Ordered dictionary.
Python Shell
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
src/odict Use class name instead of 'odict()' in __repr__. Jun 14, 2017
CHANGES.rst Bump version Jun 15, 2017
LICENSE.rst some cleanup, works with pypy Jan 3, 2015
MANIFEST.in prepare release, make checkmanifest happy Nov 25, 2016
README.rst Add coverage report to README Jun 7, 2017
bootstrap.sh Add run_coverage helper script Jun 3, 2017
buildout.cfg Minor package polishing May 18, 2017
run_coverage.sh Add run_coverage helper script Jun 3, 2017
run_tests.sh typos May 21, 2017
setup.cfg Minor package polishing May 18, 2017
setup.py Bump version Jun 15, 2017



Dictionary in which the insertion order of items is preserved (using an internal double linked list). In this implementation replacing an existing item keeps it at its original position.

Internal representation: values of the dict.

[pred_key, val, succ_key]

The sequence of elements uses as a double linked list. The links are dict keys. self.lh and self.lt are the keys of first and last element inserted in the odict.


When this package was created, collections.OrderedDict not existed yet.

Another problem is that dict cannot always be inherited from in conjunction with other base classes. This may result in instance lay-out conflicts or other errors. So odict is written in a way that let you alter the dictionary base implementation easily.

Performance (Python 3.4)

When running the benchmark script, you get results similar to the one below on recent common hardware.

Adding and deleting builtin dict objects

Add 1000 0.63800ms
Delete 1000 0.36900ms
Add 10000 5.74600ms
Delete 10000 3.97000ms
Add 100000 69.40600ms
Delete 100000 47.30000ms
Add 1000000 807.09100ms
Delete 1000000 495.33400ms

Adding and deleting collection.OrderedDict objects

Add 1000 8.15200ms
Delete 1000 0.50800ms
Add 10000 91.45800ms
Delete 10000 7.10200ms
Add 100000 982.35500ms
Delete 100000 71.02300ms
Add 1000000 10222.78300ms
Delete 1000000 715.35100ms

Adding and deleting odict objects provided by this package

Add 1000 46.75600ms
Delete 1000 0.35700ms
Add 10000 23.97900ms
Delete 10000 4.79100ms
Add 100000 276.15900ms
Delete 100000 49.02700ms
Add 1000000 3244.68600ms
Delete 1000000 539.16500ms

Relation dict : odict

creating 1000 objects 1:73.285
deleting 1000 objects 1: 0.967
creating 10000 objects 1: 4.173
deleting 10000 objects 1: 1.207
creating 100000 objects 1: 3.979
deleting 100000 objects 1: 1.037
creating 1000000 objects 1: 4.020
deleting 1000000 objects 1: 1.088

Relation OrderedDict : odict

creating 1000 objects 1: 5.736
deleting 1000 objects 1: 0.703
creating 10000 objects 1: 0.262
deleting 10000 objects 1: 0.675
creating 100000 objects 1: 0.281
deleting 100000 objects 1: 0.690
creating 1000000 objects 1: 0.317
deleting 1000000 objects 1: 0.754


Import and create ordered dictionary.

from odict import odict
od = odict()

type conversion to ordinary dict. This will fail.

>>> dict(odict([(1, 1)]))
{1: [nil, 1, nil]}

The reason for this is here -> http://bugs.python.org/issue1615701

The __init__ function of dict checks wether arg is subclass of dict, and ignores overwritten __getitem__ & co if so.

This was fixed and later reverted due to behavioural problems with pickle.

Use one of the following ways for type conversion.

>>> dict(odict([(1, 1)]).items())
{1: 1}

>>> odict([(1, 1)]).as_dict()
{1: 1}

It is possible to use abstract mixin class _odict to hook another dict base implementation. This is useful i.e. when persisting to ZODB. Inheriting from dict and Persistent at the same time fails.

from persistent.dict import PersistentDict
class podict(_odict, PersistentDict):

    def _dict_impl(self):
        return PersistentDict


In a C reimplementation of this data structure, things could be simplified (and speed up) a lot if given a value you can at the same time find its key. With that, you can use normal C pointers.



Summary of the test coverage report:

Name                    Stmts   Miss  Cover
src/odict/__init__.py       1      0   100%
src/odict/pyodict.py      237      0   100%
src/odict/tests.py        198      0   100%
TOTAL                     436      0   100%

Python Versions

  • Python 2.6+, 3.2+, pypy
  • May work with other versions (untested)


  • bearophile (Original Author)
  • Robert Niederreiter (Author)
  • Georg Bernhard
  • Florian Friesdorf
  • Jens Klein

under the Python Software Foundation License.