======================== Model instance reference ======================== .. currentmodule:: django.db.models This document describes the details of the ``Model`` API. It builds on the material presented in the :doc:`model ` and :doc:`database query ` guides, so you'll probably want to read and understand those documents before reading this one. Throughout this reference we'll use the :ref:`example Weblog models ` presented in the :doc:`database query guide `. Creating objects ================ To create a new instance of a model, just instantiate it like any other Python class: .. class:: Model(**kwargs) The keyword arguments are simply the names of the fields you've defined on your model. Note that instantiating a model in no way touches your database; for that, you need to :meth:`~Model.save()`. .. note:: You may be tempted to customize the model by overriding the ``__init__`` method. If you do so, however, take care not to change the calling signature as any change may prevent the model instance from being saved. Rather than overriding ``__init__``, try using one of these approaches: 1. Add a classmethod on the model class:: from django.db import models class Book(models.Model): title = models.CharField(max_length=100) @classmethod def create(cls, title): book = cls(title=title) # do something with the book return book book = Book.create("Pride and Prejudice") 2. Add a method on a custom manager (usually preferred):: class BookManager(models.Manager): def create_book(self, title): book = self.create(title=title) # do something with the book return book class Book(models.Model): title = models.CharField(max_length=100) objects = BookManager() book = Book.objects.create_book("Pride and Prejudice") Customizing model loading ------------------------- .. classmethod:: Model.from_db(db, field_names, values) The ``from_db()`` method can be used to customize model instance creation when loading from the database. The ``db`` argument contains the database alias for the database the model is loaded from, ``field_names`` contains the names of all loaded fields, and ``values`` contains the loaded values for each field in ``field_names``. The ``field_names`` are in the same order as the ``values``, so it is possible to use ``cls(**(zip(field_names, values)))`` to instantiate the object. If all of the model's fields are present, then ``values`` are guaranteed to be in the order ``__init__()`` expects them. That is, the instance can be created by ``cls(*values)``. It is possible to check if all fields are present by consulting ``cls._deferred`` - if ``False``, then all fields have been loaded from the database. In addition to creating the new model, the ``from_db()`` method must set the ``adding`` and ``db`` flags in the new instance's ``_state`` attribute. Below is an example showing how to record the initial values of fields that are loaded from the database:: @classmethod def from_db(cls, db, field_names, values): # default implementation of from_db() (could be replaced # with super()) if cls._deferred: instance = cls(**zip(field_names, values)) else: instance = cls(*values) instance._state.adding = False instance._state.db = db # customization to store the original field values on the instance instance._loaded_values = dict(zip(field_names, values)) return instance def save(self, *args, **kwargs): # Check how the current values differ from ._loaded_values. For example, # prevent changing the creator_id of the model. (This example doesn't # support cases where 'creator_id' is deferred). if not self._state.adding and ( self.creator_id != self._loaded_values['creator_id']): raise ValueError("Updating the value of creator isn't allowed") super(...).save(*args, **kwargs) The example above shows a full ``from_db()`` implementation to clarify how that is done. In this case it would of course be possible to just use ``super()`` call in the ``from_db()`` method. Refreshing objects from database ================================ .. method:: Model.refresh_from_db(using=None, fields=None, **kwargs) If you need to reload a model's values from the database, you can use the ``refresh_from_db()`` method. When this method is called without arguments the following is done: 1. All non-deferred fields of the model are updated to the values currently present in the database. 2. The previously loaded related instances for which the relation's value is no longer valid are removed from the reloaded instance. For example, if you have a foreign key from the reloaded instance to another model with name ``Author``, then if ``obj.author_id != obj.author.id``, ``obj.author`` will be thrown away, and when next accessed it will be reloaded with the value of ``obj.author_id``. Note that only fields of the model are reloaded from the database. Other database dependent values such as annotations are not reloaded. The reloading happens from the database the instance was loaded from, or from the default database if the instance wasn't loaded from the database. The ``using`` argument can be used to force the database used for reloading. It is