Models

A model is the single, definitive source of information about your data. Itcontains the essential fields and behaviors of the data you’re storing.Generally, each model maps to a single database table.

The basics:

  • Each model is a Python class that subclassesdjango.db.models.Model.
  • Each attribute of the model represents a database field.
  • With all of this, Django gives you an automatically-generateddatabase-access API; see Making queries.

Quick example

This example model defines a Person, which has a first_name andlast_name:

  1. from django.db import models
  2.  
  3. class Person(models.Model):
  4. first_name = models.CharField(max_length=30)
  5. last_name = models.CharField(max_length=30)

first_name and last_name are fields of the model. Each field isspecified as a class attribute, and each attribute maps to a database column.

The above Person model would create a database table like this:

  1. CREATE TABLE myapp_person (
  2. "id" serial NOT NULL PRIMARY KEY,
  3. "first_name" varchar(30) NOT NULL,
  4. "last_name" varchar(30) NOT NULL
  5. );

Some technical notes:

  • The name of the table, myapp_person, is automatically derived fromsome model metadata but can be overridden. See Table names for moredetails.
  • An id field is added automatically, but this behavior can beoverridden. See Automatic primary key fields.
  • The CREATE TABLE SQL in this example is formatted using PostgreSQLsyntax, but it’s worth noting Django uses SQL tailored to the databasebackend specified in your settings file.

Using models

Once you have defined your models, you need to tell Django you’re going to _use_those models. Do this by editing your settings file and changing theINSTALLED_APPS setting to add the name of the module that containsyour models.py.

For example, if the models for your application live in the modulemyapp.models (the package structure that is created for anapplication by the manage.py startapp script),INSTALLED_APPS should read, in part:

  1. INSTALLED_APPS = [
  2. #...
  3. 'myapp',
  4. #...
  5. ]

When you add new apps to INSTALLED_APPS, be sure to runmanage.py migrate, optionally making migrationsfor them first with manage.py makemigrations.

Fields

The most important part of a model – and the only required part of a model –is the list of database fields it defines. Fields are specified by classattributes. Be careful not to choose field names that conflict with themodels API like clean, save, ordelete.

Example:

  1. from django.db import models
  2.  
  3. class Musician(models.Model):
  4. first_name = models.CharField(max_length=50)
  5. last_name = models.CharField(max_length=50)
  6. instrument = models.CharField(max_length=100)
  7.  
  8. class Album(models.Model):
  9. artist = models.ForeignKey(Musician, on_delete=models.CASCADE)
  10. name = models.CharField(max_length=100)
  11. release_date = models.DateField()
  12. num_stars = models.IntegerField()

Field types

Each field in your model should be an instance of the appropriateField class. Django uses the field class types todetermine a few things:

  • The column type, which tells the database what kind of data to store (e.g.INTEGER, VARCHAR, TEXT).
  • The default HTML widget to use when rendering a formfield (e.g. <input type="text">, <select>).
  • The minimal validation requirements, used in Django’s admin and inautomatically-generated forms.Django ships with dozens of built-in field types; you can find the complete listin the model field reference. You can easily writeyour own fields if Django’s built-in ones don’t do the trick; seeWriting custom model fields.

Field options

Each field takes a certain set of field-specific arguments (documented in themodel field reference). For example,CharField (and its subclasses) require amax_length argument which specifies the sizeof the VARCHAR database field used to store the data.

There’s also a set of common arguments available to all field types. All areoptional. They’re fully explained in the reference, but here’s a quick summary of the most often-usedones:

  • null
  • If True, Django will store empty values as NULL in the database.Default is False.
  • blank
  • If True, the field is allowed to be blank. Default is False.

Note that this is different than null.null is purely database-related, whereasblank is validation-related. If a field hasblank=True, form validation willallow entry of an empty value. If a field has blank=False, the field will be required.

  • choices
  • A sequence of 2-tuples to use as choices for this field. If thisis given, the default form widget will be a select box instead of thestandard text field and will limit choices to the choices given.

A choices list looks like this:

  1. YEAR_IN_SCHOOL_CHOICES = [
  2. ('FR', 'Freshman'),
  3. ('SO', 'Sophomore'),
  4. ('JR', 'Junior'),
  5. ('SR', 'Senior'),
  6. ('GR', 'Graduate'),
  7. ]

Note

A new migration is created each time the order of choices changes.

The first element in each tuple is the value that will be stored in thedatabase. The second element is displayed by the field’s form widget.

Given a model instance, the display value for a field with choices canbe accessed using the get_FOO_display()method. For example:

  1. from django.db import models
  2.  
  3. class Person(models.Model):
  4. SHIRT_SIZES = (
  5. ('S', 'Small'),
  6. ('M', 'Medium'),
  7. ('L', 'Large'),
  8. )
  9. name = models.CharField(max_length=60)
  10. shirt_size = models.CharField(max_length=1, choices=SHIRT_SIZES)
  1. >>> p = Person(name="Fred Flintstone", shirt_size="L")
  2. >>> p.save()
  3. >>> p.shirt_size
  4. 'L'
  5. >>> p.get_shirt_size_display()
  6. 'Large'

You can also use enumeration classes to define choices in a conciseway:

  1. from django.db import models
  2.  
  3. class Runner(models.Model):
  4. MedalType = models.TextChoices('MedalType', 'GOLD SILVER BRONZE')
  5. name = models.CharField(max_length=60)
  6. medal = models.CharField(blank=True, choices=MedalType.choices, max_length=10)

Further examples are available in the model field reference.

