PageRenderTime 48ms CodeModel.GetById 20ms RepoModel.GetById 0ms app.codeStats 0ms

Plain Text | 1886 lines | 1284 code | 602 blank | 0 comment | 0 complexity | e4beb2f635ed95d4fb9106c7e881b6a5 MD5 | raw file
Possible License(s): BSD-3-Clause
  1. ======================
  2. QuerySet API reference
  3. ======================
  4. .. currentmodule:: django.db.models.query
  5. This document describes the details of the ``QuerySet`` API. It builds on the
  6. material presented in the :doc:`model </topics/db/models>` and :doc:`database
  7. query </topics/db/queries>` guides, so you'll probably want to read and
  8. understand those documents before reading this one.
  9. Throughout this reference we'll use the :ref:`example Weblog models
  10. <queryset-model-example>` presented in the :doc:`database query guide
  11. </topics/db/queries>`.
  12. .. _when-querysets-are-evaluated:
  13. When QuerySets are evaluated
  14. ============================
  15. Internally, a ``QuerySet`` can be constructed, filtered, sliced, and generally
  16. passed around without actually hitting the database. No database activity
  17. actually occurs until you do something to evaluate the queryset.
  18. You can evaluate a ``QuerySet`` in the following ways:
  19. * **Iteration.** A ``QuerySet`` is iterable, and it executes its database
  20. query the first time you iterate over it. For example, this will print
  21. the headline of all entries in the database::
  22. for e in Entry.objects.all():
  23. print e.headline
  24. * **Slicing.** As explained in :ref:`limiting-querysets`, a ``QuerySet`` can
  25. be sliced, using Python's array-slicing syntax. Usually slicing a
  26. ``QuerySet`` returns another (unevaluated) ``QuerySet``, but Django will
  27. execute the database query if you use the "step" parameter of slice
  28. syntax.
  29. * **Pickling/Caching.** See the following section for details of what
  30. is involved when `pickling QuerySets`_. The important thing for the
  31. purposes of this section is that the results are read from the database.
  32. * **repr().** A ``QuerySet`` is evaluated when you call ``repr()`` on it.
  33. This is for convenience in the Python interactive interpreter, so you can
  34. immediately see your results when using the API interactively.
  35. * **len().** A ``QuerySet`` is evaluated when you call ``len()`` on it.
  36. This, as you might expect, returns the length of the result list.
  37. Note: *Don't* use ``len()`` on ``QuerySet``\s if all you want to do is
  38. determine the number of records in the set. It's much more efficient to
  39. handle a count at the database level, using SQL's ``SELECT COUNT(*)``,
  40. and Django provides a ``count()`` method for precisely this reason. See
  41. ``count()`` below.
  42. * **list().** Force evaluation of a ``QuerySet`` by calling ``list()`` on
  43. it. For example::
  44. entry_list = list(Entry.objects.all())
  45. Be warned, though, that this could have a large memory overhead, because
  46. Django will load each element of the list into memory. In contrast,
  47. iterating over a ``QuerySet`` will take advantage of your database to
  48. load data and instantiate objects only as you need them.
  49. * **bool().** Testing a ``QuerySet`` in a boolean context, such as using
  50. ``bool()``, ``or``, ``and`` or an ``if`` statement, will cause the query
  51. to be executed. If there is at least one result, the ``QuerySet`` is
  52. ``True``, otherwise ``False``. For example::
  53. if Entry.objects.filter(headline="Test"):
  54. print "There is at least one Entry with the headline Test"
  55. Note: *Don't* use this if all you want to do is determine if at least one
  56. result exists, and don't need the actual objects. It's more efficient to
  57. use ``exists()`` (see below).
  58. .. _pickling QuerySets:
  59. Pickling QuerySets
  60. ------------------
  61. If you pickle_ a ``QuerySet``, this will force all the results to be loaded
  62. into memory prior to pickling. Pickling is usually used as a precursor to
  63. caching and when the cached queryset is reloaded, you want the results to
  64. already be present and ready for use (reading from the database can take some
  65. time, defeating the purpose of caching). This means that when you unpickle a
  66. ``QuerySet``, it contains the results at the moment it was pickled, rather
  67. than the results that are currently in the database.
  68. If you only want to pickle the necessary information to recreate the
  69. ``QuerySet`` from the database at a later time, pickle the ``query`` attribute
  70. of the ``QuerySet``. You can then recreate the original ``QuerySet`` (without
  71. any results loaded) using some code like this::
  72. >>> import pickle
  73. >>> query = pickle.loads(s) # Assuming 's' is the pickled string.
  74. >>> qs = MyModel.objects.all()
  75. >>> qs.query = query # Restore the original 'query'.
  76. The ``query`` attribute is an opaque object. It represents the internals of
  77. the query construction and is not part of the public API. However, it is safe
  78. (and fully supported) to pickle and unpickle the attribute's contents as
  79. described here.
  80. .. admonition:: You can't share pickles between versions
  81. Pickles of QuerySets are only valid for the version of Django that
  82. was used to generate them. If you generate a pickle using Django
  83. version N, there is no guarantee that pickle will be readable with
  84. Django version N+1. Pickles should not be used as part of a long-term
  85. archival strategy.
  86. .. _pickle:
  87. .. _queryset-api:
  88. QuerySet API
  89. ============
  90. Though you usually won't create one manually -- you'll go through a
  91. :class:`~django.db.models.Manager` -- here's the formal declaration of a
  92. ``QuerySet``:
  93. .. class:: QuerySet([model=None, query=None, using=None])
  94. Usually when you'll interact with a ``QuerySet`` you'll use it by
  95. :ref:`chaining filters <chaining-filters>`. To make this work, most
  96. ``QuerySet`` methods return new querysets. These methods are covered in
  97. detail later in this section.
  98. The ``QuerySet`` class has two public attributes you can use for
  99. introspection:
  100. .. attribute:: ordered
  101. ``True`` if the ``QuerySet`` is ordered -- i.e. has an order_by()
  102. clause or a default ordering on the model. ``False`` otherwise.
  103. .. attribute:: db
  104. The database that will be used if this query is executed now.
  105. .. note::
  106. The ``query`` parameter to :class:`QuerySet` exists so that specialized
  107. query subclasses such as
  108. :class:`~django.contrib.gis.db.models.GeoQuerySet` can reconstruct
  109. internal query state. The value of the parameter is an opaque
  110. representation of that query state and is not part of a public API.
  111. To put it simply: if you need to ask, you don't need to use it.
  112. .. currentmodule:: django.db.models.query.QuerySet
  113. Methods that return new QuerySets
  114. ---------------------------------
  115. Django provides a range of ``QuerySet`` refinement methods that modify either
  116. the types of results returned by the ``QuerySet`` or the way its SQL query is
  117. executed.
  118. filter
  119. ~~~~~~
  120. .. method:: filter(**kwargs)
  121. Returns a new ``QuerySet`` containing objects that match the given lookup
  122. parameters.
  123. The lookup parameters (``**kwargs``) should be in the format described in
  124. `Field lookups`_ below. Multiple parameters are joined via ``AND`` in the
  125. underlying SQL statement.
  126. exclude
  127. ~~~~~~~
  128. .. method:: exclude(**kwargs)
  129. Returns a new ``QuerySet`` containing objects that do *not* match the given
  130. lookup parameters.
  131. The lookup parameters (``**kwargs``) should be in the format described in
  132. `Field lookups`_ below. Multiple parameters are joined via ``AND`` in the
  133. underlying SQL statement, and the whole thing is enclosed in a ``NOT()``.
  134. This example excludes all entries whose ``pub_date`` is later than 2005-1-3
  135. AND whose ``headline`` is "Hello"::
  136. Entry.objects.exclude(, 1, 3), headline='Hello')
  137. In SQL terms, that evaluates to::
  138. SELECT ...
  139. WHERE NOT (pub_date > '2005-1-3' AND headline = 'Hello')
  140. This example excludes all entries whose ``pub_date`` is later than 2005-1-3
  141. OR whose headline is "Hello"::
  142. Entry.objects.exclude(, 1, 3)).exclude(headline='Hello')
  143. In SQL terms, that evaluates to::
  144. SELECT ...
  145. WHERE NOT pub_date > '2005-1-3'
  146. AND NOT headline = 'Hello'
  147. Note the second example is more restrictive.
  148. annotate
  149. ~~~~~~~~
  150. .. method:: annotate(*args, **kwargs)
  151. Annotates each object in the ``QuerySet`` with the provided list of
  152. aggregate values (averages, sums, etc) that have been computed over
  153. the objects that are related to the objects in the ``QuerySet``.
