def _load_comparator_classes(self):
if not self.super:
self.column_name_class = util.extract_type_name(self._cfdef.comparator_type)
self.super_column_name_class = None
else:
self.column_name_class = util.extract_type_name(self._cfdef.subcomparator_type)
self.super_column_name_class = util.extract_type_name(self._cfdef.comparator_type)
def _load_comparator_classes(self):
if not self.super:
self.column_name_class = util.extract_type_name(self._cfdef.comparator_type)
self.super_column_name_class = None
else:
self.column_name_class = util.extract_type_name(self._cfdef.subcomparator_type)
self.super_column_name_class = util.extract_type_name(self._cfdef.comparator_type)
def _load_validation_classes(self):
self.default_validation_class = util.extract_type_name(
self._cfdef.default_validation_class)
self.column_validators = {}
for name, coldef in self._cfdef.column_metadata.items():
self.column_validators[name] = util.extract_type_name(
coldef.validation_class)
def alter_column(self, keyspace, column_family, column, value_type):
"""
Sets a data type for the value of a specific column.
`value_type` is a string that determines what type the column value will be.
By default, :const:`LONG_TYPE`, :const:`INT_TYPE`,
:const:`ASCII_TYPE`, :const:`UTF8_TYPE`, :const:`TIME_UUID_TYPE`,
:const:`LEXICAL_UUID_TYPE` and :const:`BYTES_TYPE` are provided. Custom
types may be used as well by providing the class name; if the custom
comparator class is not in ``org.apache.cassandra.db.marshal``, the fully
qualified class name must be given.
For super column families, this sets the subcolumn value type for
any subcolumn named `column`, regardless of the super column name.
"""
self._conn.set_keyspace(keyspace)
cfdef = self.get_keyspace_column_families(keyspace)[column_family]
if cfdef.column_type == 'Super':
col_name_data_type = util.extract_type_name(
cfdef.subcomparator_type)
else:
col_name_data_type = util.extract_type_name(cfdef.comparator_type)
packed_column = util.pack(column, col_name_data_type)
if value_type.find('.') == -1:
value_type = 'org.apache.cassandra.db.marshal.%s' % value_type
matched = False
for c in cfdef.column_metadata:
if c.name == packed_column:
c.validation_class = value_type
matched = True
break
if not matched:
cfdef.column_metadata.append(
ColumnDef(packed_column, value_type, None, None))
self._system_update_column_family(cfdef)
def alter_column(self, keyspace, column_family, column, value_type):
"""
Sets a data type for the value of a specific column.
`value_type` is a string that determines what type the column value will be.
By default, :const:`LONG_TYPE`, :const:`INT_TYPE`,
:const:`ASCII_TYPE`, :const:`UTF8_TYPE`, :const:`TIME_UUID_TYPE`,
:const:`LEXICAL_UUID_TYPE` and :const:`BYTES_TYPE` are provided. Custom
types may be used as well by providing the class name; if the custom
comparator class is not in ``org.apache.cassandra.db.marshal``, the fully
qualified class name must be given.
For super column families, this sets the subcolumn value type for
any subcolumn named `column`, regardless of the super column name.
"""
self._conn.set_keyspace(keyspace)
cfdef = self.get_keyspace_column_families(keyspace)[column_family]
if cfdef.column_type == 'Super':
col_name_data_type = util.extract_type_name(cfdef.subcomparator_type)
else:
col_name_data_type = util.extract_type_name(cfdef.comparator_type)
packed_column = util.pack(column, col_name_data_type)
if value_type.find('.') == -1:
value_type = 'org.apache.cassandra.db.marshal.%s' % value_type
matched = False
for c in cfdef.column_metadata:
if c.name == packed_column:
c.validation_class = value_type
matched = True
break
if not matched:
cfdef.column_metadata.append(ColumnDef(packed_column, value_type, None, None))
self._system_update_column_family(cfdef)
def create_index(self,
keyspace,
column_family,
column,
value_type,
index_type=KEYS_INDEX,
index_name=None):
"""
Creates an index on a column.
This allows efficient for index usage via
:meth:`~pycassa.columnfamily.ColumnFamily.get_indexed_slices()`
`column` specifies what column to index, and `value_type` is a string
that describes that column's value's data type; see
:meth:`alter_column()` for a full description of `value_type`.
`index_type` determines how the index will be stored internally. Currently,
:const:`KEYS_INDEX` is the only option. `index_name` is an optional name
for the index.