  • default
  • The default value for the field. This can be a value or a callableobject. If callable it will be called every time a new object iscreated.
  • help_text
  • Extra “help” text to be displayed with the form widget. It’s useful fordocumentation even if your field isn’t used on a form.
  • primary_key
  • If True, this field is the primary key for the model.

If you don’t specify primary_key=True forany fields in your model, Django will automatically add anIntegerField to hold the primary key, so you don’t need to setprimary_key=True on any of your fieldsunless you want to override the default primary-key behavior. For more,see Automatic primary key fields.

The primary key field is read-only. If you change the value of the primarykey on an existing object and then save it, a new object will be createdalongside the old one. For example:

  1. from django.db import models
  2.  
  3. class Fruit(models.Model):
  4. name = models.CharField(max_length=100, primary_key=True)
  1. >>> fruit = Fruit.objects.create(name='Apple')
  2. >>> fruit.name = 'Pear'
  3. >>> fruit.save()
  4. >>> Fruit.objects.values_list('name', flat=True)
  5. <QuerySet ['Apple', 'Pear']>
  • unique
  • If True, this field must be unique throughout the table.Again, these are just short descriptions of the most common field options. Fulldetails can be found in the common model field option reference.

Automatic primary key fields

By default, Django gives each model the following field:

  1. id = models.AutoField(primary_key=True)

This is an auto-incrementing primary key.

If you’d like to specify a custom primary key, specifyprimary_key=True on one of your fields. If Djangosees you’ve explicitly set Field.primary_key, it won’t add the automaticid column.

Each model requires exactly one field to have primary_key=True (either explicitly declared or automatically added).

Verbose field names

Each field type, except for ForeignKey,ManyToManyField andOneToOneField, takes an optional first positionalargument – a verbose name. If the verbose name isn’t given, Django willautomatically create it using the field’s attribute name, converting underscoresto spaces.

In this example, the verbose name is "person's first name":

  1. first_name = models.CharField("person's first name", max_length=30)

In this example, the verbose name is "first name":

  1. first_name = models.CharField(max_length=30)

ForeignKey,ManyToManyField andOneToOneField require the first argument to be amodel class, so use the verbose_name keyword argument:

  1. poll = models.ForeignKey(
  2. Poll,
  3. on_delete=models.CASCADE,
  4. verbose_name="the related poll",
  5. )
  6. sites = models.ManyToManyField(Site, verbose_name="list of sites")
  7. place = models.OneToOneField(
  8. Place,
  9. on_delete=models.CASCADE,
  10. verbose_name="related place",
  11. )

The convention is not to capitalize the first letter of theverbose_name. Django will automatically capitalize the firstletter where it needs to.

Relationships

Clearly, the power of relational databases lies in relating tables to eachother. Django offers ways to define the three most common types of databaserelationships: many-to-one, many-to-many and one-to-one.

Many-to-one relationships

To define a many-to-one relationship, use django.db.models.ForeignKey.You use it just like any other Field type: byincluding it as a class attribute of your model.

ForeignKey requires a positional argument: the classto which the model is related.

For example, if a Car model has a Manufacturer – that is, aManufacturer makes multiple cars but each Car only has oneManufacturer – use the following definitions:

  1. from django.db import models
  2.  
  3. class Manufacturer(models.Model):
  4. # ...
  5. pass
  6.  
  7. class Car(models.Model):
  8. manufacturer = models.ForeignKey(Manufacturer, on_delete=models.CASCADE)
  9. # ...

You can also create recursive relationships (anobject with a many-to-one relationship to itself) and relationships tomodels not yet defined; see the model fieldreference for details.

It’s suggested, but not required, that the name of aForeignKey field (manufacturer in the exampleabove) be the name of the model, lowercase. You can, of course, call the fieldwhatever you want. For example:

  1. class Car(models.Model):
  2. company_that_makes_it = models.ForeignKey(
  3. Manufacturer,
  4. on_delete=models.CASCADE,
  5. )
  6. # ...

See also

ForeignKey fields accept a number of extraarguments which are explained in the model field reference. These options help define how the relationshipshould work; all are optional.

For details on accessing backwards-related objects, see theFollowing relationships backward example.

For sample code, see the Many-to-one relationship model example.

Many-to-many relationships

To define a many-to-many relationship, useManyToManyField. You use it just like any otherField type: by including it as a class attribute ofyour model.

ManyToManyField requires a positional argument: theclass to which the model is related.

For example, if a Pizza has multiple Topping objects – that is, aTopping can be on multiple pizzas and each Pizza has multiple toppings– here’s how you’d represent that:

  1. from django.db import models
  2.  
  3. class Topping(models.Model):
  4. # ...
  5. pass
  6.  
  7. class Pizza(models.Model):
  8. # ...
  9. toppings = models.ManyToManyField(Topping)

As with ForeignKey, you can also createrecursive relationships (an object with amany-to-many relationship to itself) and relationships to models not yetdefined.

It’s suggested, but not required, that the name of aManyToManyField (toppings in the example above)be a plural describing the set of related model objects.