  154. Each argument to ``annotate()`` is an annotation that will be added
  155. to each object in the ``QuerySet`` that is returned.
  156. The aggregation functions that are provided by Django are described
  157. in `Aggregation Functions`_ below.
  158. Annotations specified using keyword arguments will use the keyword as
  159. the alias for the annotation. Anonymous arguments will have an alias
  160. generated for them based upon the name of the aggregate function and
  161. the model field that is being aggregated.
  162. For example, if you were manipulating a list of blogs, you may want
  163. to determine how many entries have been made in each blog::
  164. >>> q = Blog.objects.annotate(Count('entry'))
  165. # The name of the first blog
  166. >>> q[0].name
  167. 'Blogasaurus'
  168. # The number of entries on the first blog
  169. >>> q[0].entry__count
  170. 42
  171. The ``Blog`` model doesn't define an ``entry__count`` attribute by itself,
  172. but by using a keyword argument to specify the aggregate function, you can
  173. control the name of the annotation::
  174. >>> q = Blog.objects.annotate(number_of_entries=Count('entry'))
  175. # The number of entries on the first blog, using the name provided
  176. >>> q[0].number_of_entries
  177. 42
  178. For an in-depth discussion of aggregation, see :doc:`the topic guide on
  179. Aggregation </topics/db/aggregation>`.
  180. order_by
  181. ~~~~~~~~
  182. .. method:: order_by(*fields)
  183. By default, results returned by a ``QuerySet`` are ordered by the ordering
  184. tuple given by the ``ordering`` option in the model's ``Meta``. You can
  185. override this on a per-``QuerySet`` basis by using the ``order_by`` method.
  186. Example::
  187. Entry.objects.filter(pub_date__year=2005).order_by('-pub_date', 'headline')
  188. The result above will be ordered by ``pub_date`` descending, then by
  189. ``headline`` ascending. The negative sign in front of ``"-pub_date"`` indicates
  190. *descending* order. Ascending order is implied. To order randomly, use ``"?"``,
  191. like so::
  192. Entry.objects.order_by('?')
  193. Note: ``order_by('?')`` queries may be expensive and slow, depending on the
  194. database backend you're using.
  195. To order by a field in a different model, use the same syntax as when you are
  196. querying across model relations. That is, the name of the field, followed by a
  197. double underscore (``__``), followed by the name of the field in the new model,
  198. and so on for as many models as you want to join. For example::
  199. Entry.objects.order_by('blog__name', 'headline')
  200. If you try to order by a field that is a relation to another model, Django will
  201. use the default ordering on the related model (or order by the related model's
  202. primary key if there is no ``Meta.ordering`` specified. For example::
  203. Entry.objects.order_by('blog')
  204. identical to::
  205. Entry.objects.order_by('blog__id')
  206. ...since the ``Blog`` model has no default ordering specified.
  207. Be cautious when ordering by fields in related models if you are also using
  208. ``distinct()``. See the note in :meth:`distinct` for an explanation of how
  209. related model ordering can change the expected results.
  210. It is permissible to specify a multi-valued field to order the results by (for
  211. example, a ``ManyToMany`` field). Normally this won't be a sensible thing to
  212. do and it's really an advanced usage feature. However, if you know that your
  213. queryset's filtering or available data implies that there will only be one
  214. ordering piece of data for each of the main items you are selecting, the
  215. ordering may well be exactly what you want to do. Use ordering on multi-valued
  216. fields with care and make sure the results are what you expect.
  217. There's no way to specify whether ordering should be case sensitive. With
  218. respect to case-sensitivity, Django will order results however your database
  219. backend normally orders them.
  220. If you don't want any ordering to be applied to a query, not even the default
  221. ordering, call ``order_by()`` with no parameters.
  222. You can tell if a query is ordered or not by checking the
  223. :attr:`.QuerySet.ordered` attribute, which will be ``True`` if the
  224. ``QuerySet`` has been ordered in any way.
  225. reverse
  226. ~~~~~~~
  227. .. method:: reverse()
  228. Use the ``reverse()`` method to reverse the order in which a queryset's
  229. elements are returned. Calling ``reverse()`` a second time restores the
  230. ordering back to the normal direction.
  231. To retrieve the ''last'' five items in a queryset, you could do this::
  232. my_queryset.reverse()[:5]
  233. Note that this is not quite the same as slicing from the end of a sequence in
  234. Python. The above example will return the last item first, then the
  235. penultimate item and so on. If we had a Python sequence and looked at
  236. ``seq[-5:]``, we would see the fifth-last item first. Django doesn't support
  237. that mode of access (slicing from the end), because it's not possible to do it
  238. efficiently in SQL.
  239. Also, note that ``reverse()`` should generally only be called on a
  240. ``QuerySet`` which has a defined ordering (e.g., when querying against
  241. a model which defines a default ordering, or when using
  242. ``order_by()``). If no such ordering is defined for a given
  243. ``QuerySet``, calling ``reverse()`` on it has no real effect (the
  244. ordering was undefined prior to calling ``reverse()``, and will remain
  245. undefined afterward).
  246. distinct
  247. ~~~~~~~~
  248. .. method:: distinct()
  249. Returns a new ``QuerySet`` that uses ``SELECT DISTINCT`` in its SQL query. This
  250. eliminates duplicate rows from the query results.
  251. By default, a ``QuerySet`` will not eliminate duplicate rows. In practice, this
  252. is rarely a problem, because simple queries such as ``Blog.objects.all()``
  253. don't introduce the possibility of duplicate result rows. However, if your
  254. query spans multiple tables, it's possible to get duplicate results when a
  255. ``QuerySet`` is evaluated. That's when you'd use ``distinct()``.
  256. .. note::
  257. Any fields used in an :meth:`order_by` call are included in the SQL
  258. ``SELECT`` columns. This can sometimes lead to unexpected results when
  259. used in conjunction with ``distinct()``. If you order by fields from a
  260. related model, those fields will be added to the selected columns and they
  261. may make otherwise duplicate rows appear to be distinct. Since the extra
  262. columns don't appear in the returned results (they are only there to
  263. support ordering), it sometimes looks like non-distinct results are being
  264. returned.
  265. Similarly, if you use a ``values()`` query to restrict the columns
  266. selected, the columns used in any ``order_by()`` (or default model
  267. ordering) will still be involved and may affect uniqueness of the results.
  268. The moral here is that if you are using ``distinct()`` be careful about
  269. ordering by related models. Similarly, when using ``distinct()`` and
  270. ``values()`` together, be careful when ordering by fields not in the
  271. ``values()`` call.
  272. values
  273. ~~~~~~
  274. .. method:: values(*fields)
  275. Returns a ``ValuesQuerySet`` -- a ``QuerySet`` that returns dictionaries when
  276. used as an iterable, rather than model-instance objects.
  277. Each of those dictionaries represents an object, with the keys corresponding to
  278. the attribute names of model objects.
  279. This example compares the dictionaries of ``values()`` with the normal model
  280. objects::
  281. # This list contains a Blog object.
  282. >>> Blog.objects.filter(name__startswith='Beatles')
  283. [<Blog: Beatles Blog>]
  284. # This list contains a dictionary.
  285. >>> Blog.objects.filter(name__startswith='Beatles').values()
  286. [{'id': 1, 'name': 'Beatles Blog', 'tagline': 'All the latest Beatles news.'}]
  287. ``values()`` takes optional positional arguments, ``*fields``, which specify
  288. field names to which the ``SELECT`` should be limited. If you specify the
  289. fields, each dictionary will contain only the field keys/values for the fields
  290. you specify. If you don't specify the fields, each dictionary will contain a
  291. key and value for every field in the database table.
  292. Example::
  293. >>> Blog.objects.values()
  294. [{'id': 1, 'name': 'Beatles Blog', 'tagline': 'All the latest Beatles news.'}],
  295. >>> Blog.objects.values('id', 'name')
  296. [{'id': 1, 'name': 'Beatles Blog'}]
  297. A few subtleties that are worth mentioning:
  298. * If you have a field called ``foo`` that is a
  299. :class:`~django.db.models.ForeignKey`, the default ``values()`` call
  300. will return a dictionary key called ``foo_id``, since this is the name
  301. of the hidden model attribute that stores the actual value (the ``foo``
  302. attribute refers to the related model). When you are calling
  303. ``values()`` and passing in field names, you can pass in either ``foo``
  304. or ``foo_id`` and you will get back the same thing (the dictionary key
  305. will match the field name you passed in).
