[2] | 1 | #-----------------------------------------------------------------------
|
---|
| 2 | #
|
---|
| 3 | # Copyright (C) 2000, 2001 by Autonomous Zone Industries
|
---|
| 4 | # Copyright (C) 2002 Gregory P. Smith
|
---|
| 5 | #
|
---|
| 6 | # License: This is free software. You may use this software for any
|
---|
| 7 | # purpose including modification/redistribution, so long as
|
---|
| 8 | # this header remains intact and that you do not claim any
|
---|
| 9 | # rights of ownership or authorship of this software. This
|
---|
| 10 | # software has been tested, but no warranty is expressed or
|
---|
| 11 | # implied.
|
---|
| 12 | #
|
---|
| 13 | # -- Gregory P. Smith <greg@krypto.org>
|
---|
| 14 |
|
---|
| 15 | # This provides a simple database table interface built on top of
|
---|
| 16 | # the Python Berkeley DB 3 interface.
|
---|
| 17 | #
|
---|
[391] | 18 | _cvsid = '$Id$'
|
---|
[2] | 19 |
|
---|
| 20 | import re
|
---|
| 21 | import sys
|
---|
| 22 | import copy
|
---|
| 23 | import random
|
---|
| 24 | import struct
|
---|
| 25 |
|
---|
[391] | 26 |
|
---|
| 27 | if sys.version_info[0] >= 3 :
|
---|
| 28 | import pickle
|
---|
| 29 | else :
|
---|
| 30 | if sys.version_info < (2, 6) :
|
---|
| 31 | import cPickle as pickle
|
---|
| 32 | else :
|
---|
| 33 | # When we drop support for python 2.4
|
---|
| 34 | # we could use: (in 2.5 we need a __future__ statement)
|
---|
| 35 | #
|
---|
| 36 | # with warnings.catch_warnings():
|
---|
| 37 | # warnings.filterwarnings(...)
|
---|
| 38 | # ...
|
---|
| 39 | #
|
---|
| 40 | # We can not use "with" as is, because it would be invalid syntax
|
---|
| 41 | # in python 2.4 and (with no __future__) 2.5.
|
---|
| 42 | # Here we simulate "with" following PEP 343 :
|
---|
| 43 | import warnings
|
---|
| 44 | w = warnings.catch_warnings()
|
---|
| 45 | w.__enter__()
|
---|
| 46 | try :
|
---|
| 47 | warnings.filterwarnings('ignore',
|
---|
| 48 | message='the cPickle module has been removed in Python 3.0',
|
---|
| 49 | category=DeprecationWarning)
|
---|
| 50 | import cPickle as pickle
|
---|
| 51 | finally :
|
---|
| 52 | w.__exit__()
|
---|
| 53 | del w
|
---|
| 54 |
|
---|
[2] | 55 | try:
|
---|
| 56 | # For Pythons w/distutils pybsddb
|
---|
| 57 | from bsddb3 import db
|
---|
| 58 | except ImportError:
|
---|
| 59 | # For Python 2.3
|
---|
| 60 | from bsddb import db
|
---|
| 61 |
|
---|
| 62 | class TableDBError(StandardError):
|
---|
| 63 | pass
|
---|
| 64 | class TableAlreadyExists(TableDBError):
|
---|
| 65 | pass
|
---|
| 66 |
|
---|
| 67 |
|
---|
| 68 | class Cond:
|
---|
| 69 | """This condition matches everything"""
|
---|
| 70 | def __call__(self, s):
|
---|
| 71 | return 1
|
---|
| 72 |
|
---|
| 73 | class ExactCond(Cond):
|
---|
| 74 | """Acts as an exact match condition function"""
|
---|
| 75 | def __init__(self, strtomatch):
|
---|
| 76 | self.strtomatch = strtomatch
|
---|
| 77 | def __call__(self, s):
|
---|
| 78 | return s == self.strtomatch
|
---|
| 79 |
|
---|
| 80 | class PrefixCond(Cond):
|
---|
| 81 | """Acts as a condition function for matching a string prefix"""
|
---|
| 82 | def __init__(self, prefix):
|
---|
| 83 | self.prefix = prefix
|
---|
| 84 | def __call__(self, s):
|
---|
| 85 | return s[:len(self.prefix)] == self.prefix
|
---|
| 86 |
|
---|
| 87 | class PostfixCond(Cond):
|
---|
| 88 | """Acts as a condition function for matching a string postfix"""
|
---|
| 89 | def __init__(self, postfix):
|
---|
| 90 | self.postfix = postfix
|
---|
| 91 | def __call__(self, s):
|
---|
| 92 | return s[-len(self.postfix):] == self.postfix
|
---|
| 93 |
|
---|
| 94 | class LikeCond(Cond):
|
---|
| 95 | """
|
---|
| 96 | Acts as a function that will match using an SQL 'LIKE' style
|
---|
| 97 | string. Case insensitive and % signs are wild cards.
|
---|
| 98 | This isn't perfect but it should work for the simple common cases.
|
---|
| 99 | """
|
---|
| 100 | def __init__(self, likestr, re_flags=re.IGNORECASE):
|
---|
| 101 | # escape python re characters
|
---|
| 102 | chars_to_escape = '.*+()[]?'