It doesn’t matter which model has theManyToManyField, but you should only put it in oneof the models – not both.

Generally, ManyToManyField instances should go inthe object that’s going to be edited on a form. In the above example,toppings is in Pizza (rather than Topping having a pizzasManyToManyField ) because it’s more natural to thinkabout a pizza having toppings than a topping being on multiple pizzas. The wayit’s set up above, the Pizza form would let users select the toppings.

See also

See the Many-to-many relationship model example for a full example.

ManyToManyField fields also accept a number ofextra arguments which are explained in the model field reference. These options help define how the relationshipshould work; all are optional.

Extra fields on many-to-many relationships

When you’re only dealing with many-to-many relationships such as mixing andmatching pizzas and toppings, a standardManyToManyField is all you need. However, sometimesyou may need to associate data with the relationship between two models.

For example, consider the case of an application tracking the musical groupswhich musicians belong to. There is a many-to-many relationship between a personand the groups of which they are a member, so you could use aManyToManyField to represent this relationship.However, there is a lot of detail about the membership that you might want tocollect, such as the date at which the person joined the group.

For these situations, Django allows you to specify the model that will be usedto govern the many-to-many relationship. You can then put extra fields on theintermediate model. The intermediate model is associated with theManyToManyField using thethrough argument to point to the modelthat will act as an intermediary. For our musician example, the code would looksomething like this:

  1. from django.db import models
  2.  
  3. class Person(models.Model):
  4. name = models.CharField(max_length=128)
  5.  
  6. def __str__(self):
  7. return self.name
  8.  
  9. class Group(models.Model):
  10. name = models.CharField(max_length=128)
  11. members = models.ManyToManyField(Person, through='Membership')
  12.  
  13. def __str__(self):
  14. return self.name
  15.  
  16. class Membership(models.Model):
  17. person = models.ForeignKey(Person, on_delete=models.CASCADE)
  18. group = models.ForeignKey(Group, on_delete=models.CASCADE)
  19. date_joined = models.DateField()
  20. invite_reason = models.CharField(max_length=64)

When you set up the intermediary model, you explicitly specify foreignkeys to the models that are involved in the many-to-many relationship. Thisexplicit declaration defines how the two models are related.

There are a few restrictions on the intermediate model:

  • Your intermediate model must contain one - and only one - foreign keyto the source model (this would be Group in our example), or you mustexplicitly specify the foreign keys Django should use for the relationshipusing ManyToManyField.through_fields.If you have more than one foreign key and through_fields is notspecified, a validation error will be raised. A similar restriction appliesto the foreign key to the target model (this would be Person in ourexample).
  • For a model which has a many-to-many relationship to itself through anintermediary model, two foreign keys to the same model are permitted, butthey will be treated as the two (different) sides of the many-to-manyrelationship. If there are more than two foreign keys though, youmust also specify through_fields as above, or a validation errorwill be raised.
  • When defining a many-to-many relationship from a model toitself, using an intermediary model, you must usesymmetrical=False (seethe model field reference).Now that you have set up your ManyToManyField to useyour intermediary model (Membership, in this case), you’re ready to startcreating some many-to-many relationships. You do this by creating instances ofthe intermediate model:
  1. >>> ringo = Person.objects.create(name="Ringo Starr")
  2. >>> paul = Person.objects.create(name="Paul McCartney")
  3. >>> beatles = Group.objects.create(name="The Beatles")
  4. >>> m1 = Membership(person=ringo, group=beatles,
  5. ... date_joined=date(1962, 8, 16),
  6. ... invite_reason="Needed a new drummer.")
  7. >>> m1.save()
  8. >>> beatles.members.all()
  9. <QuerySet [<Person: Ringo Starr>]>
  10. >>> ringo.group_set.all()
  11. <QuerySet [<Group: The Beatles>]>
  12. >>> m2 = Membership.objects.create(person=paul, group=beatles,
  13. ... date_joined=date(1960, 8, 1),
  14. ... invite_reason="Wanted to form a band.")
  15. >>> beatles.members.all()
  16. <QuerySet [<Person: Ringo Starr>, <Person: Paul McCartney>]>

You can also use add(), create(), or set() to create relationships,as long as you specify through_defaults for any required fields:

  1. >>> beatles.members.add(john, through_defaults={'date_joined': date(1960, 8, 1)})
  2. >>> beatles.members.create(name="George Harrison", through_defaults={'date_joined': date(1960, 8, 1)})
  3. >>> beatles.members.set([john, paul, ringo, george], through_defaults={'date_joined': date(1960, 8, 1)})

You may prefer to create instances of the intermediate model directly.