possible to force the set of fields to be loaded by using the ``fields`` argument. For example, to test that an ``update()`` call resulted in the expected update, you could write a test similar to this:: def test_update_result(self): obj = MyModel.objects.create(val=1) MyModel.objects.filter(pk=obj.pk).update(val=F('val') + 1) # At this point obj.val is still 1, but the value in the database # was updated to 2. The object's updated value needs to be reloaded # from the database. obj.refresh_from_db() self.assertEqual(obj.val, 2) Note that when deferred fields are accessed, the loading of the deferred field's value happens through this method. Thus it is possible to customize the way deferred loading happens. The example below shows how one can reload all of the instance's fields when a deferred field is reloaded:: class ExampleModel(models.Model): def refresh_from_db(self, using=None, fields=None, **kwargs): # fields contains the name of the deferred field to be # loaded. if fields is not None: fields = set(fields) deferred_fields = self.get_deferred_fields() # If any deferred field is going to be loaded if fields.intersection(deferred_fields): # then load all of them fields = fields.union(deferred_fields) super(ExampleModel, self).refresh_from_db(using, fields, **kwargs) .. method:: Model.get_deferred_fields() A helper method that returns a set containing the attribute names of all those fields that are currently deferred for this model. .. _validating-objects: Validating objects ================== There are three steps involved in validating a model: 1. Validate the model fields - :meth:`Model.clean_fields()` 2. Validate the model as a whole - :meth:`Model.clean()` 3. Validate the field uniqueness - :meth:`Model.validate_unique()` All three steps are performed when you call a model's :meth:`~Model.full_clean()` method. When you use a :class:`~django.forms.ModelForm`, the call to :meth:`~django.forms.Form.is_valid()` will perform these validation steps for all the fields that are included on the form. See the :doc:`ModelForm documentation ` for more information. You should only need to call a model's :meth:`~Model.full_clean()` method if you plan to handle validation errors yourself, or if you have excluded fields from the :class:`~django.forms.ModelForm` that require validation. .. method:: Model.full_clean(exclude=None, validate_unique=True) This method calls :meth:`Model.clean_fields()`, :meth:`Model.clean()`, and :meth:`Model.validate_unique()` (if ``validate_unique`` is ``True``), in that order and raises a :exc:`~django.core.exceptions.ValidationError` that has a ``message_dict`` attribute containing errors from all three stages. The optional ``exclude`` argument can be used to provide a list of field names that can be excluded from validation and cleaning. :class:`~django.forms.ModelForm` uses this argument to exclude fields that aren't present on your form from being validated since any errors raised could not be corrected by the user. Note that ``full_clean()`` will *not* be called automatically when you call your model's :meth:`~Model.save()` method. You'll need to call it manually when you want to run one-step model validation for your own manually created models. For example:: from django.core.exceptions import ValidationError try: article.full_clean() except ValidationError as e: # Do something based on the errors contained in e.message_dict. # Display them to a user, or handle them programmatically. pass The first step ``full_clean()`` performs is to clean each individual field. .. method:: Model.clean_fields(exclude=None) This method will validate all fields on your model. The optional ``exclude`` argument lets you provide a list of field names to exclude from validation. It will raise a :exc:`~django.core.exceptions.ValidationError` if any fields fail validation. The second step ``full_clean()`` performs is to call :meth:`Model.clean()`. This method should be overridden to perform custom validation on your model. .. method:: Model.clean() This method should be used to provide custom model validation, and to modify attributes on your model if desired. For instance, you could use it to automatically provide a value for a field, or to do validation that requires access to more than a single field:: import datetime from django.core.exceptions import ValidationError from django.db import models from django.utils.translation import ugettext_lazy as _ class Article(models.Model): ... def clean(self): # Don't allow draft entries to have a pub_date. if self.status == 'draft' and self.pub_date is not None: raise ValidationError(_('Draft entries may not have a publication date.')) # Set the pub_date for published items if it hasn't been set already. if self.status == 'published' and self.pub_date is None: self.pub_date = datetime.date.today() Note, however, that like :meth:`Model.full_clean()`, a model's ``clean()`` method is not invoked when you call your model's :meth:`~Model.save()` method. In the above example, the :exc:`~django.core.exceptions.ValidationError` exception raised by ``Model.clean()`` was