|
---|
| 103 | for char in chars_to_escape :
|
---|
| 104 | likestr = likestr.replace(char, '\\'+char)
|
---|
| 105 | # convert %s to wildcards
|
---|
| 106 | self.likestr = likestr.replace('%', '.*')
|
---|
| 107 | self.re = re.compile('^'+self.likestr+'$', re_flags)
|
---|
| 108 | def __call__(self, s):
|
---|
| 109 | return self.re.match(s)
|
---|
| 110 |
|
---|
| 111 | #
|
---|
| 112 | # keys used to store database metadata
|
---|
| 113 | #
|
---|
| 114 | _table_names_key = '__TABLE_NAMES__' # list of the tables in this db
|
---|
| 115 | _columns = '._COLUMNS__' # table_name+this key contains a list of columns
|
---|
| 116 |
|
---|
| 117 | def _columns_key(table):
|
---|
| 118 | return table + _columns
|
---|
| 119 |
|
---|
| 120 | #
|
---|
| 121 | # these keys are found within table sub databases
|
---|
| 122 | #
|
---|
| 123 | _data = '._DATA_.' # this+column+this+rowid key contains table data
|
---|
| 124 | _rowid = '._ROWID_.' # this+rowid+this key contains a unique entry for each
|
---|
| 125 | # row in the table. (no data is stored)
|
---|
| 126 | _rowid_str_len = 8 # length in bytes of the unique rowid strings
|
---|
| 127 |
|
---|
| 128 |
|
---|
| 129 | def _data_key(table, col, rowid):
|
---|
| 130 | return table + _data + col + _data + rowid
|
---|
| 131 |
|
---|
| 132 | def _search_col_data_key(table, col):
|
---|
| 133 | return table + _data + col + _data
|
---|
| 134 |
|
---|
| 135 | def _search_all_data_key(table):
|
---|
| 136 | return table + _data
|
---|
| 137 |
|
---|
| 138 | def _rowid_key(table, rowid):
|
---|
| 139 | return table + _rowid + rowid + _rowid
|
---|
| 140 |
|
---|
| 141 | def _search_rowid_key(table):
|
---|
| 142 | return table + _rowid
|
---|
| 143 |
|
---|
| 144 | def contains_metastrings(s) :
|
---|
| 145 | """Verify that the given string does not contain any
|
---|
| 146 | metadata strings that might interfere with dbtables database operation.
|
---|
| 147 | """
|
---|
| 148 | if (s.find(_table_names_key) >= 0 or
|
---|
| 149 | s.find(_columns) >= 0 or
|
---|
| 150 | s.find(_data) >= 0 or
|
---|
| 151 | s.find(_rowid) >= 0):
|
---|
| 152 | # Then
|
---|
| 153 | return 1
|
---|
| 154 | else:
|
---|
| 155 | return 0
|
---|
| 156 |
|
---|
| 157 |
|
---|
| 158 | class bsdTableDB :
|
---|
| 159 | def __init__(self, filename, dbhome, create=0, truncate=0, mode=0600,
|
---|
| 160 | recover=0, dbflags=0):
|
---|
| 161 | """bsdTableDB(filename, dbhome, create=0, truncate=0, mode=0600)
|
---|
| 162 |
|
---|
| 163 | Open database name in the dbhome Berkeley DB directory.
|
---|
| 164 | Use keyword arguments when calling this constructor.
|
---|
| 165 | """
|
---|
| 166 | self.db = None
|
---|
| 167 | myflags = db.DB_THREAD
|
---|
| 168 | if create:
|
---|
| 169 | myflags |= db.DB_CREATE
|
---|
| 170 | flagsforenv = (db.DB_INIT_MPOOL | db.DB_INIT_LOCK | db.DB_INIT_LOG |
|
---|
| 171 | db.DB_INIT_TXN | dbflags)
|
---|
| 172 | # DB_AUTO_COMMIT isn't a valid flag for env.open()
|
---|
| 173 | try:
|
---|
| 174 | dbflags |= db.DB_AUTO_COMMIT
|
---|
| 175 | except AttributeError:
|
---|
| 176 | pass
|
---|
| 177 | if recover:
|
---|
| 178 | flagsforenv = flagsforenv | db.DB_RECOVER
|
---|
| 179 | self.env = db.DBEnv()
|
---|
| 180 | # enable auto deadlock avoidance
|
---|
| 181 | self.env.set_lk_detect(db.DB_LOCK_DEFAULT)
|
---|
| 182 | self.env.open(dbhome, myflags | flagsforenv)
|
---|
| 183 | if truncate:
|
---|
| 184 | myflags |= db.DB_TRUNCATE
|
---|
| 185 | self.db = db.DB(self.env)
|
---|
| 186 | # this code relies on DBCursor.set* methods to raise exceptions
|
---|
| 187 | # rather than returning None
|
---|
| 188 | self.db.set_get_returns_none(1)
|
---|
| 189 | # allow duplicate entries [warning: be careful w/ metadata]
|
---|
| 190 | self.db.set_flags(db.DB_DUP)
|
---|
| 191 | self.db.open(filename, db.DB_BTREE, dbflags | myflags, mode)
|
---|
| 192 | self.dbfilename = filename
|
---|
| 193 |
|
---|
| 194 | if sys.version_info[0] >= 3 :
|
---|
| 195 | class cursor_py3k(object) :
|
---|
| 196 | def __init__(self, dbcursor) :
|
---|
| 197 | self._dbcursor = dbcursor
|
---|
| 198 |
|
---|
| 199 | def close(self) :
|
---|
| 200 | return self._dbcursor.close()
|
---|
| 201 |
|
---|
| 202 | def set_range(self, search) :
|
---|
| 203 | v = self._dbcursor.set_range(bytes(search, "iso8859-1"))
|
---|
[391] | 204 | if v is not None :