  306. For example::
  307. >>> Entry.objects.values()
  308. [{'blog_id': 1, 'headline': u'First Entry', ...}, ...]
  309. >>> Entry.objects.values('blog')
  310. [{'blog': 1}, ...]
  311. >>> Entry.objects.values('blog_id')
  312. [{'blog_id': 1}, ...]
  313. * When using ``values()`` together with ``distinct()``, be aware that
  314. ordering can affect the results. See the note in :meth:`distinct` for
  315. details.
  316. * If you use a ``values()`` clause after an ``extra()`` clause,
  317. any fields defined by a ``select`` argument in the ``extra()``
  318. must be explicitly included in the ``values()`` clause. However,
  319. if the ``extra()`` clause is used after the ``values()``, the
  320. fields added by the select will be included automatically.
  321. A ``ValuesQuerySet`` is useful when you know you're only going to need values
  322. from a small number of the available fields and you won't need the
  323. functionality of a model instance object. It's more efficient to select only
  324. the fields you need to use.
  325. Finally, note a ``ValuesQuerySet`` is a subclass of ``QuerySet``, so it has all
  326. methods of ``QuerySet``. You can call ``filter()`` on it, or ``order_by()``, or
  327. whatever. Yes, that means these two calls are identical::
  328. Blog.objects.values().order_by('id')
  329. Blog.objects.order_by('id').values()
  330. The people who made Django prefer to put all the SQL-affecting methods first,
  331. followed (optionally) by any output-affecting methods (such as ``values()``),
  332. but it doesn't really matter. This is your chance to really flaunt your
  333. individualism.
  334. .. versionchanged:: 1.3
  335. The ``values()`` method previously did not return anything for
  336. :class:`~django.db.models.ManyToManyField` attributes and would raise an error
  337. if you tried to pass this type of field to it.
  338. This restriction has been lifted, and you can now also refer to fields on
  339. related models with reverse relations through ``OneToOneField``, ``ForeignKey``
  340. and ``ManyToManyField`` attributes::
  341. Blog.objects.values('name', 'entry__headline')
  342. [{'name': 'My blog', 'entry__headline': 'An entry'},
  343. {'name': 'My blog', 'entry__headline': 'Another entry'}, ...]
  344. .. warning::
  345. Because :class:`~django.db.models.ManyToManyField` attributes and reverse
  346. relations can have multiple related rows, including these can have a
  347. multiplier effect on the size of your result set. This will be especially
  348. pronounced if you include multiple such fields in your ``values()`` query,
  349. in which case all possible combinations will be returned.
  350. values_list
  351. ~~~~~~~~~~~
  352. .. method:: values_list(*fields)
  353. This is similar to ``values()`` except that instead of returning dictionaries,
  354. it returns tuples when iterated over. Each tuple contains the value from the
  355. respective field passed into the ``values_list()`` call -- so the first item is
  356. the first field, etc. For example::
  357. >>> Entry.objects.values_list('id', 'headline')
  358. [(1, u'First entry'), ...]
  359. If you only pass in a single field, you can also pass in the ``flat``
  360. parameter. If ``True``, this will mean the returned results are single values,
  361. rather than one-tuples. An example should make the difference clearer::
  362. >>> Entry.objects.values_list('id').order_by('id')
  363. [(1,), (2,), (3,), ...]
  364. >>> Entry.objects.values_list('id', flat=True).order_by('id')
  365. [1, 2, 3, ...]
  366. It is an error to pass in ``flat`` when there is more than one field.
  367. If you don't pass any values to ``values_list()``, it will return all the
  368. fields in the model, in the order they were declared.
  369. dates
  370. ~~~~~
  371. .. method:: dates(field, kind, order='ASC')
  372. Returns a ``DateQuerySet`` -- a ``QuerySet`` that evaluates to a list of
  373. ``datetime.datetime`` objects representing all available dates of a particular
  374. kind within the contents of the ``QuerySet``.
  375. ``field`` should be the name of a ``DateField`` or ``DateTimeField`` of your
  376. model.
  377. ``kind`` should be either ``"year"``, ``"month"`` or ``"day"``. Each
  378. ``datetime.datetime`` object in the result list is "truncated" to the given
  379. ``type``.
  380. * ``"year"`` returns a list of all distinct year values for the field.
  381. * ``"month"`` returns a list of all distinct year/month values for the field.
  382. * ``"day"`` returns a list of all distinct year/month/day values for the field.
  383. ``order``, which defaults to ``'ASC'``, should be either ``'ASC'`` or
  384. ``'DESC'``. This specifies how to order the results.
  385. Examples::
  386. >>> Entry.objects.dates('pub_date', 'year')
  387. [datetime.datetime(2005, 1, 1)]
  388. >>> Entry.objects.dates('pub_date', 'month')
  389. [datetime.datetime(2005, 2, 1), datetime.datetime(2005, 3, 1)]
  390. >>> Entry.objects.dates('pub_date', 'day')
  391. [datetime.datetime(2005, 2, 20), datetime.datetime(2005, 3, 20)]
  392. >>> Entry.objects.dates('pub_date', 'day', order='DESC')
  393. [datetime.datetime(2005, 3, 20), datetime.datetime(2005, 2, 20)]
  394. >>> Entry.objects.filter(headline__contains='Lennon').dates('pub_date', 'day')
  395. [datetime.datetime(2005, 3, 20)]
  396. none
  397. ~~~~
  398. .. method:: none()
  399. Returns an ``EmptyQuerySet`` -- a ``QuerySet`` that always evaluates to
  400. an empty list. This can be used in cases where you know that you should
  401. return an empty result set and your caller is expecting a ``QuerySet``
  402. object (instead of returning an empty list, for example.)
  403. Examples::
  404. >>> Entry.objects.none()
  405. []
  406. all
  407. ~~~
  408. .. method:: all()
  409. Returns a *copy* of the current ``QuerySet`` (or ``QuerySet`` subclass you
  410. pass in). This can be useful in some situations where you might want to pass
  411. in either a model manager or a ``QuerySet`` and do further filtering on the
  412. result. You can safely call ``all()`` on either object and then you'll
  413. definitely have a ``QuerySet`` to work with.
  414. .. _select-related:
  415. select_related
  416. ~~~~~~~~~~~~~~
  417. .. method:: select_related()
  418. Returns a ``QuerySet`` that will automatically "follow" foreign-key
  419. relationships, selecting that additional related-object data when it executes
  420. its query. This is a performance booster which results in (sometimes much)
  421. larger queries but means later use of foreign-key relationships won't require
  422. database queries.
  423. The following examples illustrate the difference between plain lookups and
  424. ``select_related()`` lookups. Here's standard lookup::
  425. # Hits the database.
  426. e = Entry.objects.get(id=5)
  427. # Hits the database again to get the related Blog object.
  428. b =
  429. And here's ``select_related`` lookup::
  430. # Hits the database.
  431. e = Entry.objects.select_related().get(id=5)
  432. # Doesn't hit the database, because has been prepopulated
  433. # in the previous query.
  434. b =
  435. ``select_related()`` follows foreign keys as far as possible. If you have the
  436. following models::
  437. class City(models.Model):
  438. # ...
  439. class Person(models.Model):
  440. # ...
  441. hometown = models.ForeignKey(City)
  442. class Book(models.Model):
  443. # ...
  444. author = models.ForeignKey(Person)
  445. ...then a call to ``Book.objects.select_related().get(id=4)`` will cache the
  446. related ``Person`` *and* the related ``City``::
  447. b = Book.objects.select_related().get(id=4)
  448. p = # Doesn't hit the database.
  449. c = p.hometown # Doesn't hit the database.
  450. b = Book.objects.get(id=4) # No select_related() in this example.
  451. p = # Hits the database.
  452. c = p.hometown # Hits the database.
  453. Note that, by default, ``select_related()`` does not follow foreign keys that
  454. have ``null=True``.
  455. Usually, using ``select_related()`` can vastly improve performance because your
  456. app can avoid many database calls. However, in situations with deeply nested
  457. sets of relationships ``select_related()`` can sometimes end up following "too
  458. many" relations, and can generate queries so large that they end up being slow.
  459. In these situations, you can use the ``depth`` argument to ``select_related()``
  460. to control how many "levels" of relations ``select_related()`` will actually
  461. follow::
  462. b = Book.objects.select_related(depth=1).get(id=4)
  463. p = # Doesn't hit the database.
