1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162
|
Model Managers
==============
InheritanceManager
------------------
This manager (`contributed by Jeff Elmore`_) should be attached to a base model
class in a model-inheritance tree. It allows queries on that base model to
return heterogenous results of the actual proper subtypes, without any
additional queries.
For instance, if you have a ``Place`` model with subclasses ``Restaurant`` and
``Bar``, you may want to query all Places:
.. code-block:: python
nearby_places = Place.objects.filter(location='here')
But when you iterate over ``nearby_places``, you'll get only ``Place``
instances back, even for objects that are "really" ``Restaurant`` or ``Bar``.
If you attach an ``InheritanceManager`` to ``Place``, you can just call the
``select_subclasses()`` method on the ``InheritanceManager`` or any
``QuerySet`` from it, and the resulting objects will be instances of
``Restaurant`` or ``Bar``:
.. code-block:: python
from model_utils.managers import InheritanceManager
class Place(models.Model):
# ...
objects = InheritanceManager()
class Restaurant(Place):
# ...
class Bar(Place):
# ...
nearby_places = Place.objects.filter(location='here').select_subclasses()
for place in nearby_places:
# "place" will automatically be an instance of Place, Restaurant, or Bar
The database query performed will have an extra join for each subclass; if you
want to reduce the number of joins and you only need particular subclasses to
be returned as their actual type, you can pass subclass names to
``select_subclasses()``, much like the built-in ``select_related()`` method:
.. code-block:: python
nearby_places = Place.objects.select_subclasses("restaurant")
# restaurants will be Restaurant instances, bars will still be Place instances
nearby_places = Place.objects.select_subclasses("restaurant", "bar")
# all Places will be converted to Restaurant and Bar instances.
It is also possible to use the subclasses themselves as arguments to
``select_subclasses``, leaving it to calculate the relationship for you:
.. code-block:: python
nearby_places = Place.objects.select_subclasses(Restaurant)
# restaurants will be Restaurant instances, bars will still be Place instances
nearby_places = Place.objects.select_subclasses(Restaurant, Bar)
# all Places will be converted to Restaurant and Bar instances.
It is even possible to mix and match the two:
.. code-block:: python
nearby_places = Place.objects.select_subclasses(Restaurant, "bar")
# all Places will be converted to Restaurant and Bar instances.
``InheritanceManager`` also provides a subclass-fetching alternative to the
``get()`` method:
.. code-block:: python
place = Place.objects.get_subclass(id=some_id)
# "place" will automatically be an instance of Place, Restaurant, or Bar
If you don't explicitly call ``select_subclasses()`` or ``get_subclass()``,
an ``InheritanceManager`` behaves identically to a normal ``Manager``; so
it's safe to use as your default manager for the model.
.. _contributed by Jeff Elmore: http://jeffelmore.org/2010/11/11/automatic-downcasting-of-inherited-models-in-django/
JoinManager
-----------
The ``JoinManager`` will create a temporary table of your current queryset
and join that temporary table with the model of your current queryset. This can
be advantageous if you have to page through your entire DB and using django's
slice mechanism to do that. ``LIMIT .. OFFSET ..`` becomes slower the bigger
offset you use.
.. code-block:: python
sliced_qs = Place.objects.all()[2000:2010]
qs = sliced_qs.join()
# qs contains 10 objects, and there will be a much smaller performance hit
# for paging through all of first 2000 objects.
Alternatively, you can give it a queryset and the manager will create a temporary
table and join that to your current queryset. This can work as a more performant
alternative to using django's ``__in`` as described in the following
(`StackExchange answer`_).
.. code-block:: python
big_qs = Restaurant.objects.filter(menu='vegetarian')
qs = Country.objects.filter(country_code='SE').join(big_qs)
.. _StackExchange answer: https://dba.stackexchange.com/questions/91247/optimizing-a-postgres-query-with-a-large-in
.. _QueryManager:
QueryManager
------------
Many custom model managers do nothing more than return a QuerySet that
is filtered in some way. ``QueryManager`` allows you to express this
pattern with a minimum of boilerplate:
.. code-block:: python
from django.db import models
from model_utils.managers import QueryManager
class Post(models.Model):
...
published = models.BooleanField()
pub_date = models.DateField()
...
objects = models.Manager()
public = QueryManager(published=True).order_by('-pub_date')
The kwargs passed to ``QueryManager`` will be passed as-is to the
``QuerySet.filter()`` method. You can also pass a ``Q`` object to
``QueryManager`` to express more complex conditions. Note that you can
set the ordering of the ``QuerySet`` returned by the ``QueryManager``
by chaining a call to ``.order_by()`` on the ``QueryManager`` (this is
not required).
SoftDeletableManager
--------------------
Returns only model instances that have the ``is_removed`` field set
to False. Uses ``SoftDeletableQuerySet``, which ensures model instances
won't be removed in bulk, but they will be marked as removed instead.
Mixins
------
Each of the above manager classes has a corresponding mixin that can be used to
add functionality to any manager.
Note that any manager class using ``InheritanceManagerMixin`` must return a
``QuerySet`` class using ``InheritanceQuerySetMixin`` from its ``get_queryset``
method.
|