|
---|
[2] | 205 | v = (v[0].decode("iso8859-1"),
|
---|
| 206 | v[1].decode("iso8859-1"))
|
---|
| 207 | return v
|
---|
| 208 |
|
---|
| 209 | def __next__(self) :
|
---|
| 210 | v = getattr(self._dbcursor, "next")()
|
---|
[391] | 211 | if v is not None :
|
---|
[2] | 212 | v = (v[0].decode("iso8859-1"),
|
---|
| 213 | v[1].decode("iso8859-1"))
|
---|
| 214 | return v
|
---|
| 215 |
|
---|
| 216 | class db_py3k(object) :
|
---|
| 217 | def __init__(self, db) :
|
---|
| 218 | self._db = db
|
---|
| 219 |
|
---|
| 220 | def cursor(self, txn=None) :
|
---|
| 221 | return cursor_py3k(self._db.cursor(txn=txn))
|
---|
| 222 |
|
---|
| 223 | def has_key(self, key, txn=None) :
|
---|
| 224 | return getattr(self._db,"has_key")(bytes(key, "iso8859-1"),
|
---|
| 225 | txn=txn)
|
---|
| 226 |
|
---|
| 227 | def put(self, key, value, flags=0, txn=None) :
|
---|
| 228 | key = bytes(key, "iso8859-1")
|
---|
[391] | 229 | if value is not None :
|
---|
[2] | 230 | value = bytes(value, "iso8859-1")
|
---|
| 231 | return self._db.put(key, value, flags=flags, txn=txn)
|
---|
| 232 |
|
---|
| 233 | def put_bytes(self, key, value, txn=None) :
|
---|
| 234 | key = bytes(key, "iso8859-1")
|
---|
| 235 | return self._db.put(key, value, txn=txn)
|
---|
| 236 |
|
---|
| 237 | def get(self, key, txn=None, flags=0) :
|
---|
| 238 | key = bytes(key, "iso8859-1")
|
---|
| 239 | v = self._db.get(key, txn=txn, flags=flags)
|
---|
[391] | 240 | if v is not None :
|
---|
[2] | 241 | v = v.decode("iso8859-1")
|
---|
| 242 | return v
|
---|
| 243 |
|
---|
| 244 | def get_bytes(self, key, txn=None, flags=0) :
|
---|
| 245 | key = bytes(key, "iso8859-1")
|
---|
| 246 | return self._db.get(key, txn=txn, flags=flags)
|
---|
| 247 |
|
---|
| 248 | def delete(self, key, txn=None) :
|
---|
| 249 | key = bytes(key, "iso8859-1")
|
---|
| 250 | return self._db.delete(key, txn=txn)
|
---|
| 251 |
|
---|
| 252 | def close (self) :
|
---|
| 253 | return self._db.close()
|
---|
| 254 |
|
---|
| 255 | self.db = db_py3k(self.db)
|
---|
| 256 | else : # Python 2.x
|
---|
| 257 | pass
|
---|
| 258 |
|
---|
| 259 | # Initialize the table names list if this is a new database
|
---|
| 260 | txn = self.env.txn_begin()
|
---|
| 261 | try:
|
---|
| 262 | if not getattr(self.db, "has_key")(_table_names_key, txn):
|
---|
| 263 | getattr(self.db, "put_bytes", self.db.put) \
|
---|
| 264 | (_table_names_key, pickle.dumps([], 1), txn=txn)
|
---|
| 265 | # Yes, bare except
|
---|
| 266 | except:
|
---|
| 267 | txn.abort()
|
---|
| 268 | raise
|
---|
| 269 | else:
|
---|
| 270 | txn.commit()
|
---|
| 271 | # TODO verify more of the database's metadata?
|
---|
| 272 | self.__tablecolumns = {}
|
---|
| 273 |
|
---|
| 274 | def __del__(self):
|
---|
| 275 | self.close()
|
---|
| 276 |
|
---|
| 277 | def close(self):
|
---|
| 278 | if self.db is not None:
|
---|
| 279 | self.db.close()
|
---|
| 280 | self.db = None
|
---|
| 281 | if self.env is not None:
|
---|
| 282 | self.env.close()
|
---|
| 283 | self.env = None
|
---|
| 284 |
|
---|
| 285 | def checkpoint(self, mins=0):
|
---|
[391] | 286 | self.env.txn_checkpoint(mins)
|
---|
[2] | 287 |
|
---|
| 288 | def sync(self):
|
---|
[391] | 289 | self.db.sync()
|
---|
[2] | 290 |
|
---|
| 291 | def _db_print(self) :
|
---|
| 292 | """Print the database to stdout for debugging"""
|
---|
| 293 | print "******** Printing raw database for debugging ********"
|
---|
| 294 | cur = self.db.cursor()
|
---|
| 295 | try:
|
---|
| 296 | key, data = cur.first()
|
---|
| 297 | while 1:
|
---|
| 298 | print repr({key: data})
|
---|
| 299 | next = cur.next()
|
---|
| 300 | if next:
|
---|
| 301 | key, data = next
|
---|
| 302 | else:
|
---|
| 303 | cur.close()
|
---|
| 304 | return
|
---|
| 305 | except db.DBNotFoundError:
|
---|
| 306 | cur.close()
|
---|
| 307 |
|
---|
| 308 |
|
---|
| 309 | def CreateTable(self, table, columns):
|
---|
| 310 | """CreateTable(table, columns) - Create a new table in the database.
|
---|
| 311 |
|
---|
| 312 | raises TableDBError if it already exists or for other DB errors.
|
---|
| 313 | """
|
---|
| 314 | assert isinstance(columns, list)
|
---|
| 315 |
|
---|
| 316 | txn = None
|
---|
| 317 | try:
|
---|
| 318 | # checking sanity of the table and column names here on
|
---|
| 319 | # table creation will prevent problems elsewhere.
