/usr/lib/python2.7/dist-packages/pycassa/columnfamily.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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 | """
Provides an abstraction of Cassandra's data model to allow for easy
manipulation of data inside Cassandra.
.. seealso:: :mod:`pycassa.columnfamilymap`
"""
import time
import struct
from UserDict import DictMixin
from pycassa.cassandra.ttypes import Column, ColumnOrSuperColumn,\
ColumnParent, ColumnPath, ConsistencyLevel, NotFoundException,\
SlicePredicate, SliceRange, SuperColumn, KeyRange,\
IndexExpression, IndexClause, CounterColumn, Mutation
import pycassa.marshal as marshal
import pycassa.types as types
from pycassa.batch import CfMutator
try:
from collections import OrderedDict
except ImportError:
from pycassa.util import OrderedDict # NOQA
__all__ = ['gm_timestamp', 'ColumnFamily', 'PooledColumnFamily']
class ColumnValidatorDict(DictMixin):
def __init__(self, other_dict={}, name_packer=None, name_unpacker=None):
self.name_packer = name_packer or (lambda x: x)
self.name_unpacker = name_unpacker or (lambda x: x)
self.type_map = {}
self.packers = {}
self.unpackers = {}
for item, value in other_dict.items():
packed_item = self.name_packer(item)
self[packed_item] = value
def __getitem__(self, item):
packed_item = self.name_packer(item)
return self.type_map[packed_item]
def __setitem__(self, item, value):
packed_item = self.name_packer(item)
if isinstance(value, types.CassandraType):
self.type_map[packed_item] = value
self.packers[packed_item] = value.pack
self.unpackers[packed_item] = value.unpack
else:
self.type_map[packed_item] = marshal.extract_type_name(value)
self.packers[packed_item] = marshal.packer_for(value)
self.unpackers[packed_item] = marshal.unpacker_for(value)
def __delitem__(self, item):
packed_item = self.name_packer(item)
del self.type_map[packed_item]
del self.packers[packed_item]
del self.unpackers[packed_item]
def keys(self):
return map(self.name_unpacker, self.type_map.keys())
def gm_timestamp():
""" Returns the number of microseconds since the Unix Epoch. """
return int(time.time() * 1e6)
class ColumnFamily(object):
"""
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.
"""
buffer_size = 1024
""" When calling :meth:`get_range()` or :meth:`get_indexed_slices()`,
the intermediate results need to be buffered if we are fetching many
rows, otherwise performance may suffer and the Cassandra server may
overallocate memory and fail. This is the size of that buffer in number
of rows. The default is 1024. """
column_buffer_size = 1024
""" The number of columns fetched at once for :meth:`xget()` """
read_consistency_level = ConsistencyLevel.ONE
""" The default consistency level for every read operation, such as
:meth:`get` or :meth:`get_range`. This may be overridden per-operation. This should be
an instance of :class:`~pycassa.cassandra.ttypes.ConsistencyLevel`.
The default level is ``ONE``. """
write_consistency_level = ConsistencyLevel.ONE
""" The default consistency level for every write operation, such as
:meth:`insert` or :meth:`remove`. This may be overridden per-operation. This should be
an instance of :class:`.~pycassa.cassandra.ttypes.ConsistencyLevel`.
The default level is ``ONE``. """
timestamp = gm_timestamp
""" Each :meth:`insert()` or :meth:`remove` sends a timestamp with every
column. This attribute is a function that is used to get
this timestamp when needed. The default function is :meth:`gm_timestamp()`."""
dict_class = OrderedDict
""" Results are returned as dictionaries. By default, python 2.7's
:class:`collections.OrderedDict` is used if available, otherwise
:class:`~pycassa.util.OrderedDict` is used so that order is maintained.
A different class, such as :class:`dict`, may be instead by used setting
this. """
autopack_names = True
""" Controls whether column names are automatically converted to or from
their natural type to the binary string format that Cassandra uses.
The data type used is controlled by :attr:`column_name_class` for
column names and :attr:`super_column_name_class` for super column names.
By default, this is :const:`True`. """
autopack_values = True
""" Whether column values are automatically converted to or from
their natural type to the binary string format that Cassandra uses.
The data type used is controlled by :attr:`default_validation_class`
and :attr:`column_validators`.
By default, this is :const:`True`. """
autopack_keys = True
""" Whether row keys are automatically converted to or from
their natural type to the binary string format that Cassandra uses.
The data type used is controlled by :attr:`key_validation_class`.
By default, this is :const:`True`.
"""
retry_counter_mutations = False
""" Whether to retry failed counter mutations. Counter mutations are
not idempotent so retrying could result in double counting.