instantiated with a string, so it will be stored in a special error dictionary key, :data:`~django.core.exceptions.NON_FIELD_ERRORS`. This key is used for errors that are tied to the entire model instead of to a specific field:: from django.core.exceptions import ValidationError, NON_FIELD_ERRORS try: article.full_clean() except ValidationError as e: non_field_errors = e.message_dict[NON_FIELD_ERRORS] To assign exceptions to a specific field, instantiate the :exc:`~django.core.exceptions.ValidationError` with a dictionary, where the keys are the field names. We could update the previous example to assign the error to the ``pub_date`` field:: class Article(models.Model): ... def clean(self): # Don't allow draft entries to have a pub_date. if self.status == 'draft' and self.pub_date is not None: raise ValidationError({'pub_date': _('Draft entries may not have a publication date.')}) ... If you detect errors in multiple fields during ``Model.clean()``, you can also pass a dictionary mapping field names to errors:: raise ValidationError({ 'title': ValidationError(_('Missing title.'), code='required'), 'pub_date': ValidationError(_('Invalid date.'), code='invalid'), }) Finally, ``full_clean()`` will check any unique constraints on your model. .. method:: Model.validate_unique(exclude=None) This method is similar to :meth:`~Model.clean_fields`, but validates all uniqueness constraints on your model instead of individual field values. The optional ``exclude`` argument allows you to provide a list of field names to exclude from validation. It will raise a :exc:`~django.core.exceptions.ValidationError` if any fields fail validation. Note that if you provide an ``exclude`` argument to ``validate_unique()``, any :attr:`~django.db.models.Options.unique_together` constraint involving one of the fields you provided will not be checked. Saving objects ============== To save an object back to the database, call ``save()``: .. method:: Model.save(force_insert=False, force_update=False, using=DEFAULT_DB_ALIAS, update_fields=None) If you want customized saving behavior, you can override this ``save()`` method. See :ref:`overriding-model-methods` for more details. The model save process also has some subtleties; see the sections below. Auto-incrementing primary keys ------------------------------ If a model has an :class:`~django.db.models.AutoField` — an auto-incrementing primary key — then that auto-incremented value will be calculated and saved as an attribute on your object the first time you call ``save()``:: >>> b2 = Blog(name='Cheddar Talk', tagline='Thoughts on cheese.') >>> b2.id # Returns None, because b doesn't have an ID yet. >>> b2.save() >>> b2.id # Returns the ID of your new object. There's no way to tell what the value of an ID will be before you call ``save()``, because that value is calculated by your database, not by Django. For convenience, each model has an :class:`~django.db.models.AutoField` named ``id`` by default unless you explicitly specify ``primary_key=True`` on a field in your model. See the documentation for :class:`~django.db.models.AutoField` for more details. The ``pk`` property ~~~~~~~~~~~~~~~~~~~ .. attribute:: Model.pk Regardless of whether you define a primary key field yourself, or let Django supply one for you, each model will have a property called ``pk``. It behaves like a normal attribute on the model, but is actually an alias for whichever attribute is the primary key field for the model. You can read and set this value, just as you would for any other attribute, and it will update the correct field in the model. Explicitly specifying auto-primary-key values ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ If a model has an :class:`~django.db.models.AutoField` but you want to define a new object's ID explicitly when saving, just define it explicitly before saving, rather than relying on the auto-assignment of the ID:: >>> b3 = Blog(id=3, name='Cheddar Talk', tagline='Thoughts on cheese.') >>> b3.id # Returns 3. >>> b3.save() >>> b3.id # Returns 3. If you assign auto-primary-key values manually, make sure not to use an already-existing primary-key value! If you create a new object with an explicit primary-key value that already exists in the database, Django will assume you're changing the existing record rather than creating a new one. Given the above ``'Cheddar Talk'`` blog example, this example would override the previous record in the database:: b4 = Blog(id=3, name='Not Cheddar', tagline='Anything but cheese.') b4.save() # Overrides the previous blog with ID=3! See `How Django knows to UPDATE vs. INSERT`_, below, for the reason this happens. Explicitly specifying auto-primary-key values is mostly useful for bulk-saving objects, when you're confident you won't have primary-key collision. What happens when you save? --------------------------- When you save an object, Django performs the following steps: 1. **Emit a pre-save signal.** The :doc:`signal ` :attr:`django.db.models.signals.pre_save` is sent, allowing any functions listening for that signal to take some customized action. 