  464. c = p.hometown # Requires a database call.
  465. Sometimes you only want to access specific models that are related to your root
  466. model, not all of the related models. In these cases, you can pass the related
  467. field names to ``select_related()`` and it will only follow those relations.
  468. You can even do this for models that are more than one relation away by
  469. separating the field names with double underscores, just as for filters. For
  470. example, if you have this model::
  471. class Room(models.Model):
  472. # ...
  473. building = models.ForeignKey(...)
  474. class Group(models.Model):
  475. # ...
  476. teacher = models.ForeignKey(...)
  477. room = models.ForeignKey(Room)
  478. subject = models.ForeignKey(...)
  479. ...and you only needed to work with the ``room`` and ``subject`` attributes,
  480. you could write this::
  481. g = Group.objects.select_related('room', 'subject')
  482. This is also valid::
  483. g = Group.objects.select_related('room__building', 'subject')
  484. ...and would also pull in the ``building`` relation.
  485. You can refer to any ``ForeignKey`` or ``OneToOneField`` relation in
  486. the list of fields passed to ``select_related``. This includes foreign
  487. keys that have ``null=True`` (unlike the default ``select_related()``
  488. call). It's an error to use both a list of fields and the ``depth``
  489. parameter in the same ``select_related()`` call, since they are
  490. conflicting options.
  491. .. versionchanged:: 1.2
  492. You can also refer to the reverse direction of a ``OneToOneFields`` in
  493. the list of fields passed to ``select_related`` -- that is, you can traverse
  494. a ``OneToOneField`` back to the object on which the field is defined. Instead
  495. of specifying the field name, use the ``related_name`` for the field on the
  496. related object.
  497. ``OneToOneFields`` will not be traversed in the reverse direction if you
  498. are performing a depth-based ``select_related``.
  499. extra
  500. ~~~~~
  501. .. method:: extra(select=None, where=None, params=None, tables=None, order_by=None, select_params=None)
  502. Sometimes, the Django query syntax by itself can't easily express a complex
  503. ``WHERE`` clause. For these edge cases, Django provides the ``extra()``
  504. ``QuerySet`` modifier -- a hook for injecting specific clauses into the SQL
  505. generated by a ``QuerySet``.
  506. By definition, these extra lookups may not be portable to different database
  507. engines (because you're explicitly writing SQL code) and violate the DRY
  508. principle, so you should avoid them if possible.
  509. Specify one or more of ``params``, ``select``, ``where`` or ``tables``. None
  510. of the arguments is required, but you should use at least one of them.
  511. * ``select``
  512. The ``select`` argument lets you put extra fields in the ``SELECT`` clause.
  513. It should be a dictionary mapping attribute names to SQL clauses to use to
  514. calculate that attribute.
  515. Example::
  516. Entry.objects.extra(select={'is_recent': "pub_date > '2006-01-01'"})
  517. As a result, each ``Entry`` object will have an extra attribute,
  518. ``is_recent``, a boolean representing whether the entry's ``pub_date`` is
  519. greater than Jan. 1, 2006.
  520. Django inserts the given SQL snippet directly into the ``SELECT``
  521. statement, so the resulting SQL of the above example would be something
  522. like::
  523. SELECT blog_entry.*, (pub_date > '2006-01-01') AS is_recent
  524. FROM blog_entry;
  525. The next example is more advanced; it does a subquery to give each
  526. resulting ``Blog`` object an ``entry_count`` attribute, an integer count
  527. of associated ``Entry`` objects::
  528. Blog.objects.extra(
  529. select={
  530. 'entry_count': 'SELECT COUNT(*) FROM blog_entry WHERE blog_entry.blog_id ='
  531. },
  532. )
  533. (In this particular case, we're exploiting the fact that the query will
  534. already contain the ``blog_blog`` table in its ``FROM`` clause.)
  535. The resulting SQL of the above example would be::
  536. SELECT blog_blog.*, (SELECT COUNT(*) FROM blog_entry WHERE blog_entry.blog_id = AS entry_count
  537. FROM blog_blog;
  538. Note that the parenthesis required by most database engines around
  539. subqueries are not required in Django's ``select`` clauses. Also note that
  540. some database backends, such as some MySQL versions, don't support
  541. subqueries.
  542. In some rare cases, you might wish to pass parameters to the SQL fragments
  543. in ``extra(select=...)``. For this purpose, use the ``select_params``
  544. parameter. Since ``select_params`` is a sequence and the ``select``
  545. attribute is a dictionary, some care is required so that the parameters
  546. are matched up correctly with the extra select pieces. In this situation,
  547. you should use a ``django.utils.datastructures.SortedDict`` for the
  548. ``select`` value, not just a normal Python dictionary.
  549. This will work, for example::
  550. Blog.objects.extra(
  551. select=SortedDict([('a', '%s'), ('b', '%s')]),
  552. select_params=('one', 'two'))
  553. The only thing to be careful about when using select parameters in
  554. ``extra()`` is to avoid using the substring ``"%%s"`` (that's *two*
  555. percent characters before the ``s``) in the select strings. Django's
  556. tracking of parameters looks for ``%s`` and an escaped ``%`` character
  557. like this isn't detected. That will lead to incorrect results.
  558. * ``where`` / ``tables``
  559. You can define explicit SQL ``WHERE`` clauses -- perhaps to perform
  560. non-explicit joins -- by using ``where``. You can manually add tables to
  561. the SQL ``FROM`` clause by using ``tables``.
  562. ``where`` and ``tables`` both take a list of strings. All ``where``
  563. parameters are "AND"ed to any other search criteria.
  564. Example::
  565. Entry.objects.extra(where=['id IN (3, 4, 5, 20)'])
  566. ...translates (roughly) into the following SQL::
  567. SELECT * FROM blog_entry WHERE id IN (3, 4, 5, 20);
  568. Be careful when using the ``tables`` parameter if you're specifying
  569. tables that are already used in the query. When you add extra tables
  570. via the ``tables`` parameter, Django assumes you want that table included
  571. an extra time, if it is already included. That creates a problem,
  572. since the table name will then be given an alias. If a table appears
  573. multiple times in an SQL statement, the second and subsequent occurrences
  574. must use aliases so the database can tell them apart. If you're
  575. referring to the extra table you added in the extra ``where`` parameter
  576. this is going to cause errors.
  577. Normally you'll only be adding extra tables that don't already appear in
  578. the query. However, if the case outlined above does occur, there are a few
  579. solutions. First, see if you can get by without including the extra table
  580. and use the one already in the query. If that isn't possible, put your
  581. ``extra()`` call at the front of the queryset construction so that your
  582. table is the first use of that table. Finally, if all else fails, look at
  583. the query produced and rewrite your ``where`` addition to use the alias
  584. given to your extra table. The alias will be the same each time you
  585. construct the queryset in the same way, so you can rely upon the alias
  586. name to not change.
  587. * ``order_by``
  588. If you need to order the resulting queryset using some of the new fields
  589. or tables you have included via ``extra()`` use the ``order_by`` parameter
  590. to ``extra()`` and pass in a sequence of strings. These strings should
  591. either be model fields (as in the normal ``order_by()`` method on
  592. querysets), of the form ``table_name.column_name`` or an alias for a column
  593. that you specified in the ``select`` parameter to ``extra()``.
  594. For example::
  595. q = Entry.objects.extra(select={'is_recent': "pub_date > '2006-01-01'"})
  596. q = q.extra(order_by = ['-is_recent'])
  597. This would sort all the items for which ``is_recent`` is true to the front
  598. of the result set (``True`` sorts before ``False`` in a descending
  599. ordering).
  600. This shows, by the way, that you can make multiple calls to
  601. ``extra()`` and it will behave as you expect (adding new constraints each
  602. time).
  603. * ``params``
  604. The ``where`` parameter described above may use standard Python database
  605. string placeholders -- ``'%s'`` to indicate parameters the database engine
  606. should automatically quote. The ``params`` argument is a list of any extra
  607. parameters to be substituted.
  608. Example::
  609. Entry.objects.extra(where=['headline=%s'], params=['Lennon'])
  610. Always use ``params`` instead of embedding values directly into ``where``
  611. because ``params`` will ensure values are quoted correctly according to
  612. your particular backend. (For example, quotes will be escaped correctly.)