|
---|
| 320 | if contains_metastrings(table):
|
---|
| 321 | raise ValueError(
|
---|
| 322 | "bad table name: contains reserved metastrings")
|
---|
| 323 | for column in columns :
|
---|
| 324 | if contains_metastrings(column):
|
---|
| 325 | raise ValueError(
|
---|
| 326 | "bad column name: contains reserved metastrings")
|
---|
| 327 |
|
---|
| 328 | columnlist_key = _columns_key(table)
|
---|
| 329 | if getattr(self.db, "has_key")(columnlist_key):
|
---|
| 330 | raise TableAlreadyExists, "table already exists"
|
---|
| 331 |
|
---|
| 332 | txn = self.env.txn_begin()
|
---|
| 333 | # store the table's column info
|
---|
| 334 | getattr(self.db, "put_bytes", self.db.put)(columnlist_key,
|
---|
| 335 | pickle.dumps(columns, 1), txn=txn)
|
---|
| 336 |
|
---|
| 337 | # add the table name to the tablelist
|
---|
| 338 | tablelist = pickle.loads(getattr(self.db, "get_bytes",
|
---|
| 339 | self.db.get) (_table_names_key, txn=txn, flags=db.DB_RMW))
|
---|
| 340 | tablelist.append(table)
|
---|
| 341 | # delete 1st, in case we opened with DB_DUP
|
---|
| 342 | self.db.delete(_table_names_key, txn=txn)
|
---|
| 343 | getattr(self.db, "put_bytes", self.db.put)(_table_names_key,
|
---|
| 344 | pickle.dumps(tablelist, 1), txn=txn)
|
---|
| 345 |
|
---|
| 346 | txn.commit()
|
---|
| 347 | txn = None
|
---|
| 348 | except db.DBError, dberror:
|
---|
| 349 | if txn:
|
---|
| 350 | txn.abort()
|
---|
[391] | 351 | if sys.version_info < (2, 6) :
|
---|
[2] | 352 | raise TableDBError, dberror[1]
|
---|
| 353 | else :
|
---|
| 354 | raise TableDBError, dberror.args[1]
|
---|
| 355 |
|
---|
| 356 |
|
---|
| 357 | def ListTableColumns(self, table):
|
---|
| 358 | """Return a list of columns in the given table.
|
---|
| 359 | [] if the table doesn't exist.
|
---|
| 360 | """
|
---|
| 361 | assert isinstance(table, str)
|
---|
| 362 | if contains_metastrings(table):
|
---|
| 363 | raise ValueError, "bad table name: contains reserved metastrings"
|
---|
| 364 |
|
---|
| 365 | columnlist_key = _columns_key(table)
|
---|
| 366 | if not getattr(self.db, "has_key")(columnlist_key):
|
---|
| 367 | return []
|
---|
| 368 | pickledcolumnlist = getattr(self.db, "get_bytes",
|
---|
| 369 | self.db.get)(columnlist_key)
|
---|
| 370 | if pickledcolumnlist:
|
---|
| 371 | return pickle.loads(pickledcolumnlist)
|
---|
| 372 | else:
|
---|
| 373 | return []
|
---|
| 374 |
|
---|
| 375 | def ListTables(self):
|
---|
| 376 | """Return a list of tables in this database."""
|
---|
| 377 | pickledtablelist = self.db.get_get(_table_names_key)
|
---|
| 378 | if pickledtablelist:
|
---|
| 379 | return pickle.loads(pickledtablelist)
|
---|
| 380 | else:
|
---|
| 381 | return []
|
---|
| 382 |
|
---|
| 383 | def CreateOrExtendTable(self, table, columns):
|
---|
| 384 | """CreateOrExtendTable(table, columns)
|
---|
| 385 |
|
---|
| 386 | Create a new table in the database.
|
---|
| 387 |
|
---|
| 388 | If a table of this name already exists, extend it to have any
|
---|
| 389 | additional columns present in the given list as well as
|
---|
| 390 | all of its current columns.
