#-----------------------------------------------------------------------
# Copyright (C) 2000, 2001 by Autonomous Zone Industries
# Copyright (C) 2002 Gregory P. Smith
# License: This is free software. You may use this software for any
# purpose including modification/redistribution, so long as
# this header remains intact and that you do not claim any
# rights of ownership or authorship of this software. This
# software has been tested, but no warranty is expressed or
# -- Gregory P. Smith <greg@krypto.org>
# This provides a simple database table interface built on top of
# the Python Berkeley DB 3 interface.
if sys.version_info[0] >= 3 :
if sys.version_info < (2, 6) :
# When we drop support for python 2.4
# we could use: (in 2.5 we need a __future__ statement)
# with warnings.catch_warnings():
# warnings.filterwarnings(...)
# We can not use "with" as is, because it would be invalid syntax
# in python 2.4 and (with no __future__) 2.5.
# Here we simulate "with" following PEP 343 :
w = warnings.catch_warnings()
warnings.filterwarnings('ignore',
message='the cPickle module has been removed in Python 3.0',
category=DeprecationWarning)
# For Pythons w/distutils pybsddb
class TableDBError(StandardError):
class TableAlreadyExists(TableDBError):
"""This condition matches everything"""
"""Acts as an exact match condition function"""
def __init__(self, strtomatch):
self.strtomatch = strtomatch
return s == self.strtomatch
"""Acts as a condition function for matching a string prefix"""
def __init__(self, prefix):
return s[:len(self.prefix)] == self.prefix
"""Acts as a condition function for matching a string postfix"""
def __init__(self, postfix):
return s[-len(self.postfix):] == self.postfix
Acts as a function that will match using an SQL 'LIKE' style
string. Case insensitive and % signs are wild cards.
This isn't perfect but it should work for the simple common cases.
def __init__(self, likestr, re_flags=re.IGNORECASE):
# escape python re characters
chars_to_escape = '.*+()[]?'
for char in chars_to_escape :
likestr = likestr.replace(char, '\\'+char)
# convert %s to wildcards
self.likestr = likestr.replace('%', '.*')
self.re = re.compile('^'+self.likestr+'$', re_flags)
# keys used to store database metadata
_table_names_key = '__TABLE_NAMES__' # list of the tables in this db
_columns = '._COLUMNS__' # table_name+this key contains a list of columns
# these keys are found within table sub databases
_data = '._DATA_.' # this+column+this+rowid key contains table data
_rowid = '._ROWID_.' # this+rowid+this key contains a unique entry for each
# row in the table. (no data is stored)
_rowid_str_len = 8 # length in bytes of the unique rowid strings
def _data_key(table, col, rowid):
return table + _data + col + _data + rowid
def _search_col_data_key(table, col):
return table + _data + col + _data
def _search_all_data_key(table):
def _rowid_key(table, rowid):
return table + _rowid + rowid + _rowid
def _search_rowid_key(table):
def contains_metastrings(s) :
"""Verify that the given string does not contain any
metadata strings that might interfere with dbtables database operation.
if (s.find(_table_names_key) >= 0 or
def __init__(self, filename, dbhome, create=0, truncate=0, mode=0600,
"""bsdTableDB(filename, dbhome, create=0, truncate=0, mode=0600)
Open database name in the dbhome Berkeley DB directory.
Use keyword arguments when calling this constructor.
flagsforenv = (db.DB_INIT_MPOOL | db.DB_INIT_LOCK | db.DB_INIT_LOG |
db.DB_INIT_TXN | dbflags)
# DB_AUTO_COMMIT isn't a valid flag for env.open()
dbflags |= db.DB_AUTO_COMMIT
flagsforenv = flagsforenv | db.DB_RECOVER
# enable auto deadlock avoidance
self.env.set_lk_detect(db.DB_LOCK_DEFAULT)
self.env.open(dbhome, myflags | flagsforenv)
myflags |= db.DB_TRUNCATE
self.db = db.DB(self.env)
# this code relies on DBCursor.set* methods to raise exceptions
# rather than returning None
self.db.set_get_returns_none(1)
# allow duplicate entries [warning: be careful w/ metadata]
self.db.set_flags(db.DB_DUP)
self.db.open(filename, db.DB_BTREE, dbflags | myflags, mode)
self.dbfilename = filename
if sys.version_info[0] >= 3 :
class cursor_py3k(object) :
def __init__(self, dbcursor) :
self._dbcursor = dbcursor
return self._dbcursor.close()
def set_range(self, search) :
v = self._dbcursor.set_range(bytes(search, "iso8859-1"))
v = (v[0].decode("iso8859-1"),
v[1].decode("iso8859-1"))
v = getattr(self._dbcursor, "next")()
v = (v[0].decode("iso8859-1"),
v[1].decode("iso8859-1"))
def cursor(self, txn=None) :
return cursor_py3k(self._db.cursor(txn=txn))
def has_key(self, key, txn=None) :
return getattr(self._db,"has_key")(bytes(key, "iso8859-1"),
def put(self, key, value, flags=0, txn=None) :
key = bytes(key, "iso8859-1")
value = bytes(value, "iso8859-1")
return self._db.put(key, value, flags=flags, txn=txn)
def put_bytes(self, key, value, txn=None) :
key = bytes(key, "iso8859-1")
return self._db.put(key, value, txn=txn)
def get(self, key, txn=None, flags=0) :
key = bytes(key, "iso8859-1")
v = self._db.get(key, txn=txn, flags=flags)
v = v.decode("iso8859-1")
def get_bytes(self, key, txn=None, flags=0) :
key = bytes(key, "iso8859-1")
return self._db.get(key, txn=txn, flags=flags)
def delete(self, key, txn=None) :
key = bytes(key, "iso8859-1")
return self._db.delete(key, txn=txn)
self.db = db_py3k(self.db)
# Initialize the table names list if this is a new database
txn = self.env.txn_begin()
if not getattr(self.db, "has_key")(_table_names_key, txn):
getattr(self.db, "put_bytes", self.db.put) \
(_table_names_key, pickle.dumps([], 1), txn=txn)