By default, this is :const:`False`.
.. versionadded:: 1.5.0
"""
def _set_column_name_class(self, t):
if isinstance(t, types.CassandraType):
self._column_name_class = t
self._name_packer = t.pack
self._name_unpacker = t.unpack
else:
self._column_name_class = marshal.extract_type_name(t)
self._name_packer = marshal.packer_for(t)
self._name_unpacker = marshal.unpacker_for(t)
def _get_column_name_class(self):
return self._column_name_class
column_name_class = property(_get_column_name_class, _set_column_name_class)
""" The data type of column names, which pycassa will use
to determine how to pack and unpack them.
This is set automatically by inspecting the column family's
``comparator_type``, but it may also be set manually if you want
autopacking behavior without setting a ``comparator_type``. Options
include an instance of any class in :mod:`pycassa.types`, such as ``LongType()``.
"""
def _set_super_column_name_class(self, t):
if isinstance(t, types.CassandraType):
self._super_column_name_class = t
self._super_name_packer = t.pack
self._super_name_unpacker = t.unpack
else:
self._super_column_name_class = marshal.extract_type_name(t)
self._super_name_packer = marshal.packer_for(t)
self._super_name_unpacker = marshal.unpacker_for(t)
def _get_super_column_name_class(self):
return self._super_column_name_class
super_column_name_class = property(_get_super_column_name_class,
_set_super_column_name_class)
""" Like :attr:`column_name_class`, but for
super column names. """
def _set_default_validation_class(self, t):
if isinstance(t, types.CassandraType):
self._default_validation_class = t
self._default_value_packer = t.pack
self._default_value_unpacker = t.unpack
self._have_counters = isinstance(t, types.CounterColumnType)
else:
self._default_validation_class = marshal.extract_type_name(t)
self._default_value_packer = marshal.packer_for(t)
self._default_value_unpacker = marshal.unpacker_for(t)
self._have_counters = self._default_validation_class == "CounterColumnType"
if not self.super:
if self._have_counters:
def _make_counter_cosc(name, value, timestamp, ttl):
return ColumnOrSuperColumn(counter_column=CounterColumn(name, value))
self._make_cosc = _make_counter_cosc
else:
def _make_normal_cosc(name, value, timestamp, ttl):
return ColumnOrSuperColumn(Column(name, value, timestamp, ttl))
self._make_cosc = _make_normal_cosc
else:
if self._have_counters:
def _make_column(name, value, timestamp, ttl):
return CounterColumn(name, value)
self._make_column = _make_column
def _make_counter_super_cosc(scol_name, subcols):
return ColumnOrSuperColumn(counter_super_column=(SuperColumn(scol_name, subcols)))
self._make_cosc = _make_counter_super_cosc
else:
self._make_column = Column
def _make_super_cosc(scol_name, subcols):
return ColumnOrSuperColumn(super_column=(SuperColumn(scol_name, subcols)))
self._make_cosc = _make_super_cosc
def _get_default_validation_class(self):
return self._default_validation_class
default_validation_class = property(_get_default_validation_class,
_set_default_validation_class)
""" The default data type of column values, which pycassa
will use to determine how to pack and unpack them.
This is set automatically by inspecting the column family's
``default_validation_class``, but it may also be set manually if you want
autopacking behavior without setting a ``default_validation_class``. Options
include an instance of any class in :mod:`pycassa.types`, such as ``LongType()``.
"""
@property
def _allow_retries(self):
return not self._have_counters or self.retry_counter_mutations
def _set_column_validators(self, other_dict):
self._column_validators = ColumnValidatorDict(other_dict, self._pack_name, self._unpack_name)
def _get_column_validators(self):
return self._column_validators
column_validators = property(_get_column_validators, _set_column_validators)
""" Like :attr:`default_validation_class`, but is a
:class:`dict` mapping individual columns to types. """
def _set_key_validation_class(self, t):
if isinstance(t, types.CassandraType):
self._key_validation_class = t
self._key_packer = t.pack
self._key_unpacker = t.unpack
else:
self._key_validation_class = marshal.extract_type_name(t)
self._key_packer = marshal.packer_for(t)
self._key_unpacker = marshal.unpacker_for(t)
def _get_key_validation_class(self):
return self._key_validation_class
key_validation_class = property(_get_key_validation_class,
_set_key_validation_class)
""" The data type of row keys, which pycassa will use
to determine how to pack and unpack them.
This is set automatically by inspecting the column family's
``key_validation_class`` (which only exists in Cassandra 0.8 or greater),
but may be set manually if you want the autopacking behavior without
setting a ``key_validation_class`` or if you are using Cassandra 0.7.