|
---|
| 391 | """
|
---|
| 392 | assert isinstance(columns, list)
|
---|
| 393 |
|
---|
| 394 | try:
|
---|
| 395 | self.CreateTable(table, columns)
|
---|
| 396 | except TableAlreadyExists:
|
---|
| 397 | # the table already existed, add any new columns
|
---|
| 398 | txn = None
|
---|
| 399 | try:
|
---|
| 400 | columnlist_key = _columns_key(table)
|
---|
| 401 | txn = self.env.txn_begin()
|
---|
| 402 |
|
---|
| 403 | # load the current column list
|
---|
| 404 | oldcolumnlist = pickle.loads(
|
---|
| 405 | getattr(self.db, "get_bytes",
|
---|
| 406 | self.db.get)(columnlist_key, txn=txn, flags=db.DB_RMW))
|
---|
| 407 | # create a hash table for fast lookups of column names in the
|
---|
| 408 | # loop below
|
---|
| 409 | oldcolumnhash = {}
|
---|
| 410 | for c in oldcolumnlist:
|
---|
| 411 | oldcolumnhash[c] = c
|
---|
| 412 |
|
---|
| 413 | # create a new column list containing both the old and new
|
---|
| 414 | # column names
|
---|
| 415 | newcolumnlist = copy.copy(oldcolumnlist)
|
---|
| 416 | for c in columns:
|
---|
[391] | 417 | if not c in oldcolumnhash:
|
---|
[2] | 418 | newcolumnlist.append(c)
|
---|
| 419 |
|
---|
| 420 | # store the table's new extended column list
|
---|
| 421 | if newcolumnlist != oldcolumnlist :
|
---|
| 422 | # delete the old one first since we opened with DB_DUP
|
---|
| 423 | self.db.delete(columnlist_key, txn=txn)
|
---|
| 424 | getattr(self.db, "put_bytes", self.db.put)(columnlist_key,
|
---|
| 425 | pickle.dumps(newcolumnlist, 1),
|
---|
| 426 | txn=txn)
|
---|
| 427 |
|
---|
| 428 | txn.commit()
|
---|
| 429 | txn = None
|
---|
| 430 |
|
---|
| 431 | self.__load_column_info(table)
|
---|
| 432 | except db.DBError, dberror:
|
---|
| 433 | if txn:
|
---|
| 434 | txn.abort()
|
---|
[391] | 435 | if sys.version_info < (2, 6) :
|
---|
[2] | 436 | raise TableDBError, dberror[1]
|
---|
| 437 | else :
|
---|
| 438 | raise TableDBError, dberror.args[1]
|
---|
| 439 |
|
---|
| 440 |
|
---|
| 441 | def __load_column_info(self, table) :
|
---|
| 442 | """initialize the self.__tablecolumns dict"""
|
---|
| 443 | # check the column names
|
---|
| 444 | try:
|
---|
| 445 | tcolpickles = getattr(self.db, "get_bytes",
|
---|
| 446 | self.db.get)(_columns_key(table))
|
---|
| 447 | except db.DBNotFoundError:
|
---|
| 448 | raise TableDBError, "unknown table: %r" % (table,)
|
---|
| 449 | if not tcolpickles:
|
---|
| 450 | raise TableDBError, "unknown table: %r" % (table,)
|
---|
| 451 | self.__tablecolumns[table] = pickle.loads(tcolpickles)
|
---|
| 452 |
|
---|
| 453 | def __new_rowid(self, table, txn) :
|
---|
| 454 | """Create a new unique row identifier"""
|
---|
| 455 | unique = 0
|
---|
| 456 | while not unique:
|
---|
| 457 | # Generate a random 64-bit row ID string
|
---|
| 458 | # (note: might have <64 bits of true randomness
|
---|
| 459 | # but it's plenty for our database id needs!)
|
---|
| 460 | blist = []
|
---|
| 461 | for x in xrange(_rowid_str_len):
|
---|
| 462 | blist.append(random.randint(0,255))
|
---|
| 463 | newid = struct.pack('B'*_rowid_str_len, *blist)
|
---|
| 464 |
|
---|
| 465 | if sys.version_info[0] >= 3 :
|
---|
| 466 | newid = newid.decode("iso8859-1") # 8 bits
|
---|
| 467 |
|
---|
| 468 | # Guarantee uniqueness by adding this key to the database
|
---|
| 469 | try:
|
---|
| 470 | self.db.put(_rowid_key(table, newid), None, txn=txn,
|
---|
| 471 | flags=db.DB_NOOVERWRITE)
|
---|
| 472 | except db.DBKeyExistError:
|
---|
| 473 | pass
|
---|
| 474 | else:
|
---|
| 475 | unique = 1
|
---|
| 476 |
|
---|
| 477 | return newid
|
---|
| 478 |
|
---|
| 479 |
|
---|
| 480 | def Insert(self, table, rowdict) :
|
---|
| 481 | """Insert(table, datadict) - Insert a new row into the table
|
---|
| 482 | using the keys+values from rowdict as the column values.
|
---|
| 483 | """
|
---|
| 484 |
|
---|
| 485 | txn = None
|
---|
| 486 | try:
|
---|
| 487 | if not getattr(self.db, "has_key")(_columns_key(table)):
|
---|
| 488 | raise TableDBError, "unknown table"
|
---|
| 489 |
|
---|
| 490 | # check the validity of each column name
|
---|
[391] | 491 | if not table in self.__tablecolumns:
|
---|
[2] | 492 | self.__load_column_info(table)
|
---|
| 493 | for column in rowdict.keys() :
|
---|
| 494 | if not self.__tablecolumns[table].count(column):
|
---|
| 495 | raise TableDBError, "unknown column: %r" % (column,)
|
---|
| 496 |
|
---|
| 497 | # get a unique row identifier for this row
|
---|
| 498 | txn = self.env.txn_begin()
|
---|
| 499 | rowid = self.__new_rowid(table, txn=txn)
|
---|
| 500 |
|
---|
| 501 | # insert the row values into the table database
|
---|
| 502 | for column, dataitem in rowdict.items():
|
---|
| 503 | # store the value
|
---|
| 504 | self.db.put(_data_key(table, column, rowid), dataitem, txn=txn)
|
---|
| 505 |
|
---|
| 506 | txn.commit()
|
---|
| 507 | txn = None
|
---|
| 508 |
|
---|
| 509 | except db.DBError, dberror:
|
---|
| 510 | # WIBNI we could just abort the txn and re-raise the exception?
|
---|
| 511 | # But no, because TableDBError is not related to DBError via
|
---|
| 512 | # inheritance, so it would be backwards incompatible. Do the next
|
---|
| 513 | # best thing.
