Python:Extend the 'dict' class

You can either subclass dict or UserDict, since van already talked about UserDict, lets look at dict.

Type help(dict) into an interpreter and you see a big list of methods. You will need to override all the methods that modify the dict as well as the methods that iterate over the dict.

Methods that modify the dict include __delitem__,__setitem__,clear etc.

Methods that iterate the dict include __iter__,keys,values,items etc.

This should get you started

>>> class odict(dict):
...     def __init__(self, *args, **kw):
...         super(odict,self).__init__(*args, **kw)
...         self.itemlist = super(odict,self).keys()
...     def __setitem__(self, key, value):
...          # TODO: what should happen to the order if
...          #       the key is already in the dict       
...         self.itemlist.append(key)
...         super(odict,self).__setitem__(key, value)
...     def __iter__(self):
...         return iter(self.itemlist)
...     def keys(self):
...         return self.itemlist
...     def values(self):
...         return [self[key] for key in self]  
...     def itervalues(self):
...         return (self[key] for key in self)
... 
>>> od = odict(a=1,b=2)
>>> print od
{'a': 1, 'b': 2}
>>> od['d']=4
>>> od['c']=3
>>> print od   # look at the `__str__` and `__repr__` methods 
{'a': 1, 'c': 3, 'b': 2, 'd': 4}
>>> print od.keys()
['a', 'b', 'd', 'c']
>>> print od.values()
[1, 2, 4, 3]

The implementation of dict will not help you with the task. What you want is a class that has the same interface as dict, but a different implementation. That will require to implement methods like __getitem__, __setitem__, etc. If you Google for "ordereddict", you will find a lot of examples.

If you use python 2.7+, then see collections.OrderedDict.
Otherwise, backport (copy the source) or see Recipe 576693: Ordered Dictionary for Py2.4 (Python) .

But if you really need to extend the dict, then start with UserDict, source for which you can find in /lib/UserDict.py of your python distribution (Lib/collections/__init__.py with Python 3).