Options include an instance of any class in :mod:`pycassa.types`,
such as ``LongType()``.
"""
def __init__(self, pool, column_family, **kwargs):
"""
`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.
`column_family` should be the name of the column family that you
want to use in Cassandra. Note that the keyspace to be used is
determined by the pool.
"""
self.pool = pool
self.column_family = column_family
self.timestamp = gm_timestamp
self.load_schema()
recognized_kwargs = ("buffer_size", "read_consistency_level",
"write_consistency_level", "timestamp",
"dict_class", "buffer_size", "autopack_names",
"autopack_values", "autopack_keys",
"retry_counter_mutations")
for k, v in kwargs.iteritems():
if k in recognized_kwargs:
setattr(self, k, v)
else:
raise TypeError(
"ColumnFamily.__init__() got an unexpected keyword "
"argument '%s'" % (k,))
def load_schema(self):
"""
Loads the schema definition for this column family from
Cassandra and updates comparator and validation classes if
neccessary.
"""
ksdef = self.pool.execute('get_keyspace_description',
use_dict_for_col_metadata=True)
try:
self._cfdef = ksdef[self.column_family]
except KeyError:
nfe = NotFoundException()
nfe.why = 'Column family %s not found.' % self.column_family
raise nfe
self.super = self._cfdef.column_type == 'Super'
self._load_comparator_classes()
self._load_validation_classes()
self._load_key_class()
def _load_comparator_classes(self):
if not self.super:
self.column_name_class = self._cfdef.comparator_type
self.super_column_name_class = None
else:
self.column_name_class = self._cfdef.subcomparator_type
self.super_column_name_class = self._cfdef.comparator_type
def _load_validation_classes(self):
self.default_validation_class = self._cfdef.default_validation_class
self.column_validators = {}
for name, coldef in self._cfdef.column_metadata.items():
unpacked_name = self._unpack_name(name)
self.column_validators[unpacked_name] = coldef.validation_class
def _load_key_class(self):
if hasattr(self._cfdef, "key_validation_class"):
self.key_validation_class = self._cfdef.key_validation_class
else:
self.key_validation_class = 'BytesType'
def _col_to_dict(self, column, include_timestamp, include_ttl):
value = self._unpack_value(column.value, column.name)
if include_timestamp and include_ttl:
return (value, column.timestamp, column.ttl)
elif include_timestamp:
return (value, column.timestamp)
elif include_ttl:
return (value, column.ttl)
else:
return value
def _scol_to_dict(self, super_column, include_timestamp, include_ttl):
ret = self.dict_class()
for column in super_column.columns:
ret[self._unpack_name(column.name)] = self._col_to_dict(column, include_timestamp, include_ttl)
return ret
def _scounter_to_dict(self, counter_super_column):
ret = self.dict_class()
for counter in counter_super_column.columns:
ret[self._unpack_name(counter.name)] = counter.value
return ret
def _cosc_to_dict(self, list_col_or_super, include_timestamp, include_ttl):
ret = self.dict_class()
for cosc in list_col_or_super:
if cosc.column:
col = cosc.column
ret[self._unpack_name(col.name)] = self._col_to_dict(col, include_timestamp, include_ttl)
elif cosc.counter_column:
counter = cosc.counter_column
ret[self._unpack_name(counter.name)] = counter.value
elif cosc.super_column:
scol = cosc.super_column
ret[self._unpack_name(scol.name, True)] = self._scol_to_dict(scol, include_timestamp, include_ttl)
else:
scounter = cosc.counter_super_column
ret[self._unpack_name(scounter.name, True)] = self._scounter_to_dict(scounter)
return ret
def _column_path(self, super_column=None, column=None):
return ColumnPath(self.column_family,
self._pack_name(super_column, is_supercol_name=True),
self._pack_name(column, False))
def _column_parent(self, super_column=None):
return ColumnParent(column_family=self.column_family,
super_column=self._pack_name(super_column, is_supercol_name=True))
def _slice_predicate(self, columns, column_start, column_finish,
column_reversed, column_count, super_column=None, pack=True):
is_supercol_name = self.super and super_column is None
if columns is not None:
packed_cols = []
for col in columns:
packed_cols.append(self._pack_name(col, is_supercol_name=is_supercol_name))
return SlicePredicate(column_names=packed_cols)
else:
if column_start != '' and pack:
column_start = self._pack_name(column_start,
is_supercol_name=is_supercol_name,
slice_start=(not column_reversed))
if column_finish != '' and pack:
column_finish = self._pack_name(column_finish,
is_supercol_name=is_supercol_name,
slice_start=column_reversed)
sr = SliceRange(start=column_start, finish=column_finish,
reversed=column_reversed, count=column_count)
return SlicePredicate(slice_range=sr)
def _pack_name(self, value, is_supercol_name=False, slice_start=None):
if value is None:
return
if not self.autopack_names:
if not isinstance(value, basestring):
raise TypeError("A str or unicode column name was expected, " +
"but %s was received instead (%s)"
% (value.__class__.__name__, str(value)))
return value
try:
if is_supercol_name:
return self._super_name_packer(value, slice_start)
else:
return self._name_packer(value, slice_start)
except struct.error:
if is_supercol_name:
d_type = self.super_column_name_class
else:
d_type = self.column_name_class
raise TypeError("%s is not a compatible type for %s" %
(value.__class__.__name__, d_type))
def _unpack_name(self, b, is_supercol_name=False):
if not self.autopack_names:
return b
try:
if is_supercol_name:
return self._super_name_unpacker(b)
else:
return self._name_unpacker(b)
except struct.error:
if is_supercol_name:
d_type = self.super_column_name_class
else:
d_type = self.column_name_class
raise TypeError("%s cannot be converted to a type matching %s" %
(b, d_type))
def _pack_value(self, value, col_name):
if value is None:
return
if not self.autopack_values:
if not isinstance(value, basestring):
raise TypeError("A str or unicode column value was expected for " +
"column '%s', but %s was received instead (%s)"
% (str(col_name), value.__class__.__name__, str(value)))
return value
packed_col_name = self._pack_name(col_name, False)
packer = self._column_validators.packers.get(packed_col_name, self._default_value_packer)
try:
return packer(value)
except struct.error:
d_type = self.column_validators.get(col_name, self._default_validation_class)
raise TypeError("%s is not a compatible type for %s" %
(value.__class__.__name__, d_type))
def _unpack_value(self, value, col_name):
if not self.autopack_values:
return value
unpacker = self._column_validators.unpackers.get(col_name, self._default_value_unpacker)
try:
return unpacker(value)
except struct.error:
d_type = self.column_validators.get(col_name, self.default_validation_class)
raise TypeError("%s cannot be converted to a type matching %s" %
(value, d_type))
def _pack_key(self, key):
if not self.autopack_keys or key == '':
return key
try:
return self._key_packer(key)
except struct.error:
d_type = self.key_validation_class
raise TypeError("%s is not a compatible type for %s" %
(key.__class__.__name__, d_type))
def _unpack_key(self, b):
if not self.autopack_keys:
return b
try:
return self._key_unpacker(b)
except struct.error:
d_type = self.key_validation_class
raise TypeError("%s cannot be converted to a type matching %s" %
(b, d_type))
def _make_mutation_list(self, columns, timestamp, ttl):
_pack_name = self._pack_name
_pack_value = self._pack_value
if not self.super:
return map(lambda (c, v): Mutation(self._make_cosc(_pack_name(c), _pack_value(v, c), timestamp, ttl)),
columns.iteritems())
else:
mut_list = []
for super_col, subcs in columns.items():
subcols = map(lambda (c, v): self._make_column(_pack_name(c), _pack_value(v, c), timestamp, ttl),
subcs.iteritems())
mut_list.append(Mutation(self._make_cosc(_pack_name(super_col, True), subcols)))
return mut_list
def xget(self, key, column_start="", column_finish="", column_reversed=False,
column_count=None, include_timestamp=False, read_consistency_level=None,
buffer_size=None, include_ttl=False):
"""
Like :meth:`get()`, but creates a generator that pages over the columns
automatically.
The number of columns fetched at once can be controlled with the
`buffer_size` parameter. The default is :attr:`column_buffer_size`.
The generator returns `(name, value)` tuples.
