Mongita Save

"Mongita is to MongoDB as SQLite is to SQL"

Project README

Mongita Logo

Version 1.1.1 Build passing Coverage 100% License BSD MacOS|Linux

Mongita is a lightweight embedded document database that implements a commonly-used subset of the MongoDB/PyMongo interface. Mongita differs from MongoDB in that instead of being a server, Mongita is a self-contained Python library. Mongita can be configured to store its documents either on disk or in memory.

"Mongita is to MongoDB as SQLite is to SQL"

Please report any bugs. Mongita is free and open source. You can contribute!


  • Embedded database: Mongita is a good alternative to SQLite for embedded applications when a document database makes more sense than a relational one.
  • Unit testing: Mocking PyMongo/MongoDB is a pain. Worse, mocking can hide real bugs. By monkey-patching PyMongo with Mongita, unit tests can be more faithful while remaining isolated.

Design goals

  • MongoDB compatibility: Mongita implements a commonly-used subset of the PyMongo API. This allows projects to be started with Mongita and later upgraded to MongoDB once they reach an appropriate scale.
  • Embedded/self-contained: Mongita does not require a server or start a process. It is just a Python library. To use it, just add import mongita to the top of your script.
  • Speed: Mongita is within an order of magnitude of MongoDB and Sqlite in 10k document benchmarks. See the performance section below.
  • Well tested: Mongita has 100% test coverage and more test code than library code.
  • Limited dependencies: Mongita runs anywhere that Python runs. Currently the only dependencies are pymongo (for bson) and sortedcontainers (for faster indexes).
  • Thread-safe: (EXPERIMENTAL) Mongita avoids race conditions by isolating certain document modification operations.

When NOT to use Mongita

  • You need a traditional server/client relationship: Mongita is an embedded database. It is not process-safe. When you have multiple clients, a traditional server/client database is the correct choice.
  • You run a lot of uncommon commands: Mongita implements a commonly used subset of MongoDB. While the goal is to eventually implement most of it, it will take some time to get there.
  • You need extreme performance: Mongita has comparable performance to MongoDB and SQLite for common operations. However, it's possible you'll find bottlenecks - especially with uncommon operations.


pip3 install mongita

Hello world

>>> from mongita import MongitaClientDisk
>>> client = MongitaClientDisk()
>>> hello_world_db = client.hello_world_db
>>> mongoose_collection = hello_world_db.mongoose_collection
>>> mongoose_collection.insert_many([{'name': 'Meercat', 'does_not_eat': 'Snakes'},
                                     {'name': 'Yellow mongoose', 'eats': 'Termites'}])
<mongita.results.InsertManyResult object at 0x000000000>
>>> mongoose_collection.count_documents({})
>>> mongoose_collection.update_one({'name': 'Meercat'}, {'$set': {"weight": 2}})
<mongita.results.UpdateResult object at 00000000000>
>>> mongoose_collection.find({'weight': {'$gt': 1}})
<mongita.cursor.Cursor object at 00000000000>
>>> list(mongoose_collection.find({'weight': {'$gt': 1}}))
[{'_id': 'a1b2c3d4e5f6', 'name': 'Meercat', 'does_not_eat': 'Snakes', 'weight': 2}]
>>> mongoose_collection.delete_one({'name': 'Meercat'})
<mongita.results.DeleteResult object at 00000000000>


Inserts and access Finds Updates and deletes Cold start


Refer to the PyMongo docs for detailed syntax and behavior. Most named keyword parameters are not implemented. When something is not implemented, efforts are made to be loud and obvious about it.

mongita.MongitaClientMemory / mongita.MongitaClientDisk (PyMongo docs)


Note: By default, MongitaClientDisk stores its data in ~/.mongita. To use a different directory, pass host when initializing client = MongitaClientDisk(host=<db_path>).

Database (PyMongo docs)


Collection (PyMongo docs)

mongita.Collection.insert_many(documents, ordered=True)
mongita.Collection.find_one(filter, sort)
mongita.Collection.find(filter, sort, limit)
mongita.Collection.replace_one(filter, replacement, upsert=False)
mongita.Collection.update_one(filter, update)
mongita.Collection.update_many(filter, update)
mongita.Collection.distinct(key, filter)

Cursor (PyMongo docs)

mongita.Cursor.sort(key_or_list, direction=None)

CommandCursor (PyMongo docs)

errors (PyMongo docs)

mongita.errors.MongitaError (parent class of all errors)
mongita.errors.PyMongoError (alias of MongitaError)

results (PyMongo docs)


Currently implemented query operators


Currently implemented update operators



Mongita is an excellent project for open source contributors. There is a lot to do and it is easy to get started. In particular, the following tasks are high in priority:

  • More testing. Try Mongita on a project and report any bugs. The unit tests are extensive but in-the-wild bugs are still possible.
  • Windows support. Mongita fails unit tests on Windows. Fixing these and integration testing the library would be very helpful.
  • SQLite engine. SQLite beats Mongita in most benchmarks. Having it as an engine would provide a fast persistent database which would be a little less than a perfect MongoDB reproduction.
  • Cache eviction. In the disk engine, cache is never evicted after it is read or written. This could cause issues when dealing with large databases.
  • More update operators. Currently, only $set and $inc are implemented.
  • More query operators. Currently, only the "comparison operators" are implemented.
  • find_one_and_... methods.
  • Aggregation pipelines.
  • More cursor methods. Currently only sort, next, skip, and limit are implemented.

You are welcome to email me at [email protected] if you are interested.


BSD 3-clause. Mongita is free and open source for any purpose with basic restrictions related to liability, warranty, and endorsement.


Mongita was started as a component of the fastmap server. Fastmap offloads and parallelizes arbitrary Python functions on the cloud.

Similar projects

  • TinyMongo: Python library. Attempts to replicate the MongoDB interface.
  • MontyDb: Python library. Attempts to replicate the MongoDB interface.
  • UnQLite: Embedded NoSQL with Python bindings. Does not attempt to replicate the MongoDB interface. Very popular.
  • NeDB: Javascript library. Attempts to replicate the MongoDB interface. Very popular.
Open Source Agenda is not affiliated with "Mongita" Project. README Source: scottrogowski/mongita
Open Issues
Last Commit
6 months ago

Open Source Agenda Badge

Open Source Agenda Rating