  613. Bad::
  614. Entry.objects.extra(where=["headline='Lennon'"])
  615. Good::
  616. Entry.objects.extra(where=['headline=%s'], params=['Lennon'])
  617. defer
  618. ~~~~~
  619. .. method:: defer(*fields)
  620. In some complex data-modeling situations, your models might contain a lot of
  621. fields, some of which could contain a lot of data (for example, text fields),
  622. or require expensive processing to convert them to Python objects. If you are
  623. using the results of a queryset in some situation where you know you don't
  624. need those particular fields, you can tell Django not to retrieve them from
  625. the database.
  626. This is done by passing the names of the fields to not load to ``defer()``::
  627. Entry.objects.defer("headline", "body")
  628. A queryset that has deferred fields will still return model instances. Each
  629. deferred field will be retrieved from the database if you access that field
  630. (one at a time, not all the deferred fields at once).
  631. You can make multiple calls to ``defer()``. Each call adds new fields to the
  632. deferred set::
  633. # Defers both the body and headline fields.
  634. Entry.objects.defer("body").filter(rating=5).defer("headline")
  635. The order in which fields are added to the deferred set does not matter.
  636. Calling ``defer()`` with a field name that has already been deferred is
  637. harmless (the field will still be deferred).
  638. You can defer loading of fields in related models (if the related models are
  639. loading via ``select_related()``) by using the standard double-underscore
  640. notation to separate related fields::
  641. Blog.objects.select_related().defer("entry__headline", "entry__body")
  642. If you want to clear the set of deferred fields, pass ``None`` as a parameter
  643. to ``defer()``::
  644. # Load all fields immediately.
  645. my_queryset.defer(None)
  646. Some fields in a model won't be deferred, even if you ask for them. You can
  647. never defer the loading of the primary key. If you are using
  648. ``select_related()`` to retrieve other models at the same time you shouldn't
  649. defer the loading of the field that connects from the primary model to the
  650. related one (at the moment, that doesn't raise an error, but it will
  651. eventually).
  652. .. note::
  653. The ``defer()`` method (and its cousin, ``only()``, below) are only for
  654. advanced use-cases. They provide an optimization for when you have
  655. analyzed your queries closely and understand *exactly* what information
  656. you need and have measured that the difference between returning the
  657. fields you need and the full set of fields for the model will be
  658. significant. When you are initially developing your applications, don't
  659. bother using ``defer()``; leave it until your query construction has
  660. settled down and you understand where the hot-points are.
  661. only
  662. ~~~~
  663. .. method:: only(*fields)
  664. The ``only()`` method is more or less the opposite of ``defer()``. You
  665. call it with the fields that should *not* be deferred when retrieving a model.
  666. If you have a model where almost all the fields need to be deferred, using
  667. ``only()`` to specify the complementary set of fields could result in simpler
  668. code.
  669. If you have a model with fields ``name``, ``age`` and ``biography``, the
  670. following two querysets are the same, in terms of deferred fields::
  671. Person.objects.defer("age", "biography")
  672. Person.objects.only("name")
  673. Whenever you call ``only()`` it *replaces* the set of fields to load
  674. immediately. The method's name is mnemonic: **only** those fields are loaded
  675. immediately; the remainder are deferred. Thus, successive calls to ``only()``
  676. result in only the final fields being considered::
  677. # This will defer all fields except the headline.
  678. Entry.objects.only("body", "rating").only("headline")
  679. Since ``defer()`` acts incrementally (adding fields to the deferred list), you
  680. can combine calls to ``only()`` and ``defer()`` and things will behave
  681. logically::
  682. # Final result is that everything except "headline" is deferred.
  683. Entry.objects.only("headline", "body").defer("body")
  684. # Final result loads headline and body immediately (only() replaces any
  685. # existing set of fields).
  686. Entry.objects.defer("body").only("headline", "body")
  687. using
  688. ~~~~~
  689. .. method:: using(alias)
  690. .. versionadded:: 1.2
  691. This method is for controlling which database the ``QuerySet`` will be
  692. evaluated against if you are using more than one database. The only argument
  693. this method takes is the alias of a database, as defined in
  694. :setting:`DATABASES`.
  695. For example::
  696. # queries the database with the 'default' alias.
  697. >>> Entry.objects.all()
  698. # queries the database with the 'backup' alias
  699. >>> Entry.objects.using('backup')
  700. Methods that do not return QuerySets
  701. ------------------------------------
  702. The following ``QuerySet`` methods evaluate the ``QuerySet`` and return
  703. something *other than* a ``QuerySet``.
  704. These methods do not use a cache (see :ref:`caching-and-querysets`). Rather,
  705. they query the database each time they're called.
  706. get
  707. ~~~
  708. .. method:: get(**kwargs)
  709. Returns the object matching the given lookup parameters, which should be in
  710. the format described in `Field lookups`_.
  711. ``get()`` raises ``MultipleObjectsReturned`` if more than one object was
  712. found. The ``MultipleObjectsReturned`` exception is an attribute of the model
  713. class.
  714. ``get()`` raises a ``DoesNotExist`` exception if an object wasn't found for
  715. the given parameters. This exception is also an attribute of the model class.
  716. Example::
  717. Entry.objects.get(id='foo') # raises Entry.DoesNotExist
  718. The ``DoesNotExist`` exception inherits from
  719. ``django.core.exceptions.ObjectDoesNotExist``, so you can target multiple
  720. ``DoesNotExist`` exceptions. Example::
  721. from django.core.exceptions import ObjectDoesNotExist
  722. try:
  723. e = Entry.objects.get(id=3)
  724. b = Blog.objects.get(id=1)
  725. except ObjectDoesNotExist:
  726. print "Either the entry or blog doesn't exist."
  727. create
  728. ~~~~~~
  729. .. method:: create(**kwargs)
  730. A convenience method for creating an object and saving it all in one step. Thus::
  731. p = Person.objects.create(first_name="Bruce", last_name="Springsteen")
  732. and::
  733. p = Person(first_name="Bruce", last_name="Springsteen")
  735. are equivalent.
  736. The :ref:`force_insert <ref-models-force-insert>` parameter is documented
  737. elsewhere, but all it means is that a new object will always be created.
  738. Normally you won't need to worry about this. However, if your model contains a
  739. manual primary key value that you set and if that value already exists in the
  740. database, a call to ``create()`` will fail with an
  741. :exc:`~django.db.IntegrityError` since primary keys must be unique. So remember
  742. to be prepared to handle the exception if you are using manual primary keys.
  743. get_or_create
  744. ~~~~~~~~~~~~~
  745. .. method:: get_or_create(**kwargs)
  746. A convenience method for looking up an object with the given kwargs, creating
  747. one if necessary.
  748. Returns a tuple of ``(object, created)``, where ``object`` is the retrieved or
  749. created object and ``created`` is a boolean specifying whether a new object was
  750. created.
  751. This is meant as a shortcut to boilerplatish code and is mostly useful for
  752. data-import scripts. For example::
  753. try:
  754. obj = Person.objects.get(first_name='John', last_name='Lennon')
  755. except Person.DoesNotExist:
  756. obj = Person(first_name='John', last_name='Lennon', birthday=date(1940, 10, 9))
  758. This pattern gets quite unwieldy as the number of fields in a model goes up.
  759. The above example can be rewritten using ``get_or_create()`` like so::
  760. obj, created = Person.objects.get_or_create(first_name='John', last_name='Lennon',
  761. defaults={'birthday': date(1940, 10, 9)})
  762. Any keyword arguments passed to ``get_or_create()`` -- *except* an optional one
  763. called ``defaults`` -- will be used in a ``get()`` call. If an object is found,
  764. ``get_or_create()`` returns a tuple of that object and ``False``. If an object
  765. is *not* found, ``get_or_create()`` will instantiate and save a new object,
  766. returning a tuple of the new object and ``True``. The new object will be
  767. created roughly according to this algorithm::
  768. defaults = kwargs.pop('defaults', {})
  769. params = dict([(k, v) for k, v in kwargs.items() if '__' not in k])
  770. params.update(defaults)
  771. obj = self.model(**params)
  773. In English, that means start with any non-``'defaults'`` keyword argument that
  774. doesn't contain a double underscore (which would indicate a non-exact lookup).
  775. Then add the contents of ``defaults``, overriding any keys if necessary, and
  776. use the result as the keyword arguments to the model class. As hinted at
  777. above, this is a simplification of the algorithm that is used, but it contains
  778. all the pertinent details. The internal implementation has some more
  779. error-checking than this and handles some extra edge-conditions; if you're
  780. interested, read the code.