Example Usage:
.. code-block:: python
>>> from pycassa.system_manager import *
>>> sys = SystemManager('192.168.2.10:9160')
>>> sys.create_index('Keyspace1', 'Standard1', 'birthdate', LONG_TYPE, index_name='bday_index')
>>> sys.close
"""
self._conn.set_keyspace(keyspace)
cfdef = self.get_keyspace_column_families(keyspace)[column_family]
col_name_data_type = util.extract_type_name(cfdef.comparator_type)
packed_column = util.pack(column, col_name_data_type)
if value_type.find('.') == -1:
value_type = 'org.apache.cassandra.db.marshal.%s' % value_type
coldef = ColumnDef(packed_column, value_type, index_type, index_name)
for c in cfdef.column_metadata:
if c.name == packed_column:
cfdef.column_metadata.remove(c)
break
cfdef.column_metadata.append(coldef)
self._system_update_column_family(cfdef)
def create_index(self, keyspace, column_family, column, value_type,
index_type=KEYS_INDEX, index_name=None):
"""
Creates an index on a column.
This allows efficient for index usage via
:meth:`~pycassa.columnfamily.ColumnFamily.get_indexed_slices()`
`column` specifies what column to index, and `value_type` is a string
that describes that column's value's data type; see
:meth:`alter_column()` for a full description of `value_type`.
`index_type` determines how the index will be stored internally. Currently,
:const:`KEYS_INDEX` is the only option. `index_name` is an optional name
for the index.
Example Usage:
.. code-block:: python
>>> from pycassa.system_manager import *
>>> sys = SystemManager('192.168.2.10:9160')
>>> sys.create_index('Keyspace1', 'Standard1', 'birthdate', LONG_TYPE, index_name='bday_index')
>>> sys.close
"""
self._conn.set_keyspace(keyspace)
cfdef = self.get_keyspace_column_families(keyspace)[column_family]
col_name_data_type = util.extract_type_name(cfdef.comparator_type)
packed_column = util.pack(column, col_name_data_type)
if value_type.find('.') == -1:
value_type = 'org.apache.cassandra.db.marshal.%s' % value_type
coldef = ColumnDef(packed_column, value_type, index_type, index_name)
for c in cfdef.column_metadata:
if c.name == packed_column:
cfdef.column_metadata.remove(c)
break
cfdef.column_metadata.append(coldef)
self._system_update_column_family(cfdef)
def __init__(self, pool, column_family, buffer_size=1024,
read_consistency_level=ConsistencyLevel.ONE,
write_consistency_level=ConsistencyLevel.ONE,
timestamp=gm_timestamp, super=False,
dict_class=util.OrderedDict, autopack_names=True,
autopack_values=True):
"""
An abstraction of a Cassandra column family or super column family.
Operations on this, such as :meth:`get` or :meth:`insert` will get data from or
insert data into the corresponding Cassandra column family with
name `column_family`.
`pool` is a :class:`~pycassa.pool.ConnectionPool` that the column
family will use for all operations. A connection is drawn from the
pool before each operations and is returned afterwards. Note that
the keyspace to be used is determined by the pool.
When calling :meth:`get_range()` or :meth:`get_indexed_slices()`,
the intermediate results need to be buffered if we are fetching many
rows, otherwise the Cassandra server will overallocate memory and fail.
`buffer_size` is the size of that buffer in number of rows. The default
is 1024.
`read_consistency_level` and `write_consistency_level` set the default
consistency levels for every operation; these may be overridden
per-operation. These should be instances of
:class:`~pycassa.cassandra.ttypes.ConsistencyLevel`. These default
to level ``ONE``.
Each :meth:`insert()` or :meth:`remove` sends a timestamp with every
column. The `timestamp` parameter is a function that is used to get
this timestamp when needed. The default function is :meth:`gm_timestamp()`.
Results are returned as dictionaries. :class:`~pycassa.util.OrderedDict` is
used by default so that order is maintained. A different class, such as
:class:`dict` may be used instead by passing `dict_class`.
By default, column family definitions will be examined to determine
what data type Cassandra expects for column names and values. When
columns are retrieved or inserted, their names and values will be
packed or unpacked if necessary to convert them to or from their
binary representation. Automatic packing of names and values can
be individually enabled or disabled with `autopack_names` and
`autopack_values`. When using :class:`~pycassa.columnfamilymap.ColumnFamilyMap`,
these should both be set to ``False``.
"""
self.pool = pool
self._tlocal = threading.local()
self._tlocal.client = None
self.column_family = column_family
self.buffer_size = buffer_size
self.read_consistency_level = read_consistency_level
self.write_consistency_level = write_consistency_level
self.timestamp = timestamp
self.dict_class = dict_class
self.autopack_names = autopack_names
self.autopack_values = autopack_values
# Determine the ColumnFamily type to allow for auto conversion
# so that packing/unpacking doesn't need to be done manually
self.cf_data_type = None
self.col_name_data_type = None
self.supercol_name_data_type = None
self.col_type_dict = dict()
col_fam = None
try:
try:
self._obtain_connection()
col_fam = self._tlocal.client.get_keyspace_description(use_dict_for_col_metadata=True)[self.column_family]
except KeyError:
nfe = NotFoundException()
nfe.why = 'Column family %s not found.' % self.column_family
raise nfe
finally:
self._release_connection()
if col_fam is not None:
self.super = col_fam.column_type == 'Super'
if self.autopack_names:
if not self.super:
self.col_name_data_type = col_fam.comparator_type
else:
self.col_name_data_type = col_fam.subcomparator_type
self.supercol_name_data_type = util.extract_type_name(col_fam.comparator_type)
index = self.col_name_data_type = util.extract_type_name(self.col_name_data_type)
if self.autopack_values:
self.cf_data_type = util.extract_type_name(col_fam.default_validation_class)
for name, cdef in col_fam.column_metadata.items():
self.col_type_dict[name] = util.extract_type_name(cdef.validation_class)
def __init__(self,
pool,
column_family,
buffer_size=1024,
read_consistency_level=ConsistencyLevel.ONE,
write_consistency_level=ConsistencyLevel.ONE,
timestamp=gm_timestamp,
super=False,
dict_class=util.OrderedDict,
autopack_names=True,
autopack_values=True):
"""
An abstraction of a Cassandra column family or super column family.