If the custom through table defined by the intermediate model does not enforceuniqueness on the (model1, model2) pair, allowing multiple values, theremove() call willremove all intermediate model instances:

  1. >>> Membership.objects.create(person=ringo, group=beatles,
  2. ... date_joined=date(1968, 9, 4),
  3. ... invite_reason="You've been gone for a month and we miss you.")
  4. >>> beatles.members.all()
  5. <QuerySet [<Person: Ringo Starr>, <Person: Paul McCartney>, <Person: Ringo Starr>]>
  6. >>> # This deletes both of the intermediate model instances for Ringo Starr
  7. >>> beatles.members.remove(ringo)
  8. >>> beatles.members.all()
  9. <QuerySet [<Person: Paul McCartney>]>

The clear()method can be used to remove all many-to-many relationships for an instance:

  1. >>> # Beatles have broken up
  2. >>> beatles.members.clear()
  3. >>> # Note that this deletes the intermediate model instances
  4. >>> Membership.objects.all()
  5. <QuerySet []>

Once you have established the many-to-many relationships, you can issuequeries. Just as with normal many-to-many relationships, you can query usingthe attributes of the many-to-many-related model:

  1. # Find all the groups with a member whose name starts with 'Paul'
  2. >>> Group.objects.filter(members__name__startswith='Paul')
  3. <QuerySet [<Group: The Beatles>]>

As you are using an intermediate model, you can also query on its attributes:

  1. # Find all the members of the Beatles that joined after 1 Jan 1961
  2. >>> Person.objects.filter(
  3. ... group__name='The Beatles',
  4. ... membership__date_joined__gt=date(1961,1,1))
  5. <QuerySet [<Person: Ringo Starr]>

If you need to access a membership’s information you may do so by directlyquerying the Membership model:

  1. >>> ringos_membership = Membership.objects.get(group=beatles, person=ringo)
  2. >>> ringos_membership.date_joined
  3. datetime.date(1962, 8, 16)
  4. >>> ringos_membership.invite_reason
  5. 'Needed a new drummer.'

Another way to access the same information is by querying themany-to-many reverse relationship from aPerson object:

  1. >>> ringos_membership = ringo.membership_set.get(group=beatles)
  2. >>> ringos_membership.date_joined
  3. datetime.date(1962, 8, 16)
  4. >>> ringos_membership.invite_reason
  5. 'Needed a new drummer.'

One-to-one relationships

To define a one-to-one relationship, useOneToOneField. You use it just like any otherField type: by including it as a class attribute of your model.

This is most useful on the primary key of an object when that object “extends”another object in some way.

OneToOneField requires a positional argument: theclass to which the model is related.

For example, if you were building a database of “places”, you wouldbuild pretty standard stuff such as address, phone number, etc. in thedatabase. Then, if you wanted to build a database of restaurants ontop of the places, instead of repeating yourself and replicating thosefields in the Restaurant model, you could make Restaurant havea OneToOneField to Place (because arestaurant “is a” place; in fact, to handle this you’d typically useinheritance, which involves an implicitone-to-one relation).

As with ForeignKey, a recursive relationship can be defined and references to as-yetundefined models can be made.

See also

See the One-to-one relationship model example for a full example.

OneToOneField fields also accept an optionalparent_link argument.

OneToOneField classes used to automatically becomethe primary key on a model. This is no longer true (although you can manuallypass in the primary_key argument if you like).Thus, it’s now possible to have multiple fields of typeOneToOneField on a single model.

Models across files

It’s perfectly OK to relate a model to one from another app. To do this, importthe related model at the top of the file where your model is defined. Then,refer to the other model class wherever needed. For example:

  1. from django.db import models
  2. from geography.models import ZipCode
  3.  
  4. class Restaurant(models.Model):
  5. # ...
  6. zip_code = models.ForeignKey(
  7. ZipCode,
  8. on_delete=models.SET_NULL,
  9. blank=True,
  10. null=True,
  11. )

Field name restrictions

Django places some restrictions on model field names:

  • A field name cannot be a Python reserved word, because that would resultin a Python syntax error. For example:
  1. class Example(models.Model):
  2. pass = models.IntegerField() # 'pass' is a reserved word!
  • A field name cannot contain more than one underscore in a row, due tothe way Django’s query lookup syntax works. For example:
  1. class Example(models.Model):
  2. foo__bar = models.IntegerField() # 'foo__bar' has two underscores!
  • A field name cannot end with an underscore, for similar reasons.

These limitations can be worked around, though, because your field name doesn’tnecessarily have to match your database column name. See thedb_column option.

SQL reserved words, such as join, where or select, are allowed asmodel field names, because Django escapes all database table names and columnnames in every underlying SQL query. It uses the quoting syntax of yourparticular database engine.

Custom field types

If one of the existing model fields cannot be used to fit your purposes, or ifyou wish to take advantage of some less common database column types, you cancreate your own field class. Full coverage of creating your own fields isprovided in Writing custom model fields.

Meta options

Give your model metadata by using an inner class Meta, like so:

  1. from django.db import models
  2.  
  3. class Ox(models.Model):
  4. horn_length = models.IntegerField()
  5.  
  6. class Meta:
  7. ordering = ["horn_length"]
  8. verbose_name_plural = "oxen"

Model metadata is “anything that’s not a field”, such as ordering options(ordering), database table name (db_table), orhuman-readable singular and plural names (verbose_name andverbose_name_plural). None are required, and adding classMeta to a model is completely optional.

A complete list of all possible Meta options can be found in the modeloption reference.

Model attributes

  • objects
  • The most important attribute of a model is theManager. It’s the interface through whichdatabase query operations are provided to Django models and is used toretrieve the instances from the database. If nocustom Manager is defined, the default name isobjects. Managers are only accessible viamodel classes, not the model instances.