  781. If you have a field named ``defaults`` and want to use it as an exact lookup in
  782. ``get_or_create()``, just use ``'defaults__exact'``, like so::
  783. Foo.objects.get_or_create(defaults__exact='bar', defaults={'defaults': 'baz'})
  784. The ``get_or_create()`` method has similar error behavior to ``create()``
  785. when you are using manually specified primary keys. If an object needs to be
  786. created and the key already exists in the database, an ``IntegrityError`` will
  787. be raised.
  788. Finally, a word on using ``get_or_create()`` in Django views. As mentioned
  789. earlier, ``get_or_create()`` is mostly useful in scripts that need to parse
  790. data and create new records if existing ones aren't available. But if you need
  791. to use ``get_or_create()`` in a view, please make sure to use it only in
  792. ``POST`` requests unless you have a good reason not to. ``GET`` requests
  793. shouldn't have any effect on data; use ``POST`` whenever a request to a page
  794. has a side effect on your data. For more, see `Safe methods`_ in the HTTP spec.
  795. .. _Safe methods:
  796. count
  797. ~~~~~
  798. .. method:: count()
  799. Returns an integer representing the number of objects in the database matching
  800. the ``QuerySet``. ``count()`` never raises exceptions.
  801. Example::
  802. # Returns the total number of entries in the database.
  803. Entry.objects.count()
  804. # Returns the number of entries whose headline contains 'Lennon'
  805. Entry.objects.filter(headline__contains='Lennon').count()
  806. ``count()`` performs a ``SELECT COUNT(*)`` behind the scenes, so you should
  807. always use ``count()`` rather than loading all of the record into Python
  808. objects and calling ``len()`` on the result (unless you need to load the
  809. objects into memory anyway, in which case ``len()`` will be faster).
  810. Depending on which database you're using (e.g. PostgreSQL vs. MySQL),
  811. ``count()`` may return a long integer instead of a normal Python integer. This
  812. is an underlying implementation quirk that shouldn't pose any real-world
  813. problems.
  814. in_bulk
  815. ~~~~~~~
  816. .. method:: in_bulk(id_list)
  817. Takes a list of primary-key values and returns a dictionary mapping each
  818. primary-key value to an instance of the object with the given ID.
  819. Example::
  820. >>> Blog.objects.in_bulk([1])
  821. {1: <Blog: Beatles Blog>}
  822. >>> Blog.objects.in_bulk([1, 2])
  823. {1: <Blog: Beatles Blog>, 2: <Blog: Cheddar Talk>}
  824. >>> Blog.objects.in_bulk([])
  825. {}
  826. If you pass ``in_bulk()`` an empty list, you'll get an empty dictionary.
  827. iterator
  828. ~~~~~~~~
  829. .. method:: iterator()
  830. Evaluates the ``QuerySet`` (by performing the query) and returns an
  831. `iterator`_ over the results. A ``QuerySet`` typically caches its
  832. results internally so that repeated evaluations do not result in
  833. additional queries; ``iterator()`` will instead read results directly,
  834. without doing any caching at the ``QuerySet`` level. For a
  835. ``QuerySet`` which returns a large number of objects, this often
  836. results in better performance and a significant reduction in memory
  837. Note that using ``iterator()`` on a ``QuerySet`` which has already
  838. been evaluated will force it to evaluate again, repeating the query.
  839. .. _iterator:
  840. latest
  841. ~~~~~~
  842. .. method:: latest(field_name=None)
  843. Returns the latest object in the table, by date, using the ``field_name``
  844. provided as the date field.
  845. This example returns the latest ``Entry`` in the table, according to the
  846. ``pub_date`` field::
  847. Entry.objects.latest('pub_date')
  848. If your model's ``Meta`` specifies ``get_latest_by``, you can leave off the
  849. ``field_name`` argument to ``latest()``. Django will use the field specified in
  850. ``get_latest_by`` by default.
  851. Like ``get()``, ``latest()`` raises ``DoesNotExist`` if an object doesn't
  852. exist with the given parameters.
  853. Note ``latest()`` exists purely for convenience and readability.
  854. aggregate
  855. ~~~~~~~~~
  856. .. method:: aggregate(*args, **kwargs)
  857. Returns a dictionary of aggregate values (averages, sums, etc) calculated
  858. over the ``QuerySet``. Each argument to ``aggregate()`` specifies
  859. a value that will be included in the dictionary that is returned.
  860. The aggregation functions that are provided by Django are described
  861. in `Aggregation Functions`_ below.
  862. Aggregates specified using keyword arguments will use the keyword as
  863. the name for the annotation. Anonymous arguments will have an name
  864. generated for them based upon the name of the aggregate function and
  865. the model field that is being aggregated.
  866. For example, if you were manipulating blog entries, you may want to know
  867. the number of authors that have contributed blog entries::
  868. >>> q = Blog.objects.aggregate(Count('entry'))
  869. {'entry__count': 16}
  870. By using a keyword argument to specify the aggregate function, you can
  871. control the name of the aggregation value that is returned::
  872. >>> q = Blog.objects.aggregate(number_of_entries=Count('entry'))
  873. {'number_of_entries': 16}
  874. For an in-depth discussion of aggregation, see :doc:`the topic guide on
  875. Aggregation </topics/db/aggregation>`.
  876. exists
  877. ~~~~~~
  878. .. method:: exists()
  879. .. versionadded:: 1.2
  880. Returns ``True`` if the :class:`.QuerySet` contains any results, and ``False``
  881. if not. This tries to perform the query in the simplest and fastest way
  882. possible, but it *does* execute nearly the same query. This means that calling
  883. :meth:`.QuerySet.exists` is faster than ``bool(some_query_set)``, but not by
  884. a large degree. If ``some_query_set`` has not yet been evaluated, but you know
  885. that it will be at some point, then using ``some_query_set.exists()`` will do
  886. more overall work (an additional query) than simply using
  887. ``bool(some_query_set)``.
  888. update
  889. ~~~~~~
  890. .. method:: update(**kwargs)
  891. Performs an SQL update query for the specified fields, and returns
  892. the number of rows affected. The ``update()`` method is applied instantly and
  893. the only restriction on the :class:`.QuerySet` that is updated is that it can
  894. only update columns in the model's main table. Filtering based on related
  895. fields is still possible. You cannot call ``update()`` on a
  896. :class:`.QuerySet` that has had a slice taken or can otherwise no longer be
  897. filtered.
  898. For example, if you wanted to update all the entries in a particular blog
  899. to use the same headline::
  900. >>> b = Blog.objects.get(pk=1)
  901. # Update all the headlines belonging to this Blog.
  902. >>> Entry.objects.select_related().filter(blog=b).update(headline='Everything is the same')
  903. The ``update()`` method does a bulk update and does not call any ``save()``
  904. methods on your models, nor does it emit the ``pre_save`` or ``post_save``
  905. signals (which are a consequence of calling ``save()``).
  906. delete
  907. ~~~~~~
  908. .. method:: delete()
  909. Performs an SQL delete query on all rows in the :class:`.QuerySet`. The
  910. ``delete()`` is applied instantly. You cannot call ``delete()`` on a
  911. :class:`.QuerySet` that has had a slice taken or can otherwise no longer be
  912. filtered.
  913. For example, to delete all the entries in a particular blog::
  914. >>> b = Blog.objects.get(pk=1)
  915. # Delete all the entries belonging to this Blog.
  916. >>> Entry.objects.filter(blog=b).delete()
  917. By default, Django's :class:`~django.db.models.ForeignKey` emulates the SQL
  918. constraint ``ON DELETE CASCADE`` -- in other words, any objects with foreign
  919. keys pointing at the objects to be deleted will be deleted along with them.
  920. For example::
  921. blogs = Blog.objects.all()
  922. # This will delete all Blogs and all of their Entry objects.
  923. blogs.delete()
  924. .. versionadded:: 1.3
  925. This cascade behavior is customizable via the
  926. :attr:`~django.db.models.ForeignKey.on_delete` argument to the
  927. :class:`~django.db.models.ForeignKey`.
  928. The ``delete()`` method does a bulk delete and does not call any ``delete()``
  929. methods on your models. It does, however, emit the
  930. :data:`~django.db.models.signals.pre_delete` and
  931. :data:`~django.db.models.signals.post_delete` signals for all deleted objects
  932. (including cascaded deletions).
  933. .. _field-lookups:
  934. Field lookups
  935. -------------
  936. Field lookups are how you specify the meat of an SQL ``WHERE`` clause. They're
  937. specified as keyword arguments to the ``QuerySet`` methods ``filter()``,
  938. ``exclude()`` and ``get()``.
