[ZODB-Dev] Object uniqueness

Markus Läll markus3 at ut.ee
Fri Oct 3 08:57:11 EDT 2008


I think I know now -- objects compare identical and equal only if they are
in the same container object. (Check the code below)

I don't mean to be negative, but why can't ZODB be like a big RAM? Like
when i store an object to multiple places on the same DB, then it would
actually be the same in each place. (Like if you comment out the "Commit,
close and reload" part of the below code, all the comparisons evaluate to
True)

Markus

# -*- coding: cp1252 -*-
# ZODB
from ZODB import FileStorage, DB
import transaction

# Trees
from BTrees.IOBTree import IOBTree as IO
from BTrees.OOBTree import OOBTree as OO, OOSet as OSet

# Open db
storage = FileStorage.FileStorage('test.fs')
db = DB(storage)
conn = db.open()
root = conn.root()

# The  one and only class
class X(object):
    attr1 = 1
    def __init__(self, attr2):
        self.attr2 = attr2

# Initiate three containers, OOBTree and OOSet
root[0] = OO()
root[1] = OSet()
root[0]['subBTree'] = OO()

# The only object to be stored...
x = X('abc')

# ...and storing it
root[0]['a'] = x
root[0]['b'] = x
root[0]['subBTree']['a'] = x
root[1].update([ x ])

# Commit, close and reload
transaction.commit()
db.close()
storage = FileStorage.FileStorage('test.fs')
db = DB(storage)
conn = db.open()
root = conn.root()

# Shorthands...
a = root[0]['a']
asub = root[0]['subBTree']['a']
b = root[0]['b']
c = iter(root[1]).next()

# ...and the comparison
print 'a is asub:', a is asub
print 'a == asub:', a == asub

print 'a is b:', a is b
print 'a == b:', a == b

print 'a is c:', a is c
print 'a == c:', a == c




More information about the ZODB-Dev mailing list