Model methods

Define custom methods on a model to add custom “row-level” functionality to yourobjects. Whereas Manager methods are intended to do“table-wide” things, model methods should act on a particular model instance.

This is a valuable technique for keeping business logic in one place – themodel.

For example, this model has a few custom methods:

  1. from django.db import models
  2.  
  3. class Person(models.Model):
  4. first_name = models.CharField(max_length=50)
  5. last_name = models.CharField(max_length=50)
  6. birth_date = models.DateField()
  7.  
  8. def baby_boomer_status(self):
  9. "Returns the person's baby-boomer status."
  10. import datetime
  11. if self.birth_date < datetime.date(1945, 8, 1):
  12. return "Pre-boomer"
  13. elif self.birth_date < datetime.date(1965, 1, 1):
  14. return "Baby boomer"
  15. else:
  16. return "Post-boomer"
  17.  
  18. @property
  19. def full_name(self):
  20. "Returns the person's full name."
  21. return '%s %s' % (self.first_name, self.last_name)

The last method in this example is a property.

The model instance reference has a complete listof methods automatically given to each model.You can override most of these – see overriding predefined model methods,below – but there are a couple that you’ll almost always want to define:

  • str()
  • A Python “magic method” that returns a string representation of anyobject. This is what Python and Django will use whenever a modelinstance needs to be coerced and displayed as a plain string. Mostnotably, this happens when you display an object in an interactiveconsole or in the admin.

You’ll always want to define this method; the default isn’t very helpfulat all.

  • get_absolute_url()
  • This tells Django how to calculate the URL for an object. Django usesthis in its admin interface, and any time it needs to figure out a URLfor an object.

Any object that has a URL that uniquely identifies it should define thismethod.

Overriding predefined model methods

There’s another set of model methods thatencapsulate a bunch of database behavior that you’ll want to customize. Inparticular you’ll often want to change the way save() anddelete() work.

You’re free to override these methods (and any other model method) to alterbehavior.

A classic use-case for overriding the built-in methods is if you want somethingto happen whenever you save an object. For example (seesave() for documentation of the parameters it accepts):

  1. from django.db import models
  2.  
  3. class Blog(models.Model):
  4. name = models.CharField(max_length=100)
  5. tagline = models.TextField()
  6.  
  7. def save(self, *args, **kwargs):
  8. do_something()
  9. super().save(*args, **kwargs) # Call the "real" save() method.
  10. do_something_else()

You can also prevent saving:

  1. from django.db import models
  2.  
  3. class Blog(models.Model):
  4. name = models.CharField(max_length=100)
  5. tagline = models.TextField()
  6.  
  7. def save(self, *args, **kwargs):
  8. if self.name == "Yoko Ono's blog":
  9. return # Yoko shall never have her own blog!
  10. else:
  11. super().save(*args, **kwargs) # Call the "real" save() method.

It’s important to remember to call the superclass method – that’sthat super().save(args, *kwargs) business – to ensurethat the object still gets saved into the database. If you forget tocall the superclass method, the default behavior won’t happen and thedatabase won’t get touched.

It’s also important that you pass through the arguments that can bepassed to the model method – that’s what the args, **kwargs bitdoes. Django will, from time to time, extend the capabilities ofbuilt-in model methods, adding new arguments. If you use args,**kwargs in your method definitions, you are guaranteed that yourcode will automatically support those arguments when they are added.

Overridden model methods are not called on bulk operations

Note that the delete() method for an object is notnecessarily called when deleting objects in bulk using aQuerySet or as a result of a cascadingdelete. To ensure customizeddelete logic gets executed, you can usepre_delete and/orpost_delete signals.

Unfortunately, there isn’t a workaround whencreating orupdating objects in bulk,since none of save(),pre_save, andpost_save are called.

Executing custom SQL

Another common pattern is writing custom SQL statements in model methods andmodule-level methods. For more details on using raw SQL, see the documentationon using raw SQL.

Model inheritance

Model inheritance in Django works almost identically to the way normalclass inheritance works in Python, but the basics at the beginning of the pageshould still be followed. That means the base class should subclassdjango.db.models.Model.

The only decision you have to make is whether you want the parent models to bemodels in their own right (with their own database tables), or if the parentsare just holders of common information that will only be visible through thechild models.

There are three styles of inheritance that are possible in Django.

  • Often, you will just want to use the parent class to hold information thatyou don’t want to have to type out for each child model. This class isn’tgoing to ever be used in isolation, so Abstract base classes arewhat you’re after.
  • If you’re subclassing an existing model (perhaps something from anotherapplication entirely) and want each model to have its own database table,Multi-table inheritance is the way to go.
  • Finally, if you only want to modify the Python-level behavior of a model,without changing the models fields in any way, you can useProxy models.

Abstract base classes

Abstract base classes are useful when you want to put some commoninformation into a number of other models. You write your base classand put abstract=True in the Metaclass. This model will then not be used to create any databasetable. Instead, when it is used as a base class for other models, itsfields will be added to those of the child class.

An example:

  1. from django.db import models
  2.  
  3. class CommonInfo(models.Model):
  4. name = models.CharField(max_length=100)
  5. age = models.PositiveIntegerField()
  6.  
  7. class Meta:
  8. abstract = True
  9.  
  10. class Student(CommonInfo):
  11. home_group = models.CharField(max_length=5)

The Student model will have three fields: name, age andhome_group. The CommonInfo model cannot be used as a normal Djangomodel, since it is an abstract base class. It does not generate a databasetable or have a manager, and cannot be instantiated or saved directly.