2. **Pre-process the data.** Each field on the object is asked to perform any automated data modification that the field may need to perform. Most fields do *no* pre-processing — the field data is kept as-is. Pre-processing is only used on fields that have special behavior. For example, if your model has a :class:`~django.db.models.DateField` with ``auto_now=True``, the pre-save phase will alter the data in the object to ensure that the date field contains the current date stamp. (Our documentation doesn't yet include a list of all the fields with this "special behavior.") 3. **Prepare the data for the database.** Each field is asked to provide its current value in a data type that can be written to the database. Most fields require *no* data preparation. Simple data types, such as integers and strings, are 'ready to write' as a Python object. However, more complex data types often require some modification. For example, :class:`~django.db.models.DateField` fields use a Python ``datetime`` object to store data. Databases don't store ``datetime`` objects, so the field value must be converted into an ISO-compliant date string for insertion into the database. 4. **Insert the data into the database.** The pre-processed, prepared data is then composed into an SQL statement for insertion into the database. 5. **Emit a post-save signal.** The signal :attr:`django.db.models.signals.post_save` is sent, allowing any functions listening for that signal to take some customized action. How Django knows to UPDATE vs. INSERT ------------------------------------- You may have noticed Django database objects use the same ``save()`` method for creating and changing objects. Django abstracts the need to use ``INSERT`` or ``UPDATE`` SQL statements. Specifically, when you call ``save()``, Django follows this algorithm: * If the object's primary key attribute is set to a value that evaluates to ``True`` (i.e., a value other than ``None`` or the empty string), Django executes an ``UPDATE``. * If the object's primary key attribute is *not* set or if the ``UPDATE`` didn't update anything, Django executes an ``INSERT``. The one gotcha here is that you should be careful not to specify a primary-key value explicitly when saving new objects, if you cannot guarantee the primary-key value is unused. For more on this nuance, see `Explicitly specifying auto-primary-key values`_ above and `Forcing an INSERT or UPDATE`_ below. In Django 1.5 and earlier, Django did a ``SELECT`` when the primary key attribute was set. If the ``SELECT`` found a row, then Django did an ``UPDATE``, otherwise it did an ``INSERT``. The old algorithm results in one more query in the ``UPDATE`` case. There are some rare cases where the database doesn't report that a row was updated even if the database contains a row for the object's primary key value. An example is the PostgreSQL ``ON UPDATE`` trigger which returns ``NULL``. In such cases it is possible to revert to the old algorithm by setting the :attr:`~django.db.models.Options.select_on_save` option to ``True``. .. _ref-models-force-insert: Forcing an INSERT or UPDATE ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In some rare circumstances, it's necessary to be able to force the :meth:`~Model.save()` method to perform an SQL ``INSERT`` and not fall back to doing an ``UPDATE``. Or vice-versa: update, if possible, but not insert a new row. In these cases you can pass the ``force_insert=True`` or ``force_update=True`` parameters to the :meth:`~Model.save()` method. Obviously, passing both parameters is an error: you cannot both insert *and* update at the same time! It should be very rare that you'll need to use these parameters. Django will almost always do the right thing and trying to override that will lead to errors that are difficult to track down. This feature is for advanced use only. Using ``update_fields`` will force an update similarly to ``force_update``. .. _ref-models-field-updates-using-f-expressions: Updating attributes based on existing fields -------------------------------------------- Sometimes you'll need to perform a simple arithmetic task on a field, such as incrementing or decrementing the current value. The obvious way to achieve this is to do something like:: >>> product = Product.objects.get(name='Venezuelan Beaver Cheese') >>> product.number_sold += 1 >>> product.save() If the old ``number_sold`` value retrieved from the database was 10, then the value of 11 will be written back to the database. The process can be made robust, :ref:`avoiding a race condition `, as well as slightly faster by expressing the update relative to the original field value, rather than as an explicit assignment of a new value. Django provides :class:`F expressions ` for performing this kind of relative update. Using :class:`F expressions `, the previous example is expressed as:: >>> from django.db.models import F >>> product = Product.objects.get(name='Venezuelan Beaver Cheese') >>> product.number_sold = F('number_sold') + 1 >>> product.save() For more details, see the documentation