|
---|
| 514 | info = sys.exc_info()
|
---|
| 515 | if txn:
|
---|
| 516 | txn.abort()
|
---|
| 517 | self.db.delete(_rowid_key(table, rowid))
|
---|
[391] | 518 | if sys.version_info < (2, 6) :
|
---|
[2] | 519 | raise TableDBError, dberror[1], info[2]
|
---|
| 520 | else :
|
---|
| 521 | raise TableDBError, dberror.args[1], info[2]
|
---|
| 522 |
|
---|
| 523 |
|
---|
| 524 | def Modify(self, table, conditions={}, mappings={}):
|
---|
| 525 | """Modify(table, conditions={}, mappings={}) - Modify items in rows matching 'conditions' using mapping functions in 'mappings'
|
---|
| 526 |
|
---|
| 527 | * table - the table name
|
---|
| 528 | * conditions - a dictionary keyed on column names containing
|
---|
| 529 | a condition callable expecting the data string as an
|
---|
| 530 | argument and returning a boolean.
|
---|
| 531 | * mappings - a dictionary keyed on column names containing a
|
---|
| 532 | condition callable expecting the data string as an argument and
|
---|
| 533 | returning the new string for that column.
|
---|
| 534 | """
|
---|
| 535 |
|
---|
| 536 | try:
|
---|
| 537 | matching_rowids = self.__Select(table, [], conditions)
|
---|
| 538 |
|
---|
| 539 | # modify only requested columns
|
---|
| 540 | columns = mappings.keys()
|
---|
| 541 | for rowid in matching_rowids.keys():
|
---|
| 542 | txn = None
|
---|
| 543 | try:
|
---|
| 544 | for column in columns:
|
---|
| 545 | txn = self.env.txn_begin()
|
---|
| 546 | # modify the requested column
|
---|
| 547 | try:
|
---|
| 548 | dataitem = self.db.get(
|
---|
| 549 | _data_key(table, column, rowid),
|
---|
| 550 | txn=txn)
|
---|
| 551 | self.db.delete(
|
---|
| 552 | _data_key(table, column, rowid),
|
---|
| 553 | txn=txn)
|
---|
| 554 | except db.DBNotFoundError:
|
---|
| 555 | # XXXXXXX row key somehow didn't exist, assume no
|
---|
| 556 | # error
|
---|
| 557 | dataitem = None
|
---|
| 558 | dataitem = mappings[column](dataitem)
|
---|
[391] | 559 | if dataitem is not None:
|
---|
[2] | 560 | self.db.put(
|
---|
| 561 | _data_key(table, column, rowid),
|
---|
| 562 | dataitem, txn=txn)
|
---|
| 563 | txn.commit()
|
---|
| 564 | txn = None
|
---|
| 565 |
|
---|
| 566 | # catch all exceptions here since we call unknown callables
|
---|
| 567 | except:
|
---|
| 568 | if txn:
|
---|
| 569 | txn.abort()
|
---|
| 570 | raise
|
---|
| 571 |
|
---|
| 572 | except db.DBError, dberror:
|
---|
[391] | 573 | if sys.version_info < (2, 6) :
|
---|
[2] | 574 | raise TableDBError, dberror[1]
|
---|
| 575 | else :
|
---|
| 576 | raise TableDBError, dberror.args[1]
|
---|
| 577 |
|
---|
| 578 | def Delete(self, table, conditions={}):
|
---|
| 579 | """Delete(table, conditions) - Delete items matching the given
|
---|
| 580 | conditions from the table.
|
---|
| 581 |
|
---|
| 582 | * conditions - a dictionary keyed on column names containing
|
---|
| 583 | condition functions expecting the data string as an
|
---|
| 584 | argument and returning a boolean.
|
---|
| 585 | """
|
---|
| 586 |
|
---|
| 587 | try:
|
---|
| 588 | matching_rowids = self.__Select(table, [], conditions)
|
---|
| 589 |
|
---|
| 590 | # delete row data from all columns
|
---|
| 591 | columns = self.__tablecolumns[table]
|
---|
| 592 | for rowid in matching_rowids.keys():
|
---|
| 593 | txn = None
|
---|
| 594 | try:
|
---|
| 595 | txn = self.env.txn_begin()
|
---|
| 596 | for column in columns:
|
---|
| 597 | # delete the data key
|
---|
| 598 | try:
|
---|
| 599 | self.db.delete(_data_key(table, column, rowid),
|
---|
| 600 | txn=txn)
|
---|
| 601 | except db.DBNotFoundError:
|
---|
| 602 | # XXXXXXX column may not exist, assume no error
|
---|
| 603 | pass
|
---|
| 604 |
|
---|
| 605 | try:
|
---|
| 606 | self.db.delete(_rowid_key(table, rowid), txn=txn)
|
---|
| 607 | except db.DBNotFoundError:
|
---|
| 608 | # XXXXXXX row key somehow didn't exist, assume no error
|
---|
| 609 | pass
|
---|
| 610 | txn.commit()
|
---|
| 611 | txn = None
|
---|
| 612 | except db.DBError, dberror:
|
---|
| 613 | if txn:
|
---|
| 614 | txn.abort()
|
---|
| 615 | raise
|
---|
| 616 | except db.DBError, dberror:
|
---|
[391] | 617 | if sys.version_info < (2, 6) :
|
---|
[2] | 618 | raise TableDBError, dberror[1]
|
---|
| 619 | else :
|
---|
| 620 | raise TableDBError, dberror.args[1]
|
---|
| 621 |
|
---|
| 622 |
|
---|
| 623 | def Select(self, table, columns, conditions={}):
|
---|
| 624 | """Select(table, columns, conditions) - retrieve specific row data
|
---|
| 625 | Returns a list of row column->value mapping dictionaries.
|
---|
| 626 |
|
---|
| 627 | * columns - a list of which column data to return. If
|
---|
| 628 | columns is None, all columns will be returned.
|
---|
| 629 | * conditions - a dictionary keyed on column names
|
---|
| 630 | containing callable conditions expecting the data string as an
|
---|
| 631 | argument and returning a boolean.