Fields inherited from abstract base classes can be overridden with anotherfield or value, or be removed with None.

For many uses, this type of model inheritance will be exactly what you want.It provides a way to factor out common information at the Python level, whilestill only creating one database table per child model at the database level.

Meta inheritance

When an abstract base class is created, Django makes any Metainner class you declared in the base class available as anattribute. If a child class does not declare its own Metaclass, it will inherit the parent’s Meta. If the child wants toextend the parent’s Meta class, it can subclass it. For example:

  1. from django.db import models
  2.  
  3. class CommonInfo(models.Model):
  4. # ...
  5. class Meta:
  6. abstract = True
  7. ordering = ['name']
  8.  
  9. class Student(CommonInfo):
  10. # ...
  11. class Meta(CommonInfo.Meta):
  12. db_table = 'student_info'

Django does make one adjustment to the Meta class of an abstract baseclass: before installing the Meta attribute, it sets abstract=False.This means that children of abstract base classes don’t automatically becomeabstract classes themselves. Of course, you can make an abstract base classthat inherits from another abstract base class. You just need to remember toexplicitly set abstract=True each time.

Some attributes won’t make sense to include in the Meta class of anabstract base class. For example, including db_table would mean that allthe child classes (the ones that don’t specify their own Meta) would usethe same database table, which is almost certainly not what you want.

If you are using related_name orrelated_query_name on a ForeignKey orManyToManyField, you must always specify a unique reverse name and queryname for the field. This would normally cause a problem in abstract baseclasses, since the fields on this class are included into each of the childclasses, with exactly the same values for the attributes (includingrelated_name andrelated_query_name) each time.

To work around this problem, when you are usingrelated_name orrelated_query_name in an abstract baseclass (only), part of the value should contain '%(app_label)s' and'%(class)s'.

  • '%(class)s' is replaced by the lowercased name of the child class thatthe field is used in.
  • '%(app_label)s' is replaced by the lowercased name of the app the childclass is contained within. Each installed application name must be unique andthe model class names within each app must also be unique, therefore theresulting name will end up being different.For example, given an app common/models.py:
  1. from django.db import models
  2.  
  3. class Base(models.Model):
  4. m2m = models.ManyToManyField(
  5. OtherModel,
  6. related_name="%(app_label)s_%(class)s_related",
  7. related_query_name="%(app_label)s_%(class)ss",
  8. )
  9.  
  10. class Meta:
  11. abstract = True
  12.  
  13. class ChildA(Base):
  14. pass
  15.  
  16. class ChildB(Base):
  17. pass

Along with another app rare/models.py:

  1. from common.models import Base
  2.  
  3. class ChildB(Base):
  4. pass

The reverse name of the common.ChildA.m2m field will becommon_childa_related and the reverse query name will be common_childas.The reverse name of the common.ChildB.m2m field will becommon_childb_related and the reverse query name will becommon_childbs. Finally, the reverse name of the rare.ChildB.m2m fieldwill be rare_childb_related and the reverse query name will berare_childbs. It’s up to you how you use the '%(class)s' and'%(app_label)s' portion to construct your related name or related query namebut if you forget to use it, Django will raise errors when you perform systemchecks (or run migrate).

If you don’t specify a related_nameattribute for a field in an abstract base class, the default reverse name willbe the name of the child class followed by '_set', just as it normallywould be if you’d declared the field directly on the child class. For example,in the above code, if the related_nameattribute was omitted, the reverse name for the m2m field would bechilda_set in the ChildA case and childb_set for the ChildBfield.

Multi-table inheritance

The second type of model inheritance supported by Django is when each model inthe hierarchy is a model all by itself. Each model corresponds to its owndatabase table and can be queried and created individually. The inheritancerelationship introduces links between the child model and each of its parents(via an automatically-created OneToOneField).For example:

  1. from django.db import models
  2.  
  3. class Place(models.Model):
  4. name = models.CharField(max_length=50)
  5. address = models.CharField(max_length=80)
  6.  
  7. class Restaurant(Place):
  8. serves_hot_dogs = models.BooleanField(default=False)
  9. serves_pizza = models.BooleanField(default=False)

All of the fields of Place will also be available in Restaurant,although the data will reside in a different database table. So these are bothpossible:

  1. >>> Place.objects.filter(name="Bob's Cafe")
  2. >>> Restaurant.objects.filter(name="Bob's Cafe")

If you have a Place that is also a Restaurant, you can get from thePlace object to the Restaurant object by using the lowercase version ofthe model name:

  1. >>> p = Place.objects.get(id=12)
  2. # If p is a Restaurant object, this will give the child class:
  3. >>> p.restaurant
  4. <Restaurant: ...>

However, if p in the above example was not a Restaurant (it had beencreated directly as a Place object or was the parent of some other class),referring to p.restaurant would raise a Restaurant.DoesNotExistexception.

The automatically-created OneToOneField onRestaurant that links it to Place looks like this:

  1. place_ptr = models.OneToOneField(
  2. Place, on_delete=models.CASCADE,
  3. parent_link=True,
  4. )

You can override that field by declaring your ownOneToOneField with parent_link=True on Restaurant.