  939. For an introduction, see :ref:`field-lookups-intro`.
  940. .. fieldlookup:: exact
  941. exact
  942. ~~~~~
  943. Exact match. If the value provided for comparison is ``None``, it will
  944. be interpreted as an SQL ``NULL`` (See isnull_ for more details).
  945. Examples::
  946. Entry.objects.get(id__exact=14)
  947. Entry.objects.get(id__exact=None)
  948. SQL equivalents::
  949. SELECT ... WHERE id = 14;
  950. SELECT ... WHERE id IS NULL;
  951. .. admonition:: MySQL comparisons
  952. In MySQL, a database table's "collation" setting determines whether
  953. ``exact`` comparisons are case-sensitive. This is a database setting, *not*
  954. a Django setting. It's possible to configure your MySQL tables to use
  955. case-sensitive comparisons, but some trade-offs are involved. For more
  956. information about this, see the :ref:`collation section <mysql-collation>`
  957. in the :doc:`databases </ref/databases>` documentation.
  958. .. fieldlookup:: iexact
  959. iexact
  960. ~~~~~~
  961. Case-insensitive exact match.
  962. Example::
  963. Blog.objects.get(name__iexact='beatles blog')
  964. SQL equivalent::
  965. SELECT ... WHERE name ILIKE 'beatles blog';
  966. Note this will match ``'Beatles Blog'``, ``'beatles blog'``, ``'BeAtLes
  967. BLoG'``, etc.
  968. .. admonition:: SQLite users
  969. When using the SQLite backend and Unicode (non-ASCII) strings, bear in
  970. mind the :ref:`database note <sqlite-string-matching>` about string
  971. comparisons. SQLite does not do case-insensitive matching for Unicode
  972. strings.
  973. .. fieldlookup:: contains
  974. contains
  975. ~~~~~~~~
  976. Case-sensitive containment test.
  977. Example::
  978. Entry.objects.get(headline__contains='Lennon')
  979. SQL equivalent::
  980. SELECT ... WHERE headline LIKE '%Lennon%';
  981. Note this will match the headline ``'Today Lennon honored'`` but not
  982. ``'today lennon honored'``.
  983. SQLite doesn't support case-sensitive ``LIKE`` statements; ``contains`` acts
  984. like ``icontains`` for SQLite.
  985. .. fieldlookup:: icontains
  986. icontains
  987. ~~~~~~~~~
  988. Case-insensitive containment test.
  989. Example::
  990. Entry.objects.get(headline__icontains='Lennon')
  991. SQL equivalent::
  992. SELECT ... WHERE headline ILIKE '%Lennon%';
  993. .. admonition:: SQLite users
  994. When using the SQLite backend and Unicode (non-ASCII) strings, bear in
  995. mind the :ref:`database note <sqlite-string-matching>` about string
  996. comparisons.
  997. .. fieldlookup:: in
  998. in
  999. ~~
  1000. In a given list.
  1001. Example::
  1002. Entry.objects.filter(id__in=[1, 3, 4])
  1003. SQL equivalent::
  1004. SELECT ... WHERE id IN (1, 3, 4);
  1005. You can also use a queryset to dynamically evaluate the list of values
  1006. instead of providing a list of literal values::
  1007. inner_qs = Blog.objects.filter(name__contains='Cheddar')
  1008. entries = Entry.objects.filter(blog__in=inner_qs)
  1009. This queryset will be evaluated as subselect statement::
  1010. SELECT ... WHERE IN (SELECT id FROM ... WHERE NAME LIKE '%Cheddar%')
  1011. The above code fragment could also be written as follows::
  1012. inner_q = Blog.objects.filter(name__contains='Cheddar').values('pk').query
  1013. entries = Entry.objects.filter(blog__in=inner_q)
  1014. This second form is a bit less readable and unnatural to write, since it
  1015. accesses the internal ``query`` attribute and requires a ``ValuesQuerySet``.
  1016. If your code doesn't require compatibility with Django 1.0, use the first
  1017. form, passing in a queryset directly.
  1018. If you pass in a ``ValuesQuerySet`` or ``ValuesListQuerySet`` (the result of
  1019. calling ``values()`` or ``values_list()`` on a queryset) as the value to an
  1020. ``__in`` lookup, you need to ensure you are only extracting one field in the
  1021. result. For example, this will work (filtering on the blog names)::
  1022. inner_qs = Blog.objects.filter(name__contains='Ch').values('name')
  1023. entries = Entry.objects.filter(blog__name__in=inner_qs)
  1024. This example will raise an exception, since the inner query is trying to
  1025. extract two field values, where only one is expected::
  1026. # Bad code! Will raise a TypeError.
  1027. inner_qs = Blog.objects.filter(name__contains='Ch').values('name', 'id')
  1028. entries = Entry.objects.filter(blog__name__in=inner_qs)
  1029. .. warning::
  1030. This ``query`` attribute should be considered an opaque internal attribute.
  1031. It's fine to use it like above, but its API may change between Django
  1032. versions.
  1033. .. admonition:: Performance considerations
  1034. Be cautious about using nested queries and understand your database
  1035. server's performance characteristics (if in doubt, benchmark!). Some
  1036. database backends, most notably MySQL, don't optimize nested queries very
  1037. well. It is more efficient, in those cases, to extract a list of values
  1038. and then pass that into the second query. That is, execute two queries
  1039. instead of one::
  1040. values = Blog.objects.filter(
  1041. name__contains='Cheddar').values_list('pk', flat=True)
  1042. entries = Entry.objects.filter(blog__in=list(values))
  1043. Note the ``list()`` call around the Blog ``QuerySet`` to force execution of
  1044. the first query. Without it, a nested query would be executed, because
  1045. :ref:`querysets-are-lazy`.
  1046. .. fieldlookup:: gt
  1047. gt
  1048. ~~
  1049. Greater than.
  1050. Example::
  1051. Entry.objects.filter(id__gt=4)
  1052. SQL equivalent::
  1053. SELECT ... WHERE id > 4;
  1054. .. fieldlookup:: gte
  1055. gte
  1056. ~~~
  1057. Greater than or equal to.
  1058. .. fieldlookup:: lt
  1059. lt
  1060. ~~
  1061. Less than.
  1062. .. fieldlookup:: lte
  1063. lte
  1064. ~~~
  1065. Less than or equal to.
  1066. .. fieldlookup:: startswith
  1067. startswith
  1068. ~~~~~~~~~~
  1069. Case-sensitive starts-with.
  1070. Example::
  1071. Entry.objects.filter(headline__startswith='Will')
  1072. SQL equivalent::
  1073. SELECT ... WHERE headline LIKE 'Will%';
  1074. SQLite doesn't support case-sensitive ``LIKE`` statements; ``startswith`` acts
  1075. like ``istartswith`` for SQLite.
  1076. .. fieldlookup:: istartswith
  1077. istartswith
  1078. ~~~~~~~~~~~
  1079. Case-insensitive starts-with.
  1080. Example::
  1081. Entry.objects.filter(headline__istartswith='will')
  1082. SQL equivalent::
  1083. SELECT ... WHERE headline ILIKE 'Will%';
  1084. .. admonition:: SQLite users
  1085. When using the SQLite backend and Unicode (non-ASCII) strings, bear in
  1086. mind the :ref:`database note <sqlite-string-matching>` about string
  1087. comparisons.
  1088. .. fieldlookup:: endswith
  1089. endswith
  1090. ~~~~~~~~
  1091. Case-sensitive ends-with.
  1092. Example::
  1093. Entry.objects.filter(headline__endswith='cats')
  1094. SQL equivalent::
  1095. SELECT ... WHERE headline LIKE '%cats';
  1096. SQLite doesn't support case-sensitive ``LIKE`` statements; ``endswith`` acts
  1097. like ``iendswith`` for SQLite.
  1098. .. fieldlookup:: iendswith
  1099. iendswith
  1100. ~~~~~~~~~
  1101. Case-insensitive ends-with.
  1102. Example::
  1103. Entry.objects.filter(headline__iendswith='will')
  1104. SQL equivalent::
  1105. SELECT ... WHERE headline ILIKE '%will'
  1106. .. admonition:: SQLite users
  1107. When using the SQLite backend and Unicode (non-ASCII) strings, bear in
  1108. mind the :ref:`database note <sqlite-string-matching>` about string
  1109. comparisons.
