Blitz is a document-oriented database for Python that is backend-agnostic. It comes with a flat-file database for JSON documents and provides MongoDB-like querying capabilities.





GitHub Stars



Last Commit

2yrs ago











Build Status PyPI Code Issues Python 3

BlitzDB, or just Blitz is a document-based, object-oriented, transactional database written purely in Python. Among other things, it provides a powerful querying language, deep indexing of documents, compressed data storage and automatic referencing of embedded documents. It is reasonably fast, can be easily embedded in any Python application and does not have any external dependencies (except when using a third-party backend). In addition, you can use it as a frontend to other database engines such as MongoDB in case you should need more power.

Go To Main Documentation

Key Features

  • Document-based, object-oriented interface.
  • Powerful and rich querying language.
  • Deep document indexes on arbitrary fields.
  • Compressed storage of documents.
  • Support for multiple backends (e.g. file-based storage, MongoDB).
  • Support for database transactions (currently only for the file-based backend).

Use Cases

Blitz can be used as a standalone document store for client application. Originally blitz was designed for use with the checkmate Python code analysis toolkit, where it stores statistical data. Since blitz stores all documents as single JSON files, it is possible to put the whole database under version-control.


The easiest way to install Blitz is through pip or easy_install

pip install blitzdb
easy_install blitzdb

For more detailed installation instructions, have a look at the documentation.

Detailed Documentation

The detailed documentation for this project is hosted on ReadTheDocs, feel free to take a look!


  • 0.4.4: SQL backend: Do not coerce server_default values via a CAST, as this can cause incompatibilities.
  • 0.4.3: Many small improvements to the SQL backend.
  • 0.3.0: Fully functional SQL backend.
  • 0.2.12: Added support for proper attribute iteration to Document.
  • 0.2.11: Allow setting the collection parameter through a Document.Meta attribute.
  • 0.2.10: Bugfix-Release: Fix Python 3 compatibility issue.
  • 0.2.9: Bugfix-Release: Fix serialization problem with file backend.
  • 0.2.8: Added get, has_key and clear methods to Document class
  • 0.2.7: Fixed problem with unicode function in Python 3.
  • 0.2.6: Bugfix-Release: Fixed an issue with the $exists operator for the file backend.
  • 0.2.5: Bugfix-Release
  • 0.2.4: Added support for projections and update operations to the MongoDB backend.
  • 0.2.3: Bugfix-Release: Fixed bug in transaction data caching in MongoDB backend.
  • 0.2.2: Fix for slice operators in MongoDB backend.
  • 0.2.1: Better tests.
  • 0.2.0: Support for including additional information in DB references. Support for accessing document attributes as dictionary items. Added $regex parameter that allows to use regular expressions in queries.
  • 0.1.5: MongoDB backend now supports database transactions. Database operations are now read-isolated by default, i.e. uncommitted operations will not affect database queries before they are committed.
  • 0.1.4: Improved indexing of objects for the file backend, added support for automatic serialization/deserialization of object attributes when adding keys to or querying an index.
  • 0.1.3: Sorting of query sets is now supported (still experimental)
  • 0.1.2: Small bugfixes, BlitzDB version number now contained in DB config dict
  • 0.1.1: BlitzDB is now Python3 compatible (thanks to David Koblas)

Contributors (in alphabetical order)

  • @bwiessneth
  • Florian Lehmann - @cashaddy
  • Karskrin - @cBrauge
  • Chris Mutel - @cmutel
  • Cecil Woebker - @cwoebker
  • Ethan Blackburn - @EthanBlackburn
  • Javier Collado - @jcollado
  • Jason Xie - @jxieeducation
  • David Koblas - @koblas
  • Stéphane Wirtel - @matrixise
  • Victor Miclovich - @miclovich
  • Dmytro Kyrychuk - @orgkhnargh
  • Christoph Neumann - @programmdesign
  • Dale - @puredistortion
  • tjado - @tejado
  • Thomas Ballinger - @thomasballinger
  • Tyler Kennedy - @TkTech
  • Toby Champion - @tobych

Thanks for all your contributions, without you BlitzDB wouldn't be what it is today :)

Third-Party Contributions

  • Flask-BlitzDB Flask adapter for BlitzDB. Blitz + Flask = Awesome!


To get an idea of what you can do with Blitz, here are some examples.

Creating objects

from blitzdb import Document

class Movie(Document):

class Actor(Document):

the_godfather = Movie({'name': 'The Godfather','year':1972,'pk':1L})

marlon_brando = Actor({'name':'Marlon Brando','pk':1L})
al_pacino = Actor({'name' : 'Al Pacino','pk':1L})

Storing objects in the database:

from blitzdb import FileBackend

backend = FileBackend("/path/to/my/db")

Retrieving objects from the database:

the_godfather = backend.get(Movie,{'pk':1L})
the_godfather = backend.get(Movie,{'name' : 'The Godfather'})

Filtering objects

movies_from_1972 = backend.filter(Movie,{'year' : 1972})

Working with transactions

the_godfather.director = 'Roland Emmerich' #oops...
backend.rollback() #undo the changes...

Creating nested object references

the_godfather.cast = {'Don Vito Corleone' : marlon_brando, 'Michael Corleone' : al_pacino}

#Documents stored within other objects will be automatically converted to database references.

marlon_brando.performances = [the_godfather]
al_pacino.performances = [the_godfather]
#Will store references to the movies within the documents in the DB

Creation of database indexes and advanced querying

#Will create an index on the 'performances' field, for fast querying

godfather_cast = backend.filter(Actor,{'movies' : the_godfather})
#Will return 'Al Pacino' and 'Marlon Brando'

Arbitrary filter expressions

star_wars_iv = Movie({'name' : 'Star Wars - Episode IV: A New Hope','year': 1977})

movies_from_the_seventies = backend.filter(Movie,{'year': lambda year : year >= 1970 and year < 1980})
#Will return Star Wars & The Godfather (man, what a decade!)

Rate & Review

Great Documentation0
Easy to Use0
Highly Customizable0
Bleeding Edge0
Responsive Maintainers0
Poor Documentation0
Hard to Use0
Unwelcoming Community0