"""
packed_key = self._pack_key(key)
cp = self._column_parent(None)
rcl = read_consistency_level or self.read_consistency_level
if buffer_size is None:
buffer_size = self.column_buffer_size
count = i = 0
last_name = finish = ""
if column_start != "":
last_name = self._pack_name(column_start,
is_supercol_name=self.super,
slice_start=(not column_reversed))
if column_finish != "":
finish = self._pack_name(column_finish,
is_supercol_name=self.super,
slice_start=column_reversed)
while True:
if column_count is not None:
if i == 0 and column_count <= buffer_size:
buffer_size = column_count
else:
buffer_size = min(column_count - count + 1, buffer_size)
sp = self._slice_predicate(None, last_name, finish,
column_reversed, buffer_size, None, pack=False)
list_cosc = self.pool.execute('get_slice', packed_key, cp, sp, rcl)
if not list_cosc:
return
for j, cosc in enumerate(list_cosc):
if j == 0 and i != 0:
continue
if self.super:
if self._have_counters:
scol = cosc.counter_super_column
else:
scol = cosc.super_column
yield (self._unpack_name(scol.name, True), self._scol_to_dict(scol, include_timestamp, include_ttl))
else:
if self._have_counters:
col = cosc.counter_column
else:
col = cosc.column
yield (self._unpack_name(col.name, False), self._col_to_dict(col, include_timestamp, include_ttl))
count += 1
if column_count is not None and count >= column_count:
return
if len(list_cosc) != buffer_size:
return
if self.super:
if self._have_counters:
last_name = list_cosc[-1].counter_super_column.name
else:
last_name = list_cosc[-1].super_column.name
else:
if self._have_counters:
last_name = list_cosc[-1].counter_column.name
else:
last_name = list_cosc[-1].column.name
i += 1
def get(self, key, columns=None, column_start="", column_finish="",
column_reversed=False, column_count=100, include_timestamp=False,
super_column=None, read_consistency_level=None, include_ttl=False):
"""
Fetches all or part of the row with key `key`.
The columns fetched may be limited to a specified list of column names
using `columns`.
Alternatively, you may fetch a slice of columns or super columns from a row
using `column_start`, `column_finish`, and `column_count`.
Setting these will cause columns or super columns to be fetched starting with
`column_start`, continuing until `column_count` columns or super columns have
been fetched or `column_finish` is reached. If `column_start` is left as the
empty string, the slice will begin with the start of the row; leaving
`column_finish` blank will cause the slice to extend to the end of the row.
Note that `column_count` defaults to 100, so rows over this size will not be
completely fetched by default.
If `column_reversed` is ``True``, columns are fetched in reverse sorted order,
beginning with `column_start`. In this case, if `column_start` is the empty
string, the slice will begin with the end of the row.
You may fetch all or part of only a single super column by setting `super_column`.
If this is set, `column_start`, `column_finish`, `column_count`, and `column_reversed`
will apply to the subcolumns of `super_column`.
To include every column's timestamp in the result set, set `include_timestamp` to
``True``. Results will include a ``(value, timestamp)`` tuple for each column.
To include every column's ttl in the result set, set `include_ttl` to
``True``. Results will include a ``(value, ttl)`` tuple for each column.
If this is a standard column family, the return type is of the form
``{column_name: column_value}``. If this is a super column family and `super_column`
is not specified, the results are of the form
``{super_column_name: {column_name, column_value}}``. If `super_column` is set,
the super column name will be excluded and the results are of the form
``{column_name: column_value}``.
"""
packed_key = self._pack_key(key)
single_column = columns is not None and len(columns) == 1
if (not self.super and single_column) or \
(self.super and super_column is not None and single_column):
column = None
if self.super and super_column is None:
super_column = columns[0]
else:
column = columns[0]
cp = self._column_path(super_column, column)
col_or_super = self.pool.execute('get', packed_key, cp,
read_consistency_level or self.read_consistency_level)
return self._cosc_to_dict([col_or_super], include_timestamp, include_ttl)
else:
cp = self._column_parent(super_column)
sp = self._slice_predicate(columns, column_start, column_finish,
column_reversed, column_count, super_column)
list_col_or_super = self.pool.execute('get_slice', packed_key, cp, sp,
read_consistency_level or self.read_consistency_level)
if len(list_col_or_super) == 0:
raise NotFoundException()
return self._cosc_to_dict(list_col_or_super, include_timestamp, include_ttl)
def get_indexed_slices(self, index_clause, columns=None, column_start="", column_finish="",
column_reversed=False, column_count=100, include_timestamp=False,
read_consistency_level=None, buffer_size=None, include_ttl=False):
"""
Similar to :meth:`get_range()`, but an :class:`~pycassa.cassandra.ttypes.IndexClause`
is used instead of a key range.
`index_clause` limits the keys that are returned based on expressions
that compare the value of a column to a given value. At least one of the
expressions in the :class:`.IndexClause` must be on an indexed column.
Note that Cassandra does not support secondary indexes or get_indexed_slices()
for super column families.