Meta and multi-table inheritance

In the multi-table inheritance situation, it doesn’t make sense for a childclass to inherit from its parent’s Meta class. All the Meta optionshave already been applied to the parent class and applying them again wouldnormally only lead to contradictory behavior (this is in contrast with theabstract base class case, where the base class doesn’t exist in its ownright).

So a child model does not have access to its parent’s Meta class. However, there are a few limited cases where the childinherits behavior from the parent: if the child does not specify anordering attribute or aget_latest_by attribute, it will inheritthese from its parent.

If the parent has an ordering and you don’t want the child to have any naturalordering, you can explicitly disable it:

  1. class ChildModel(ParentModel):
  2. # ...
  3. class Meta:
  4. # Remove parent's ordering effect
  5. ordering = []

Inheritance and reverse relations

Because multi-table inheritance uses an implicitOneToOneField to link the child andthe parent, it’s possible to move from the parent down to the child,as in the above example. However, this uses up the name that is thedefault related_name value forForeignKey andManyToManyField relations. If youare putting those types of relations on a subclass of the parent model, youmust specify the related_nameattribute on each such field. If you forget, Django will raise a validationerror.

For example, using the above Place class again, let’s create anothersubclass with a ManyToManyField:

  1. class Supplier(Place):
  2. customers = models.ManyToManyField(Place)

This results in the error:

  1. Reverse query name for 'Supplier.customers' clashes with reverse query
  2. name for 'Supplier.place_ptr'.
  3.  
  4. HINT: Add or change a related_name argument to the definition for
  5. 'Supplier.customers' or 'Supplier.place_ptr'.

Adding related_name to the customers field as follows would resolve theerror: models.ManyToManyField(Place, related_name='provider').

As mentioned, Django will automatically create aOneToOneField linking your childclass back to any non-abstract parent models. If you want to control thename of the attribute linking back to the parent, you can create yourown OneToOneField and setparent_link=Trueto indicate that your field is the link back to the parent class.

Proxy models

When using multi-table inheritance, a newdatabase table is created for each subclass of a model. This is usually thedesired behavior, since the subclass needs a place to store any additionaldata fields that are not present on the base class. Sometimes, however, youonly want to change the Python behavior of a model – perhaps to change thedefault manager, or add a new method.

This is what proxy model inheritance is for: creating a proxy for theoriginal model. You can create, delete and update instances of the proxy modeland all the data will be saved as if you were using the original (non-proxied)model. The difference is that you can change things like the default modelordering or the default manager in the proxy, without having to alter theoriginal.

Proxy models are declared like normal models. You tell Django that it’s aproxy model by setting the proxy attribute ofthe Meta class to True.

For example, suppose you want to add a method to the Person model. You can do it like this:

  1. from django.db import models
  2.  
  3. class Person(models.Model):
  4. first_name = models.CharField(max_length=30)
  5. last_name = models.CharField(max_length=30)
  6.  
  7. class MyPerson(Person):
  8. class Meta:
  9. proxy = True
  10.  
  11. def do_something(self):
  12. # ...
  13. pass

The MyPerson class operates on the same database table as its parentPerson class. In particular, any new instances of Person will also beaccessible through MyPerson, and vice-versa:

  1. >>> p = Person.objects.create(first_name="foobar")
  2. >>> MyPerson.objects.get(first_name="foobar")
  3. <MyPerson: foobar>

You could also use a proxy model to define a different default ordering ona model. You might not always want to order the Person model, but regularlyorder by the last_name attribute when you use the proxy:

  1. class OrderedPerson(Person):
  2. class Meta:
  3. ordering = ["last_name"]
  4. proxy = True

Now normal Person queries will be unorderedand OrderedPerson queries will be ordered by last_name.

Proxy models inherit Meta attributes in the same way as regularmodels.

QuerySets still return the model that was requested

There is no way to have Django return, say, a MyPerson object whenever youquery for Person objects. A queryset for Person objects will returnthose types of objects. The whole point of proxy objects is that code relyingon the original Person will use those and your own code can use theextensions you included (that no other code is relying on anyway). It is nota way to replace the Person (or any other) model everywhere with somethingof your own creation.

Base class restrictions

A proxy model must inherit from exactly one non-abstract model class. Youcan’t inherit from multiple non-abstract models as the proxy model doesn’tprovide any connection between the rows in the different database tables. Aproxy model can inherit from any number of abstract model classes, providingthey do not define any model fields. A proxy model may also inherit from anynumber of proxy models that share a common non-abstract parent class.

Proxy model managers

If you don’t specify any model managers on a proxy model, it inherits themanagers from its model parents. If you define a manager on the proxy model,it will become the default, although any managers defined on the parentclasses will still be available.

Continuing our example from above, you could change the default manager usedwhen you query the Person model like this:

  1. from django.db import models
  2.  
  3. class NewManager(models.Manager):
  4. # ...
  5. pass
  6.  
  7. class MyPerson(Person):
  8. objects = NewManager()
  9.  
  10. class Meta:
  11. proxy = True

If you wanted to add a new manager to the Proxy, without replacing theexisting default, you can use the techniques described in the custommanager documentation: create a base classcontaining the new managers and inherit that after the primary base class:

  1. # Create an abstract class for the new manager.
  2. class ExtraManagers(models.Model):
  3. secondary = NewManager()
  4.  
  5. class Meta:
  6. abstract = True
  7.  
  8. class MyPerson(Person, ExtraManagers):
  9. class Meta:
  10. proxy = True

You probably won’t need to do this very often, but, when you do, it’spossible.