on :class:`F expressions ` and their :ref:`use in update queries `. Specifying which fields to save ------------------------------- If ``save()`` is passed a list of field names in keyword argument ``update_fields``, only the fields named in that list will be updated. This may be desirable if you want to update just one or a few fields on an object. There will be a slight performance benefit from preventing all of the model fields from being updated in the database. For example:: product.name = 'Name changed again' product.save(update_fields=['name']) The ``update_fields`` argument can be any iterable containing strings. An empty ``update_fields`` iterable will skip the save. A value of None will perform an update on all fields. Specifying ``update_fields`` will force an update. When saving a model fetched through deferred model loading (:meth:`~django.db.models.query.QuerySet.only()` or :meth:`~django.db.models.query.QuerySet.defer()`) only the fields loaded from the DB will get updated. In effect there is an automatic ``update_fields`` in this case. If you assign or change any deferred field value, the field will be added to the updated fields. Deleting objects ================ .. method:: Model.delete(using=DEFAULT_DB_ALIAS, keep_parents=False) Issues an SQL ``DELETE`` for the object. This only deletes the object in the database; the Python instance will still exist and will still have data in its fields. This method returns the number of objects deleted and a dictionary with the number of deletions per object type. For more details, including how to delete objects in bulk, see :ref:`topics-db-queries-delete`. If you want customized deletion behavior, you can override the ``delete()`` method. See :ref:`overriding-model-methods` for more details. Sometimes with :ref:`multi-table inheritance ` you may want to delete only a child model's data. Specifying ``keep_parents=True`` will keep the parent model's data. .. versionchanged:: 1.9 The ``keep_parents`` parameter was added. .. versionchanged:: 1.9 The return value describing the number of objects deleted was added. Pickling objects ================ When you :mod:`pickle` a model, its current state is pickled. When you unpickle it, it'll contain the model instance at the moment it was pickled, rather than the data that's currently in the database. .. admonition:: You can't share pickles between versions Pickles of models are only valid for the version of Django that was used to generate them. If you generate a pickle using Django version N, there is no guarantee that pickle will be readable with Django version N+1. Pickles should not be used as part of a long-term archival strategy. Since pickle compatibility errors can be difficult to diagnose, such as silently corrupted objects, a ``RuntimeWarning`` is raised when you try to unpickle a model in a Django version that is different than the one in which it was pickled. .. _model-instance-methods: Other model instance methods ============================ A few object methods have special purposes. ``__str__`` ----------- .. method:: Model.__str__() The ``__str__()`` method is called whenever you call ``str()`` on an object. Django uses ``str(obj)`` in a number of places. Most notably, to display an object in the Django admin site and as the value inserted into a template when it displays an object. Thus, you should always return a nice, human-readable representation of the model from the ``__str__()`` method. For example:: from django.db import models from django.utils.encoding import python_2_unicode_compatible @python_2_unicode_compatible # only if you need to support Python 2 class Person(models.Model): first_name = models.CharField(max_length=50) last_name = models.CharField(max_length=50) def __str__(self): return '%s %s' % (self.first_name, self.last_name) If you'd like compatibility with Python 2, you can decorate your model class with :func:`~django.utils.encoding.python_2_unicode_compatible` as shown above. ``__eq__`` ---------- .. method:: Model.__eq__() The equality method is defined such that instances with the same primary key value and the same concrete class are considered equal. For proxy models, concrete class is defined as the model's first non-proxy parent; for all other models it is simply the model's class. For example:: from django.db import models class MyModel(models.Model): id = models.AutoField(primary_key=True) class MyProxyModel(MyModel): class Meta: proxy = True class MultitableInherited(MyModel): pass MyModel(id=1) == MyModel(id=1) MyModel(id=1) == MyProxyModel(id=1) MyModel(id=1) != MultitableInherited(id=1) MyModel(id=1) != MyModel(id=2) ``__hash__`` ------------ .. method:: Model.__hash__() The ``__hash__`` method is based on the instance's primary key value. It is effectively hash(obj.pk). If the instance doesn't have a primary key value then a ``TypeError`` will be raised (otherwise the ``__hash__`` method would return different values before and after the instance is saved, but changing the ``__hash__`` value of an instance `is forbidden in Python`_). .. _is forbidden in Python: https://docs.python.org/reference/datamodel.html#object.