# TODO verify more of the database's metadata?
def checkpoint(self, mins=0):
self.env.txn_checkpoint(mins)
"""Print the database to stdout for debugging"""
print "******** Printing raw database for debugging ********"
except db.DBNotFoundError:
def CreateTable(self, table, columns):
"""CreateTable(table, columns) - Create a new table in the database.
raises TableDBError if it already exists or for other DB errors.
assert isinstance(columns, list)
# checking sanity of the table and column names here on
# table creation will prevent problems elsewhere.
if contains_metastrings(table):
"bad table name: contains reserved metastrings")
if contains_metastrings(column):
"bad column name: contains reserved metastrings")
columnlist_key = _columns_key(table)
if getattr(self.db, "has_key")(columnlist_key):
raise TableAlreadyExists, "table already exists"
txn = self.env.txn_begin()
# store the table's column info
getattr(self.db, "put_bytes", self.db.put)(columnlist_key,
pickle.dumps(columns, 1), txn=txn)
# add the table name to the tablelist
tablelist = pickle.loads(getattr(self.db, "get_bytes",
self.db.get) (_table_names_key, txn=txn, flags=db.DB_RMW))
# delete 1st, in case we opened with DB_DUP
self.db.delete(_table_names_key, txn=txn)
getattr(self.db, "put_bytes", self.db.put)(_table_names_key,
pickle.dumps(tablelist, 1), txn=txn)
except db.DBError, dberror:
if sys.version_info < (2, 6) :
raise TableDBError, dberror[1]
raise TableDBError, dberror.args[1]
def ListTableColumns(self, table):
"""Return a list of columns in the given table.
[] if the table doesn't exist.
assert isinstance(table, str)
if contains_metastrings(table):
raise ValueError, "bad table name: contains reserved metastrings"
columnlist_key = _columns_key(table)
if not getattr(self.db, "has_key")(columnlist_key):
pickledcolumnlist = getattr(self.db, "get_bytes",
self.db.get)(columnlist_key)
return pickle.loads(pickledcolumnlist)
"""Return a list of tables in this database."""
pickledtablelist = self.db.get_get(_table_names_key)
return pickle.loads(pickledtablelist)
def CreateOrExtendTable(self, table, columns):
"""CreateOrExtendTable(table, columns)
Create a new table in the database.
If a table of this name already exists, extend it to have any
additional columns present in the given list as well as
all of its current columns.
assert isinstance(columns, list)
self.CreateTable(table, columns)
except TableAlreadyExists:
# the table already existed, add any new columns
columnlist_key = _columns_key(table)
txn = self.env.txn_begin()
# load the current column list
oldcolumnlist = pickle.loads(
getattr(self.db, "get_bytes",
self.db.get)(columnlist_key, txn=txn, flags=db.DB_RMW))
# create a hash table for fast lookups of column names in the
# create a new column list containing both the old and new
newcolumnlist = copy.copy(oldcolumnlist)
if not c in oldcolumnhash:
# store the table's new extended column list
if newcolumnlist != oldcolumnlist :
# delete the old one first since we opened with DB_DUP
self.db.delete(columnlist_key, txn=txn)
getattr(self.db, "put_bytes", self.db.put)(columnlist_key,
pickle.dumps(newcolumnlist, 1),
self.__load_column_info(table)
except db.DBError, dberror:
if sys.version_info < (2, 6) :
raise TableDBError, dberror[1]
raise TableDBError, dberror.args[1]
def __load_column_info(self, table) :
"""initialize the self.__tablecolumns dict"""
tcolpickles = getattr(self.db, "get_bytes",
self.db.get)(_columns_key(table))
except db.DBNotFoundError:
raise TableDBError, "unknown table: %r" % (table,)
raise TableDBError, "unknown table: %r" % (table,)
self.__tablecolumns[table] = pickle.loads(tcolpickles)
def __new_rowid(self, table, txn) :
"""Create a new unique row identifier"""
# Generate a random 64-bit row ID string
# (note: might have <64 bits of true randomness
# but it's plenty for our database id needs!)
for x in xrange(_rowid_str_len):
blist.append(random.randint(0,255))
newid = struct.pack('B'*_rowid_str_len, *blist)
if sys.version_info[0] >= 3 :
newid = newid.decode("iso8859-1") # 8 bits
# Guarantee uniqueness by adding this key to the database
self.db.put(_rowid_key(table, newid), None, txn=txn,
except db.DBKeyExistError:
def Insert(self, table, rowdict) :
"""Insert(table, datadict) - Insert a new row into the table
using the keys+values from rowdict as the column values.
if not getattr(self.db, "has_key")(_columns_key(table)):
raise TableDBError, "unknown table"
# check the validity of each column name
if not table in self.__tablecolumns:
self.__load_column_info(table)
for column in rowdict.keys() :
if not self.__tablecolumns[table].count(column):
raise TableDBError, "unknown column: %r" % (column,)
# get a unique row identifier for this row
txn = self.env.txn_begin()
rowid = self.__new_rowid(table, txn=txn)
It is recommended that you Edit text format, this type of Fix handles quite a lot in one request