  1110. .. fieldlookup:: range
  1111. range
  1112. ~~~~~
  1113. Range test (inclusive).
  1114. Example::
  1115. start_date =, 1, 1)
  1116. end_date =, 3, 31)
  1117. Entry.objects.filter(pub_date__range=(start_date, end_date))
  1118. SQL equivalent::
  1119. SELECT ... WHERE pub_date BETWEEN '2005-01-01' and '2005-03-31';
  1120. You can use ``range`` anywhere you can use ``BETWEEN`` in SQL -- for dates,
  1121. numbers and even characters.
  1122. .. fieldlookup:: year
  1123. year
  1124. ~~~~
  1125. For date/datetime fields, exact year match. Takes a four-digit year.
  1126. Example::
  1127. Entry.objects.filter(pub_date__year=2005)
  1128. SQL equivalent::
  1129. SELECT ... WHERE pub_date BETWEEN '2005-01-01' AND '2005-12-31 23:59:59.999999';
  1130. (The exact SQL syntax varies for each database engine.)
  1131. .. fieldlookup:: month
  1132. month
  1133. ~~~~~
  1134. For date/datetime fields, exact month match. Takes an integer 1 (January)
  1135. through 12 (December).
  1136. Example::
  1137. Entry.objects.filter(pub_date__month=12)
  1138. SQL equivalent::
  1139. SELECT ... WHERE EXTRACT('month' FROM pub_date) = '12';
  1140. (The exact SQL syntax varies for each database engine.)
  1141. .. fieldlookup:: day
  1142. day
  1143. ~~~
  1144. For date/datetime fields, exact day match.
  1145. Example::
  1146. Entry.objects.filter(pub_date__day=3)
  1147. SQL equivalent::
  1148. SELECT ... WHERE EXTRACT('day' FROM pub_date) = '3';
  1149. (The exact SQL syntax varies for each database engine.)
  1150. Note this will match any record with a pub_date on the third day of the month,
  1151. such as January 3, July 3, etc.
  1152. .. fieldlookup:: week_day
  1153. week_day
  1154. ~~~~~~~~
  1155. For date/datetime fields, a 'day of the week' match.
  1156. Takes an integer value representing the day of week from 1 (Sunday) to 7
  1157. (Saturday).
  1158. Example::
  1159. Entry.objects.filter(pub_date__week_day=2)
  1160. (No equivalent SQL code fragment is included for this lookup because
  1161. implementation of the relevant query varies among different database engines.)
  1162. Note this will match any record with a pub_date that falls on a Monday (day 2
  1163. of the week), regardless of the month or year in which it occurs. Week days
  1164. are indexed with day 1 being Sunday and day 7 being Saturday.
  1165. .. fieldlookup:: isnull
  1166. isnull
  1167. ~~~~~~
  1168. Takes either ``True`` or ``False``, which correspond to SQL queries of
  1169. ``IS NULL`` and ``IS NOT NULL``, respectively.
  1170. Example::
  1171. Entry.objects.filter(pub_date__isnull=True)
  1172. SQL equivalent::
  1173. SELECT ... WHERE pub_date IS NULL;
  1174. .. fieldlookup:: search
  1175. search
  1176. ~~~~~~
  1177. A boolean full-text search, taking advantage of full-text indexing. This is
  1178. like ``contains`` but is significantly faster due to full-text indexing.
  1179. Example::
  1180. Entry.objects.filter(headline__search="+Django -jazz Python")
  1181. SQL equivalent::
  1182. SELECT ... WHERE MATCH(tablename, headline) AGAINST (+Django -jazz Python IN BOOLEAN MODE);
  1183. Note this is only available in MySQL and requires direct manipulation of the
  1184. database to add the full-text index. By default Django uses BOOLEAN MODE for
  1185. full text searches. `See the MySQL documentation for additional details.
  1186. <>`_
  1187. .. fieldlookup:: regex
  1188. regex
  1189. ~~~~~
  1190. Case-sensitive regular expression match.
  1191. The regular expression syntax is that of the database backend in use.
  1192. In the case of SQLite, which has no built in regular expression support,
  1193. this feature is provided by a (Python) user-defined REGEXP function, and
  1194. the regular expression syntax is therefore that of Python's ``re`` module.
  1195. Example::
  1196. Entry.objects.get(title__regex=r'^(An?|The) +')
  1197. SQL equivalents::
  1198. SELECT ... WHERE title REGEXP BINARY '^(An?|The) +'; -- MySQL
  1199. SELECT ... WHERE REGEXP_LIKE(title, '^(an?|the) +', 'c'); -- Oracle
  1200. SELECT ... WHERE title ~ '^(An?|The) +'; -- PostgreSQL
  1201. SELECT ... WHERE title REGEXP '^(An?|The) +'; -- SQLite
  1202. Using raw strings (e.g., ``r'foo'`` instead of ``'foo'``) for passing in the
  1203. regular expression syntax is recommended.
  1204. .. fieldlookup:: iregex
  1205. iregex
  1206. ~~~~~~
  1207. Case-insensitive regular expression match.
  1208. Example::
  1209. Entry.objects.get(title__iregex=r'^(an?|the) +')
  1210. SQL equivalents::
  1211. SELECT ... WHERE title REGEXP '^(an?|the) +'; -- MySQL
  1212. SELECT ... WHERE REGEXP_LIKE(title, '^(an?|the) +', 'i'); -- Oracle
  1213. SELECT ... WHERE title ~* '^(an?|the) +'; -- PostgreSQL
  1214. SELECT ... WHERE title REGEXP '(?i)^(an?|the) +'; -- SQLite
  1215. .. _aggregation-functions:
  1216. Aggregation functions
  1217. ---------------------
  1218. .. currentmodule:: django.db.models
  1219. Django provides the following aggregation functions in the
  1220. ``django.db.models`` module. For details on how to use these
  1221. aggregate functions, see
  1222. :doc:`the topic guide on aggregation </topics/db/aggregation>`.
  1223. Avg
  1224. ~~~
  1225. .. class:: Avg(field)
  1226. Returns the mean value of the given field.
  1227. * Default alias: ``<field>__avg``
  1228. * Return type: float
  1229. Count
  1230. ~~~~~
  1231. .. class:: Count(field, distinct=False)
  1232. Returns the number of objects that are related through the provided field.
  1233. * Default alias: ``<field>__count``
  1234. * Return type: integer
  1235. Has one optional argument:
  1236. .. attribute:: distinct
  1237. If distinct=True, the count will only include unique instances. This has
  1238. the SQL equivalent of ``COUNT(DISTINCT field)``. Default value is ``False``.
  1239. Max
  1240. ~~~
  1241. .. class:: Max(field)
  1242. Returns the maximum value of the given field.
  1243. * Default alias: ``<field>__max``
  1244. * Return type: same as input field
  1245. Min
  1246. ~~~
  1247. .. class:: Min(field)
  1248. Returns the minimum value of the given field.
  1249. * Default alias: ``<field>__min``
  1250. * Return type: same as input field
  1251. StdDev
  1252. ~~~~~~
  1253. .. class:: StdDev(field, sample=False)
  1254. Returns the standard deviation of the data in the provided field.
  1255. * Default alias: ``<field>__stddev``
  1256. * Return type: float
  1257. Has one optional argument:
  1258. .. attribute:: sample
  1259. By default, ``StdDev`` returns the population standard deviation. However,
  1260. if ``sample=True``, the return value will be the sample standard deviation.
  1261. .. admonition:: SQLite
  1262. SQLite doesn't provide ``StdDev`` out of the box. An implementation is
  1263. available as an extension module for SQLite. Consult the SQlite
  1264. documentation for instructions on obtaining and installing this extension.
  1265. Sum
  1266. ~~~
  1267. .. class:: Sum(field)
  1268. Computes the sum of all values of the given field.
  1269. * Default alias: ``<field>__sum``
  1270. * Return type: same as input field
  1271. Variance
  1272. ~~~~~~~~
  1273. .. class:: Variance(field, sample=False)
  1274. Returns the variance of the data in the provided field.
  1275. * Default alias: ``<field>__variance``
  1276. * Return type: float
  1277. Has one optional argument:
  1278. .. attribute:: sample
  1279. By default, ``Variance`` returns the population variance. However,
  1280. if ``sample=True``, the return value will be the sample variance.
  1281. .. admonition:: SQLite
  1282. SQLite doesn't provide ``Variance`` out of the box. An implementation is
  1283. available as an extension module for SQLite. Consult the SQlite
  1284. documentation for instructions on obtaining and installing this extension.