__hash__ ``get_absolute_url`` -------------------- .. method:: Model.get_absolute_url() Define a ``get_absolute_url()`` method to tell Django how to calculate the canonical URL for an object. To callers, this method should appear to return a string that can be used to refer to the object over HTTP. For example:: def get_absolute_url(self): return "/people/%i/" % self.id (Whilst this code is correct and simple, it may not be the most portable way to write this kind of method. The :func:`~django.core.urlresolvers.reverse` function is usually the best approach.) For example:: def get_absolute_url(self): from django.core.urlresolvers import reverse return reverse('people.views.details', args=[str(self.id)]) One place Django uses ``get_absolute_url()`` is in the admin app. If an object defines this method, the object-editing page will have a "View on site" link that will jump you directly to the object's public view, as given by ``get_absolute_url()``. Similarly, a couple of other bits of Django, such as the :doc:`syndication feed framework `, use ``get_absolute_url()`` when it is defined. If it makes sense for your model's instances to each have a unique URL, you should define ``get_absolute_url()``. .. warning:: You should avoid building the URL from unvalidated user input, in order to reduce possibilities of link or redirect poisoning:: def get_absolute_url(self): return '/%s/' % self.name If ``self.name`` is ``'/example.com'`` this returns ``'//example.com/'`` which, in turn, is a valid schema relative URL but not the expected ``'/%2Fexample.com/'``. It's good practice to use ``get_absolute_url()`` in templates, instead of hard-coding your objects' URLs. For example, this template code is bad: .. code-block:: html+django {{ object.name }} This template code is much better: .. code-block:: html+django {{ object.name }} The logic here is that if you change the URL structure of your objects, even for something simple such as correcting a spelling error, you don't want to have to track down every place that the URL might be created. Specify it once, in ``get_absolute_url()`` and have all your other code call that one place. .. note:: The string you return from ``get_absolute_url()`` **must** contain only ASCII characters (required by the URI specification, :rfc:`2396`) and be URL-encoded, if necessary. Code and templates calling ``get_absolute_url()`` should be able to use the result directly without any further processing. You may wish to use the ``django.utils.encoding.iri_to_uri()`` function to help with this if you are using unicode strings containing characters outside the ASCII range at all. Extra instance methods ====================== In addition to :meth:`~Model.save()`, :meth:`~Model.delete()`, a model object might have some of the following methods: .. method:: Model.get_FOO_display() For every field that has :attr:`~django.db.models.Field.choices` set, the object will have a ``get_FOO_display()`` method, where ``FOO`` is the name of the field. This method returns the "human-readable" value of the field. For example:: from django.db import models class Person(models.Model): SHIRT_SIZES = ( ('S', 'Small'), ('M', 'Medium'), ('L', 'Large'), ) name = models.CharField(max_length=60) shirt_size = models.CharField(max_length=2, choices=SHIRT_SIZES) :: >>> p = Person(name="Fred Flintstone", shirt_size="L") >>> p.save() >>> p.shirt_size 'L' >>> p.get_shirt_size_display() 'Large' .. method:: Model.get_next_by_FOO(\**kwargs) .. method:: Model.get_previous_by_FOO(\**kwargs) For every :class:`~django.db.models.DateField` and :class:`~django.db.models.DateTimeField` that does not have :attr:`null=True `, the object will have ``get_next_by_FOO()`` and ``get_previous_by_FOO()`` methods, where ``FOO`` is the name of the field. This returns the next and previous object with respect to the date field, raising a :exc:`~django.db.models.Model.DoesNotExist` exception when appropriate. Both of these methods will perform their queries using the default manager for the model. If you need to emulate filtering used by a custom manager, or want to perform one-off custom filtering, both methods also accept optional keyword arguments, which should be in the format described in :ref:`Field lookups `. Note that in the case of identical date values, these methods will use the primary key as a tie-breaker. This guarantees that no records are skipped or duplicated. That also means you cannot use those methods on unsaved objects. Other attributes ================ ``DoesNotExist`` ---------------- .. exception:: Model.DoesNotExist This exception is raised by the ORM in a couple places, for example by :meth:`QuerySet.get() ` when an object is not found for the given query parameters. Django provides a ``DoesNotExist`` exception as an attribute of each model class to identify the class of object that could not be found and to allow you to catch a particular model class with ``try/except``. The exception is a subclass of :exc:`django.core.exceptions.ObjectDoesNotExist`.