.. seealso:: :meth:`~pycassa.index.create_index_clause()` and
:meth:`~pycassa.index.create_index_expression()`
"""
assert not self.super, "get_indexed_slices() is not " \
"supported by super column families"
cl = read_consistency_level or self.read_consistency_level
cp = self._column_parent()
sp = self._slice_predicate(columns, column_start, column_finish,
column_reversed, column_count)
new_exprs = []
# Pack the values in the index clause expressions
for expr in index_clause.expressions:
value = self._pack_value(expr.value, expr.column_name)
name = self._pack_name(expr.column_name)
new_exprs.append(IndexExpression(name, expr.op, value))
packed_start_key = self._pack_key(index_clause.start_key)
clause = IndexClause(new_exprs, packed_start_key, index_clause.count)
# Figure out how we will chunk the request
if buffer_size is None:
buffer_size = self.buffer_size
row_count = clause.count
count = 0
i = 0
last_key = clause.start_key
while True:
if row_count is not None:
if i == 0 and row_count <= buffer_size:
# We don't need to chunk, grab exactly the number of rows
buffer_size = row_count
else:
buffer_size = min(row_count - count + 1, buffer_size)
clause.count = buffer_size
clause.start_key = last_key
key_slices = self.pool.execute('get_indexed_slices', cp, clause, sp, cl)
if key_slices is None:
return
for j, key_slice in enumerate(key_slices):
# Ignore the first element after the first iteration
# because it will be a duplicate.
if j == 0 and i != 0:
continue
unpacked_key = self._unpack_key(key_slice.key)
yield (unpacked_key,
self._cosc_to_dict(key_slice.columns, include_timestamp, include_ttl))
count += 1
if row_count is not None and count >= row_count:
return
if len(key_slices) != buffer_size:
return
last_key = key_slices[-1].key
i += 1
def multiget(self, keys, columns=None, column_start="", column_finish="",
column_reversed=False, column_count=100, include_timestamp=False,
super_column=None, read_consistency_level=None, buffer_size=None, include_ttl=False):
"""
Fetch multiple rows from a Cassandra server.
`keys` should be a list of keys to fetch.
`buffer_size` is the number of rows from the total list to fetch at a time.
If left as ``None``, the ColumnFamily's :attr:`buffer_size` will be used.
All other parameters are the same as :meth:`get()`, except that a list of keys may
be passed in.
Results will be returned in the form: ``{key: {column_name: column_value}}``. If
an OrderedDict is used, the rows will have the same order as `keys`.
"""
packed_keys = map(self._pack_key, keys)
cp = self._column_parent(super_column)
sp = self._slice_predicate(columns, column_start, column_finish,
column_reversed, column_count, super_column)
consistency = read_consistency_level or self.read_consistency_level
buffer_size = buffer_size or self.buffer_size
offset = 0
keymap = {}
while offset < len(packed_keys):
new_keymap = self.pool.execute('multiget_slice',
packed_keys[offset:offset + buffer_size], cp, sp, consistency)
keymap.update(new_keymap)
offset += buffer_size
ret = self.dict_class()
# Keep the order of keys
for key in keys:
ret[key] = None
empty_keys = []
for packed_key, columns in keymap.iteritems():
unpacked_key = self._unpack_key(packed_key)
if len(columns) > 0:
ret[unpacked_key] = self._cosc_to_dict(columns, include_timestamp, include_ttl)
else:
empty_keys.append(unpacked_key)
for key in empty_keys:
try:
del ret[key]
except KeyError:
pass
return ret
MAX_COUNT = 2 ** 31 - 1
def get_count(self, key, super_column=None, read_consistency_level=None,
columns=None, column_start="", column_finish="",
column_reversed=False, max_count=None):
"""
Count the number of columns in the row with key `key`.
You may limit the columns or super columns counted to those in `columns`.
Additionally, you may limit the columns or super columns counted to
only those between `column_start` and `column_finish`.
You may also count only the number of subcolumns in a single super column
using `super_column`. If this is set, `columns`, `column_start`, and
`column_finish` only apply to the subcolumns of `super_column`.
To put an upper bound on the number of columns that are counted,
set `max_count`.
"""
if max_count is None:
max_count = self.MAX_COUNT
packed_key = self._pack_key(key)
cp = self._column_parent(super_column)
sp = self._slice_predicate(columns, column_start, column_finish,
column_reversed, max_count, super_column)
return self.pool.execute('get_count', packed_key, cp, sp,
read_consistency_level or self.read_consistency_level)
def multiget_count(self, keys, super_column=None,
read_consistency_level=None,
columns=None, column_start="",
column_finish="", buffer_size=None,
column_reversed=False, max_count=None):
"""
Perform a column count in parallel on a set of rows.
The parameters are the same as for :meth:`multiget()`, except that a list
of keys may be used. A dictionary of the form ``{key: int}`` is
returned.
`buffer_size` is the number of rows from the total list to count at a time.
If left as ``None``, the ColumnFamily's :attr:`buffer_size` will be used.
To put an upper bound on the number of columns that are counted,
set `max_count`.
