dictionary referencing error

[ckbcouture]

BPO	1012249
Nosy	@rhettinger

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2004-08-19.17:14:08.000>
created_at = <Date 2004-08-19.14:57:23.000>
labels = ['interpreter-core', 'invalid']
title = 'dictionary referencing error'
updated_at = <Date 2004-08-19.17:14:08.000>
user = 'https://bugs.python.org/ckbcouture'

bugs.python.org fields:

activity = <Date 2004-08-19.17:14:08.000>
actor = 'rhettinger'
assignee = 'none'
closed = True
closed_date = None
closer = None
components = ['Interpreter Core']
creation = <Date 2004-08-19.14:57:23.000>
creator = 'ckbcouture'
dependencies = []
files = []
hgrepos = []
issue_num = 1012249
keywords = []
message_count = 5.0
messages = ['22136', '22137', '22138', '22139', '22140']
nosy_count = 2.0
nosy_names = ['rhettinger', 'ckbcouture']
pr_nums = []
priority = 'normal'
resolution = 'not a bug'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue1012249'
versions = ['Python 2.3']

[ckbcouture]

When I write a class as such:

class MyObject:

   myDict = {}

...

Each new instance refers to the SAME dictionary, not a
newly minted one. However, this does not hold true for
other data types.

This has forced me to place dictionary declarations in
the __init__ method to ensure uniqueness.

Unless this is intended for some obscure reason, it
should be fixed.

Thanks,
Cameron

[rhettinger]

Logged In: YES
user_id=80475

That is the intended behavior. Attribute definitions inside
the class definition but outside the method definitions are
essentially class variables. They are useful because each
instance does not need to have its own copy. For cases
where an instance needs its own copy, your solution using
__init__ was the right thing to do.

In time, this will be obvious and all will feel as right as
rain.

[ckbcouture]

Logged In: YES
user_id=1096248

The problem is that this only seems to work for dictionaries,
tuples and lists. it DOES NOT work for numbers and strings.

Try the following:

class Test:

        mint = 1
        mchar = "A"
        mdict = {}
        mlist = []
        mtup = ()

        def __init__( self ):
                self.mint += 1
                self.mchar += self.mchar
                self.mdict[ self.mchar ] = self.mint
                self.mlist.append( self.mdict )
                self.mtup += ( self.mlist, )
                print self.mtup


for i in range(10):
        Test()

For me it outputs:

([{'AA': 2}],)
([{'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)

Which means that it is working as you say for some, but not
all data types. Is THAT intended?

Thanks,
Cameron

[ckbcouture]

Logged In: YES
user_id=1096248

The problem is that this only seems to work for dictionaries,
tuples and lists. it DOES NOT work for numbers and strings.

Try the following:

class Test:

        mint = 1
        mchar = "A"
        mdict = {}
        mlist = []
        mtup = ()

        def __init__( self ):
                self.mint += 1
                self.mchar += self.mchar
                self.mdict[ self.mchar ] = self.mint
                self.mlist.append( self.mdict )
                self.mtup += ( self.mlist, )
                print self.mtup


for i in range(10):
        Test()

For me it outputs:

([{'AA': 2}],)
([{'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)
([{'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA':
2}, {'AA': 2}, {'AA': 2}, {'AA': 2}, {'AA': 2}],)

Which means that it is working as you say for some, but not
all data types. Is THAT intended?

Thanks,
Cameron

[rhettinger]

Logged In: YES
user_id=80475

This is also the way it is supposed to be. At issue is
whether the object is mutable or not. If mutable, something
like .append() alters the object inplace (at the class
level). If immutable, something like a = a + b reads a from
the class level, adds b creating a newobject, and stores the
new object, a, back at the instance level.

Please work through the tutorial or make a newsgroup posting
requesting better explanations. This isn't a bug, it is
just something intrinsic to Python and is a minor step that
everyone has to go through on the path to mastering the
language.