|
---|
| 632 | """
|
---|
| 633 | try:
|
---|
[391] | 634 | if not table in self.__tablecolumns:
|
---|
[2] | 635 | self.__load_column_info(table)
|
---|
| 636 | if columns is None:
|
---|
| 637 | columns = self.__tablecolumns[table]
|
---|
| 638 | matching_rowids = self.__Select(table, columns, conditions)
|
---|
| 639 | except db.DBError, dberror:
|
---|
[391] | 640 | if sys.version_info < (2, 6) :
|
---|
[2] | 641 | raise TableDBError, dberror[1]
|
---|
| 642 | else :
|
---|
| 643 | raise TableDBError, dberror.args[1]
|
---|
| 644 | # return the matches as a list of dictionaries
|
---|
| 645 | return matching_rowids.values()
|
---|
| 646 |
|
---|
| 647 |
|
---|
| 648 | def __Select(self, table, columns, conditions):
|
---|
| 649 | """__Select() - Used to implement Select and Delete (above)
|
---|
| 650 | Returns a dictionary keyed on rowids containing dicts
|
---|
| 651 | holding the row data for columns listed in the columns param
|
---|
| 652 | that match the given conditions.
|
---|
| 653 | * conditions is a dictionary keyed on column names
|
---|
| 654 | containing callable conditions expecting the data string as an
|
---|
| 655 | argument and returning a boolean.
|
---|
| 656 | """
|
---|
| 657 | # check the validity of each column name
|
---|
[391] | 658 | if not table in self.__tablecolumns:
|
---|
[2] | 659 | self.__load_column_info(table)
|
---|
| 660 | if columns is None:
|
---|
| 661 | columns = self.tablecolumns[table]
|
---|
| 662 | for column in (columns + conditions.keys()):
|
---|
| 663 | if not self.__tablecolumns[table].count(column):
|
---|
| 664 | raise TableDBError, "unknown column: %r" % (column,)
|
---|
| 665 |
|
---|
| 666 | # keyed on rows that match so far, containings dicts keyed on
|
---|
| 667 | # column names containing the data for that row and column.
|
---|
| 668 | matching_rowids = {}
|
---|
| 669 | # keys are rowids that do not match
|
---|
| 670 | rejected_rowids = {}
|
---|
| 671 |
|
---|
| 672 | # attempt to sort the conditions in such a way as to minimize full
|
---|
| 673 | # column lookups
|
---|
| 674 | def cmp_conditions(atuple, btuple):
|
---|
| 675 | a = atuple[1]
|
---|
| 676 | b = btuple[1]
|
---|
| 677 | if type(a) is type(b):
|
---|
[391] | 678 |
|
---|
| 679 | # Needed for python 3. "cmp" vanished in 3.0.1
|
---|
| 680 | def cmp(a, b) :
|
---|
| 681 | if a==b : return 0
|
---|
| 682 | if a<b : return -1
|
---|
| 683 | return 1
|
---|
| 684 |
|
---|
[2] | 685 | if isinstance(a, PrefixCond) and isinstance(b, PrefixCond):
|
---|
| 686 | # longest prefix first
|
---|
| 687 | return cmp(len(b.prefix), len(a.prefix))
|
---|
| 688 | if isinstance(a, LikeCond) and isinstance(b, LikeCond):
|
---|
| 689 | # longest likestr first
|
---|
| 690 | return cmp(len(b.likestr), len(a.likestr))
|
---|
| 691 | return 0
|
---|
| 692 | if isinstance(a, ExactCond):
|
---|
| 693 | return -1
|
---|
| 694 | if isinstance(b, ExactCond):
|
---|
| 695 | return 1
|
---|
| 696 | if isinstance(a, PrefixCond):
|
---|
| 697 | return -1
|
---|
| 698 | if isinstance(b, PrefixCond):
|
---|
| 699 | return 1
|
---|
| 700 | # leave all unknown condition callables alone as equals
|
---|
| 701 | return 0
|
---|
| 702 |
|
---|
[391] | 703 | if sys.version_info < (2, 6) :
|
---|
[2] | 704 | conditionlist = conditions.items()
|
---|
| 705 | conditionlist.sort(cmp_conditions)
|
---|
| 706 | else : # Insertion Sort. Please, improve
|
---|
| 707 | conditionlist = []
|
---|
| 708 | for i in conditions.items() :
|
---|
| 709 | for j, k in enumerate(conditionlist) :
|
---|
| 710 | r = cmp_conditions(k, i)
|
---|
| 711 | if r == 1 :
|
---|
| 712 | conditionlist.insert(j, i)
|
---|
| 713 | break
|
---|
| 714 | else :
|
---|
| 715 | conditionlist.append(i)
|
---|
| 716 |
|
---|
| 717 | # Apply conditions to column data to find what we want
|
---|
| 718 | cur = self.db.cursor()
|
---|
| 719 | column_num = -1
|
---|
| 720 | for column, condition in conditionlist:
|
---|
| 721 | column_num = column_num + 1
|
---|
| 722 | searchkey = _search_col_data_key(table, column)
|
---|
| 723 | # speedup: don't linear search columns within loop
|
---|
| 724 | if column in columns:
|
---|
| 725 | savethiscolumndata = 1 # save the data for return
|
---|
| 726 | else:
|
---|
| 727 | savethiscolumndata = 0 # data only used for selection
|
---|
| 728 |
|
---|
| 729 | try:
|
---|
| 730 | key, data = cur.set_range(searchkey)
|
---|
| 731 | while key[:len(searchkey)] == searchkey:
|
---|
| 732 | # extract the rowid from the key
|
---|
| 733 | rowid = key[-_rowid_str_len:]
|
---|
| 734 |
|
---|
[391] | 735 | if not rowid in rejected_rowids:
|
---|
[2] | 736 | # if no condition was specified or the condition
|
---|
| 737 | # succeeds, add row to our match list.