"""
if max_count is None:
max_count = self.MAX_COUNT
packed_keys = map(self._pack_key, keys)
cp = self._column_parent(super_column)
sp = self._slice_predicate(columns, column_start, column_finish,
column_reversed, max_count, super_column)
consistency = read_consistency_level or self.read_consistency_level
buffer_size = buffer_size or self.buffer_size
offset = 0
keymap = {}
while offset < len(packed_keys):
new_keymap = self.pool.execute('multiget_count',
packed_keys[offset:offset + buffer_size], cp, sp, consistency)
keymap.update(new_keymap)
offset += buffer_size
ret = self.dict_class()
# Keep the order of keys
for key in keys:
ret[key] = None
for packed_key, count in keymap.iteritems():
ret[self._unpack_key(packed_key)] = count
return ret
def get_range(self, start="", finish="", columns=None, column_start="",
column_finish="", column_reversed=False, column_count=100,
row_count=None, include_timestamp=False,
super_column=None, read_consistency_level=None,
buffer_size=None, filter_empty=True, include_ttl=False,
start_token=None, finish_token=None):
"""
Get an iterator over rows in a specified key range.
The key range begins with `start` and ends with `finish`. If left
as empty strings, these extend to the beginning and end, respectively.
Note that if RandomPartitioner is used, rows are stored in the
order of the MD5 hash of their keys, so getting a lexicographical range
of keys is not feasible.
In place of `start` and `finish`, you may use `start_token` and
`finish_token` or a combination of `start` and `finish_token`. In this
case, you are specifying a token range to fetch instead of a key
range. This can be useful for fetching all data owned
by a node or for parallelizing a full data set scan. Otherwise,
you should typically just use `start` and `finish`. When using
RandomPartitioner or Murmur3Partitioner, `start_token`
and `finish_token` should be string versions of the numeric tokens;
for ByteOrderedPartitioner, they should be hex-encoded string versions
of the token.
The `row_count` parameter limits the total number of rows that may be
returned. If left as ``None``, the number of rows that may be returned
is unlimited (this is the default).
When calling `get_range()`, 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. If left as ``None``, the
ColumnFamily's :attr:`buffer_size` attribute will be used.
When `filter_empty` is left as ``True``, empty rows (including
`range ghosts <http://wiki.apache.org/cassandra/FAQ#range_ghosts>`_)
will be skipped and will not count towards `row_count`.
All other parameters are the same as those of :meth:`get()`.
A generator over ``(key, {column_name: column_value})`` is returned.
To convert this to a list, use ``list()`` on the result.
"""
cl = read_consistency_level or self.read_consistency_level
cp = self._column_parent(super_column)
sp = self._slice_predicate(columns, column_start, column_finish,
column_reversed, column_count, super_column)
kr_args = {}
count = 0
i = 0
if start_token is not None and (start not in ("", None) or finish not in ("", None)):
raise ValueError(
"ColumnFamily.get_range() received incompatible arguments: "
"'start_token' may not be used with 'start' or 'finish'")
if finish_token is not None and finish not in ("", None):
raise ValueError(
"ColumnFamily.get_range() received incompatible arguments: "
"'finish_token' may not be used with 'finish'")
if start_token is not None:
kr_args['start_token'] = start_token
kr_args['end_token'] = "" if finish_token is None else finish_token
elif finish_token is not None:
kr_args['start_key'] = self._pack_key(start)
kr_args['end_token'] = finish_token
else:
kr_args['start_key'] = self._pack_key(start)
kr_args['end_key'] = self._pack_key(finish)
if buffer_size is None:
buffer_size = self.buffer_size
while True:
if row_count is not None:
if i == 0 and row_count <= buffer_size:
# We don't need to chunk, grab exactly the number of rows
buffer_size = row_count
else:
buffer_size = min(row_count - count + 1, buffer_size)
kr_args['count'] = buffer_size
key_range = KeyRange(**kr_args)
key_slices = self.pool.execute('get_range_slices', cp, sp, key_range, cl)
# This may happen if nothing was ever inserted
if key_slices is None:
return
for j, key_slice in enumerate(key_slices):
# Ignore the first element after the first iteration
# because it will be a duplicate.
if j == 0 and i != 0:
continue
if filter_empty and not key_slice.columns:
continue
yield (self._unpack_key(key_slice.key),
self._cosc_to_dict(key_slice.columns, include_timestamp, include_ttl))
count += 1
if row_count is not None and count >= row_count:
return
if len(key_slices) != buffer_size:
return
if 'start_token' in kr_args:
del kr_args['start_token']
kr_args['start_key'] = key_slices[-1].key
i += 1
def insert(self, key, columns, timestamp=None, ttl=None,
write_consistency_level=None):
"""
Insert or update columns in the row with key `key`.