Differences between proxy inheritance and unmanaged models

Proxy model inheritance might look fairly similar to creating an unmanagedmodel, using the managed attribute on amodel’s Meta class.

With careful setting of Meta.db_table you could create an unmanaged model thatshadows an existing model and adds Python methods to it. However, that would bevery repetitive and fragile as you need to keep both copies synchronized if youmake any changes.

On the other hand, proxy models are intended to behave exactly like the modelthey are proxying for. They are always in sync with the parent model since theydirectly inherit its fields and managers.

The general rules are:

  • If you are mirroring an existing model or database table and don’t wantall the original database table columns, use Meta.managed=False.That option is normally useful for modeling database views and tablesnot under the control of Django.
  • If you are wanting to change the Python-only behavior of a model, butkeep all the same fields as in the original, use Meta.proxy=True.This sets things up so that the proxy model is an exact copy of thestorage structure of the original model when data is saved.

Multiple inheritance

Just as with Python’s subclassing, it’s possible for a Django model to inheritfrom multiple parent models. Keep in mind that normal Python name resolutionrules apply. The first base class that a particular name (e.g. Meta) appears in will be the one that is used; for example, thismeans that if multiple parents contain a Meta class,only the first one is going to be used, and all others will be ignored.

Generally, you won’t need to inherit from multiple parents. The main use-casewhere this is useful is for “mix-in” classes: adding a particular extrafield or method to every class that inherits the mix-in. Try to keep yourinheritance hierarchies as simple and straightforward as possible so that youwon’t have to struggle to work out where a particular piece of information iscoming from.

Note that inheriting from multiple models that have a common id primarykey field will raise an error. To properly use multiple inheritance, you canuse an explicit AutoField in the base models:

  1. class Article(models.Model):
  2. article_id = models.AutoField(primary_key=True)
  3. ...
  4.  
  5. class Book(models.Model):
  6. book_id = models.AutoField(primary_key=True)
  7. ...
  8.  
  9. class BookReview(Book, Article):
  10. pass

Or use a common ancestor to hold the AutoField. Thisrequires using an explicit OneToOneField from eachparent model to the common ancestor to avoid a clash between the fields thatare automatically generated and inherited by the child:

  1. class Piece(models.Model):
  2. pass
  3.  
  4. class Article(Piece):
  5. article_piece = models.OneToOneField(Piece, on_delete=models.CASCADE, parent_link=True)
  6. ...
  7.  
  8. class Book(Piece):
  9. book_piece = models.OneToOneField(Piece, on_delete=models.CASCADE, parent_link=True)
  10. ...
  11.  
  12. class BookReview(Book, Article):
  13. pass

Field name “hiding” is not permitted

In normal Python class inheritance, it is permissible for a child class tooverride any attribute from the parent class. In Django, this isn’t usuallypermitted for model fields. If a non-abstract model base class has a fieldcalled author, you can’t create another model field or definean attribute called author in any class that inherits from that base class.

This restriction doesn’t apply to model fields inherited from an abstractmodel. Such fields may be overridden with another field or value, or be removedby setting field_name = None.

Warning

Model managers are inherited from abstract base classes. Overriding aninherited field which is referenced by an inheritedManager may cause subtle bugs. See custommanagers and model inheritance.

Note

Some fields define extra attributes on the model, e.g. aForeignKey defines an extra attribute with_id appended to the field name, as well as related_name andrelated_query_name on the foreign model.

These extra attributes cannot be overridden unless the field that definesit is changed or removed so that it no longer defines the extra attribute.

Overriding fields in a parent model leads to difficulties in areas such asinitializing new instances (specifying which field is being initialized inModel.init) and serialization. These are features which normal Pythonclass inheritance doesn’t have to deal with in quite the same way, so thedifference between Django model inheritance and Python class inheritance isn’tarbitrary.

This restriction only applies to attributes which areField instances. Normal Python attributescan be overridden if you wish. It also only applies to the name of theattribute as Python sees it: if you are manually specifying the databasecolumn name, you can have the same column name appearing in both a child andan ancestor model for multi-table inheritance (they are columns in twodifferent database tables).

Django will raise a FieldError if you overrideany model field in any ancestor model.

Organizing models in a package

The manage.py startapp command creates an applicationstructure that includes a models.py file. If you have many models,organizing them in separate files may be useful.

To do so, create a models package. Remove models.py and create amyapp/models/ directory with an init.py file and the files tostore your models. You must import the models in the init.py file.

For example, if you had organic.py and synthetic.py in the modelsdirectory:

myapp/models/init.py

  1. from .organic import Person
  2. from .synthetic import Robot

Explicitly importing each model rather than using from .models import *has the advantages of not cluttering the namespace, making code more readable,and keeping code analysis tools useful.

See also

  • The Models Reference
  • Covers all the model related APIs including model fields, relatedobjects, and QuerySet.