefault(key, {}).setdefault(column_family, []).extend(cols)
if mutations:
conn = self.pool.get()
mutatefn = conn.atomic_batch_mutate if atomic else conn.batch_mutate
mutatefn(mutations, write_consistency_level,
allow_retries=self.allow_retries)
self._buffer = []
finally:
if conn:
conn.return_to_pool()
self._lock.release()
def insert(self, column_family, key, columns, timestamp=None, ttl=None):
"""
Adds a single row insert to the batch.
`column_family` is the :class:`~pycassa.columnfamily.ColumnFamily`
that the insert will be executed on.
If this is used on a counter column family, integers may be used for
column values, and they will be taken as counter adjustments.
"""
if columns:
if timestamp is None:
timestamp = column_family.timestamp()
packed_key = column_family._pack_key(key)
mut_list = column_family._make_mutation_list(columns, timestamp, ttl)
self._enqueue(packed_key, column_family, mut_list)
return self
def remove(self, column_family, key, columns=None, super_column=None, timestamp=None):
"""
Adds a single row remove to the batch.
`column_family` is the :class:`~pycassa.columnfamily.ColumnFamily`
that the remove will be executed on.
"""
if timestamp is None:
timestamp = column_family.timestamp()
deletion = Deletion(timestamp=timestamp)
_pack_name = column_family._pack_name
if super_column is not None:
deletion.super_column = _pack_name(super_column, True)
if columns is not None:
is_super = column_family.super and super_column is None
packed_cols = [_pack_name(col, is_super) for col in columns]
deletion.predicate = SlicePredicate(column_names=packed_cols)
mutation = Mutation(deletion=deletion)
packed_key = column_family._pack_key(key)
self._enqueue(packed_key, column_family, (mutation,))
return self
class CfMutator(Mutator):
"""
A :class:`~pycassa.batch.Mutator` that deals only with one column family.
"""
def __init__(self, column_family, queue_size=100, write_consistency_level=None,
allow_retries=True, atomic=False):
"""
`column_family` is the :class:`~pycassa.columnfamily.ColumnFamily`
that all operations will be executed on.
"""
wcl = write_consistency_level or column_family.write_consistency_level
Mutator.__init__(self, column_family.pool, queue_size, wcl, allow_retries, atomic)
self._column_family = column_family
def insert(self, key, cols, timestamp=None, ttl=None):
""" Adds a single row insert to the batch. """
return Mutator.insert(self, self._column_family, key, cols, timestamp, ttl)
def remove(self, key, columns=None, super_column=None, timestamp=None):
""" Adds a single row remove to the batch. """
return Mutator.remove(self, self._column_family, key,
columns, super_column, timestamp)
|