Operations on this, such as :meth:`get` or :meth:`insert` will get data from or
insert data into the corresponding Cassandra column family with
name `column_family`.
`pool` is a :class:`~pycassa.pool.ConnectionPool` that the column
family will use for all operations. A connection is drawn from the
pool before each operations and is returned afterwards. Note that
the keyspace to be used is determined by the pool.
When calling :meth:`get_range()` or :meth:`get_indexed_slices()`,
the intermediate results need to be buffered if we are fetching many
rows, otherwise the Cassandra server will overallocate memory and fail.
`buffer_size` is the size of that buffer in number of rows. The default
is 1024.
`read_consistency_level` and `write_consistency_level` set the default
consistency levels for every operation; these may be overridden
per-operation. These should be instances of
:class:`~pycassa.cassandra.ttypes.ConsistencyLevel`. These default
to level ``ONE``.
Each :meth:`insert()` or :meth:`remove` sends a timestamp with every
column. The `timestamp` parameter is a function that is used to get
this timestamp when needed. The default function is :meth:`gm_timestamp()`.
Results are returned as dictionaries. :class:`~pycassa.util.OrderedDict` is
used by default so that order is maintained. A different class, such as
:class:`dict` may be used instead by passing `dict_class`.
By default, column family definitions will be examined to determine
what data type Cassandra expects for column names and values. When
columns are retrieved or inserted, their names and values will be
packed or unpacked if necessary to convert them to or from their
binary representation. Automatic packing of names and values can
be individually enabled or disabled with `autopack_names` and
`autopack_values`. When using :class:`~pycassa.columnfamilymap.ColumnFamilyMap`,
these should both be set to ``False``.
"""
self.pool = pool
self._tlocal = threading.local()
self._tlocal.client = None
self.column_family = column_family
self.buffer_size = buffer_size
self.read_consistency_level = read_consistency_level
self.write_consistency_level = write_consistency_level
self.timestamp = timestamp
self.dict_class = dict_class
self.autopack_names = autopack_names
self.autopack_values = autopack_values
# Determine the ColumnFamily type to allow for auto conversion
# so that packing/unpacking doesn't need to be done manually
self.cf_data_type = None
self.col_name_data_type = None
self.supercol_name_data_type = None
self.col_type_dict = dict()
col_fam = None
try:
try:
self._obtain_connection()
col_fam = self._tlocal.client.get_keyspace_description(
use_dict_for_col_metadata=True)[self.column_family]
except KeyError:
nfe = NotFoundException()
nfe.why = 'Column family %s not found.' % self.column_family
raise nfe
finally:
self._release_connection()
if col_fam is not None:
self.super = col_fam.column_type == 'Super'
if self.autopack_names:
if not self.super:
self.col_name_data_type = col_fam.comparator_type
else:
self.col_name_data_type = col_fam.subcomparator_type
self.supercol_name_data_type = util.extract_type_name(
col_fam.comparator_type)
index = self.col_name_data_type = util.extract_type_name(
self.col_name_data_type)
if self.autopack_values:
self.cf_data_type = util.extract_type_name(
col_fam.default_validation_class)
for name, cdef in col_fam.column_metadata.items():
self.col_type_dict[name] = util.extract_type_name(
cdef.validation_class)
def _load_key_class(self):
if hasattr(self._cfdef, "key_validation_class"):
self.key_validation_class = util.extract_type_name(self._cfdef.key_validation_class)
else:
self.key_validation_class = "BytesType"
def _load_validation_classes(self):
self.default_validation_class = util.extract_type_name(self._cfdef.default_validation_class)
self.column_validators = {}
for name, coldef in self._cfdef.column_metadata.items():
self.column_validators[name] = util.extract_type_name(coldef.validation_class)
def _load_key_class(self):
if hasattr(self._cfdef, "key_validation_class"):
self.key_validation_class = util.extract_type_name(self._cfdef.key_validation_class)
else:
self.key_validation_class = 'BytesType'
