A Quick OmMongo Introduction


This tutorial is incomplete, but hopefully has enough detail that the places to look in the API documentation for the rest is clear.

In order to use OmMongo to interact with a Mongo database, two things are needed:

  • a Session object to handle queuing,
    sending, and receiving data from the database.
  • a subclass of Document to define a mapping
    between a mongo object and python class. The class serves as the mapping, and instances of the class are the python classes that are saved or loaded into the database

This tutorial is going go through the basics of each of these, as well as querying and updating in a way which is programmatic, type safe, and value checked.

Creating a Mapping Object

>>> from ommongo.document import Document
>>> from ommongo.fields import *
>>> class BloodDonor(Document):
...     first_name = StringField()
...     last_name = StringField()
...     age = IntField(min_value=0)
...     gender = EnumField(StringField(), 'male', 'female')
...     blood_type = EnumField(StringField(), 'O+','A+','B+','AB+','O-','A-','B-','AB-')
...     def __str__(self):
...         return '%s %s (%s; Age: %d; Type: %s)' % (self.first_name, self.last_name,
...             self.gender, self.age, self.blood_type)

The above code creates a class BloodDoner which will has all of the information necessary to create python objects to save, load database objects (though there are none at the moment), and to construct queries for these objects.

It also introduces several field subclasses which are used to define the key/value pairs in the mongo Document. The StringField and IntField are fairly self-explanatory. The ommongo.fields.EnumField is more complex. It takes a field as its first argument followed by any number of values of the type accepted by that field. The EnumField will check both that the value it is constructed with is of the correct type, but also that its value is one of the values given in the constructor.

See also

Document and fields


>>> from ommongo.session import Session
>>> session = Session.connect('ommongo-tutorial')
>>> session.clear_collection(BloodDonor)

The above code creates a session object by connecting to a local mongo server and accessing the ommongo-tutorial database. Any arguments after the database name will be used as arguments to pymongo’s connection function. It’s also possible to directly construct a session using a pymongo database object, allowing one connection to be used for multiple sessions. The last line clears all objects from the collection used for BloodDonor objects, in case this tutorial code had run before.

>>> donor = BloodDonor(first_name='Jeff', last_name='Jenkins',
...             age=28, blood_type='O+', gender='male')
>>> session.save(donor)

When save is called on the donor object the session serializes it into a form that the Mongo database understands, and then inserts it. Once the save command is complete the _id field of donor object is set.

To make the next section more interesting more objects are needed, so we’ll add some of the cast of Community:

>>> session.save(BloodDonor(first_name='Jeff', last_name='Winger', age=38, blood_type='O+', gender='male'))
>>> session.save(BloodDonor(first_name='Britta', last_name='Perry', age=27, blood_type='A+', gender='female'))
>>> session.save(BloodDonor(first_name='Abed', last_name='Nadir', age=29, blood_type='O+', gender='male'))
>>> session.save(BloodDonor(first_name='Shirley', last_name='Bennett', age=39, blood_type='O-', gender='female'))


>>> for donor in session.query(BloodDonor).filter(BloodDonor.first_name == 'Jeff'):
>>>    print donor
Jeff Jenkins (male; Age: 28; Type: O+)
Jeff Winger (male; Age: 38; Type: O+)

The above code uses the query() method of the session object to start a query on the BloodDonor collection. The filter function on a query object allows constraints to be added to the results returned. In this case all donors who have the name Jeff are being printed. The attributes of a Document subclass are used to access the names of fields in such a way that they generate query expressions which filter can use.

Multiple filters can be applied by chaining calls to filter or by adding comma-separated query expressions inside a single call. The following two examples return the same results:

>>> query = session.query(BloodDonor)
>>> for donor in query.filter(BloodDonor.first_name == 'Jeff', BloodDonor.age < 30):
>>>    print donor
Jeff Jenkins (male; Age: 28; Type: O+)
>>> query = session.query(BloodDonor)
>>> for donor in query.filter(BloodDonor.first_name == 'Jeff').filter(BloodDonor.age < 30):
>>>    print donor
Jeff Jenkins (male; Age: 28; Type: O+)

Instead of getting elements by iterating on the query, the one() and first() methods can be used. one returns the first result and raises an exception if there is not exactly one returned result. first returns the first result if there is one, and None if there is not one.

See also

Mongo Query Expression Language for all available types of query expressions


There are a number of methods available for updating results of a query ( rather than updating results by loading an object and re-inserting it). Here’s an example where my age changed and it turned out I had the wrong blood type:

>>> query = session.query(BloodDonor).filter(BloodDonor.first_name == 'Jeff', BloodDonor.last_name == 'Jenkins')
>>> query.inc(BloodDonor.age, 1).set(BloodDonor.blood_type, 'O-').execute()
>>> query.one()
Jeff Jenkins (male; Age: 29; Type: O-)

inc() and set() are two of the methods which can be used to do updates on a query. Once an update method has been called on a query any further chaining is on a UpdateExpression, so no further filtering of results will be possible. The execute() method causes the update to actually happen in the database.

It is also possible to call update() on the session to automatically update a particular object with an update command instead of either saving the object again (which is synchronous) or constructing the update expression yourself as above.

By default all fields use the set operation to update their values, but that can be overriden by passing the on_update operator to the field constructor, or a keyword argument to update. The value of these arguments is currently a string with the mongo operation you want to do instead. For example, IntField(on_update='$inc') is an IntField which will increment when update is called. Session.update is still experimental, so watch out for side-effects

See also



still to come!