|
---|
| 738 | if not condition or condition(data):
|
---|
[391] | 739 | if not rowid in matching_rowids:
|
---|
[2] | 740 | matching_rowids[rowid] = {}
|
---|
| 741 | if savethiscolumndata:
|
---|
| 742 | matching_rowids[rowid][column] = data
|
---|
| 743 | else:
|
---|
[391] | 744 | if rowid in matching_rowids:
|
---|
[2] | 745 | del matching_rowids[rowid]
|
---|
| 746 | rejected_rowids[rowid] = rowid
|
---|
| 747 |
|
---|
| 748 | key, data = cur.next()
|
---|
| 749 |
|
---|
| 750 | except db.DBError, dberror:
|
---|
[391] | 751 | if dberror.args[0] != db.DB_NOTFOUND:
|
---|
| 752 | raise
|
---|
[2] | 753 | continue
|
---|
| 754 |
|
---|
| 755 | cur.close()
|
---|
| 756 |
|
---|
| 757 | # we're done selecting rows, garbage collect the reject list
|
---|
| 758 | del rejected_rowids
|
---|
| 759 |
|
---|
| 760 | # extract any remaining desired column data from the
|
---|
| 761 | # database for the matching rows.
|
---|
| 762 | if len(columns) > 0:
|
---|
| 763 | for rowid, rowdata in matching_rowids.items():
|
---|
| 764 | for column in columns:
|
---|
[391] | 765 | if column in rowdata:
|
---|
[2] | 766 | continue
|
---|
| 767 | try:
|
---|
| 768 | rowdata[column] = self.db.get(
|
---|
| 769 | _data_key(table, column, rowid))
|
---|
| 770 | except db.DBError, dberror:
|
---|
[391] | 771 | if sys.version_info < (2, 6) :
|
---|
[2] | 772 | if dberror[0] != db.DB_NOTFOUND:
|
---|
| 773 | raise
|
---|
| 774 | else :
|
---|
| 775 | if dberror.args[0] != db.DB_NOTFOUND:
|
---|
| 776 | raise
|
---|
| 777 | rowdata[column] = None
|
---|
| 778 |
|
---|
| 779 | # return the matches
|
---|
| 780 | return matching_rowids
|
---|
| 781 |
|
---|
| 782 |
|
---|
| 783 | def Drop(self, table):
|
---|
| 784 | """Remove an entire table from the database"""
|
---|
| 785 | txn = None
|
---|
| 786 | try:
|
---|
| 787 | txn = self.env.txn_begin()
|
---|
| 788 |
|
---|
| 789 | # delete the column list
|
---|
| 790 | self.db.delete(_columns_key(table), txn=txn)
|
---|
| 791 |
|
---|
| 792 | cur = self.db.cursor(txn)
|
---|
| 793 |
|
---|
| 794 | # delete all keys containing this tables column and row info
|
---|
| 795 | table_key = _search_all_data_key(table)
|
---|
| 796 | while 1:
|
---|
| 797 | try:
|
---|
| 798 | key, data = cur.set_range(table_key)
|
---|
| 799 | except db.DBNotFoundError:
|
---|
| 800 | break
|
---|
| 801 | # only delete items in this table
|
---|
| 802 | if key[:len(table_key)] != table_key:
|
---|
| 803 | break
|
---|
| 804 | cur.delete()
|
---|
| 805 |
|
---|
| 806 | # delete all rowids used by this table
|
---|
| 807 | table_key = _search_rowid_key(table)
|
---|
| 808 | while 1:
|
---|
| 809 | try:
|
---|
| 810 | key, data = cur.set_range(table_key)
|
---|
| 811 | except db.DBNotFoundError:
|
---|
| 812 | break
|
---|
| 813 | # only delete items in this table
|
---|
| 814 | if key[:len(table_key)] != table_key:
|
---|
| 815 | break
|
---|
| 816 | cur.delete()
|
---|
| 817 |
|
---|
| 818 | cur.close()
|
---|
| 819 |
|
---|
| 820 | # delete the tablename from the table name list
|
---|
| 821 | tablelist = pickle.loads(
|
---|
| 822 | getattr(self.db, "get_bytes", self.db.get)(_table_names_key,
|
---|
| 823 | txn=txn, flags=db.DB_RMW))
|
---|
| 824 | try:
|
---|
| 825 | tablelist.remove(table)
|
---|
| 826 | except ValueError:
|
---|
| 827 | # hmm, it wasn't there, oh well, that's what we want.
|
---|
| 828 | pass
|
---|
| 829 | # delete 1st, incase we opened with DB_DUP
|
---|
| 830 | self.db.delete(_table_names_key, txn=txn)
|
---|
| 831 | getattr(self.db, "put_bytes", self.db.put)(_table_names_key,
|
---|
| 832 | pickle.dumps(tablelist, 1), txn=txn)
|
---|
| 833 |
|
---|
| 834 | txn.commit()
|
---|
| 835 | txn = None
|
---|
| 836 |
|
---|
[391] | 837 | if table in self.__tablecolumns:
|
---|
[2] | 838 | del self.__tablecolumns[table]
|
---|
| 839 |
|
---|
| 840 | except db.DBError, dberror:
|
---|
| 841 | if txn:
|
---|
| 842 | txn.abort()
|
---|
[391] | 843 | raise TableDBError(dberror.args[1])
|
---|