`columns` should be a dictionary of columns or super columns to insert
or update. If this is a standard column family, `columns` should
look like ``{column_name: column_value}``. If this is a super
column family, `columns` should look like
``{super_column_name: {sub_column_name: value}}``. If this is a
counter column family, you may use integers as values and those will
be used as counter adjustments.
A timestamp may be supplied for all inserted columns with `timestamp`.
`ttl` sets the "time to live" in number of seconds for the inserted
columns. After this many seconds, Cassandra will mark the columns as
deleted.
The timestamp Cassandra reports as being used for insert is returned.
"""
if timestamp is None:
timestamp = self.timestamp()
packed_key = self._pack_key(key)
mut_list = self._make_mutation_list(columns, timestamp, ttl)
mutations = {packed_key: {self.column_family: mut_list}}
self.pool.execute('batch_mutate', mutations,
write_consistency_level or self.write_consistency_level,
allow_retries=self._allow_retries)
return timestamp
def batch_insert(self, rows, timestamp=None, ttl=None, write_consistency_level=None):
"""
Like :meth:`insert()`, but multiple rows may be inserted at once.
The `rows` parameter should be of the form ``{key: {column_name: column_value}}``
if this is a standard column family or
``{key: {super_column_name: {column_name: column_value}}}`` if this is a super
column family.
"""
if timestamp == None:
timestamp = self.timestamp()
cf = self.column_family
mutations = {}
for key, columns in rows.iteritems():
packed_key = self._pack_key(key)
mut_list = self._make_mutation_list(columns, timestamp, ttl)
mutations[packed_key] = {cf: mut_list}
if mutations:
self.pool.execute('batch_mutate', mutations,
write_consistency_level or self.write_consistency_level,
allow_retries=self._allow_retries)
return timestamp
def add(self, key, column, value=1, super_column=None, write_consistency_level=None):
"""
Increment or decrement a counter.
`value` should be an integer, either positive or negative, to be added
to a counter column. By default, `value` is 1.
.. versionadded:: 1.1.0
Available in Cassandra 0.8.0 and later.
"""
packed_key = self._pack_key(key)
cp = self._column_parent(super_column)
column = self._pack_name(column)
self.pool.execute('add', packed_key, cp, CounterColumn(column, value),
write_consistency_level or self.write_consistency_level,
allow_retries=self._allow_retries)
def remove(self, key, columns=None, super_column=None,
write_consistency_level=None, timestamp=None, counter=None):
"""
Remove a specified row or a set of columns within the row with key `key`.
A set of columns or super columns to delete may be specified using
`columns`.
A single super column may be deleted by setting `super_column`. If
`super_column` is specified, `columns` will apply to the subcolumns
of `super_column`.
If `columns` and `super_column` are both ``None``, the entire row is
removed.
The timestamp used for the mutation is returned.
"""
if timestamp is None:
timestamp = self.timestamp()
batch = self.batch(write_consistency_level=write_consistency_level)
batch.remove(key, columns, super_column, timestamp)
batch.send()
return timestamp
def remove_counter(self, key, column, super_column=None, write_consistency_level=None):
"""
Remove a counter at the specified location.
Note that counters have limited support for deletes: if you remove a
counter, you must wait to issue any following update until the delete
has reached all the nodes and all of them have been fully compacted.
.. versionadded:: 1.1.0
Available in Cassandra 0.8.0 and later.
"""
packed_key = self._pack_key(key)
cp = self._column_path(super_column, column)
self.pool.execute('remove_counter', packed_key, cp,
write_consistency_level or self.write_consistency_level)
def batch(self, queue_size=100, write_consistency_level=None, atomic=None):
"""
Create batch mutator for doing multiple insert, update, and remove
operations using as few roundtrips as possible.
The `queue_size` parameter sets the max number of mutations per request.
A :class:`~pycassa.batch.CfMutator` is returned.
"""
return CfMutator(self, queue_size,
write_consistency_level or self.write_consistency_level,
allow_retries=self._allow_retries,
atomic=atomic)
def truncate(self):
"""
Marks the entire ColumnFamily as deleted.
From the user's perspective, a successful call to ``truncate`` will
result complete data deletion from this column family. Internally,
however, disk space will not be immediately released, as with all
deletes in Cassandra, this one only marks the data as deleted.
The operation succeeds only if all hosts in the cluster at available
and will throw an :exc:`.UnavailableException` if some hosts are
down.
"""
self.pool.execute('truncate', self.column_family)
PooledColumnFamily = ColumnFamily
|