Model field reference

This document contains all the API references of Field including the field options and field types Django offers.

See also

If the built-in fields don’t do the trick, you can try django-localflavor (documentation), which contains assorted pieces of code that are useful for particular countries and cultures.

Also, you can easily write your own custom model fields.


Technically, these models are defined in django.db.models.fields, but for convenience they’re imported into django.db.models; the standard convention is to use from django.db import models and refer to fields as models.<Foo>Field.

Field options

The following arguments are available to all field types. All are optional.



If True, Django will store empty values as NULL in the database. Default is False.

Avoid using null on string-based fields such as CharField and TextField. If a string-based field has null=True, that means it has two possible values for “no data”: NULL, and the empty string. In most cases, it’s redundant to have two possible values for “no data;” the Django convention is to use the empty string, not NULL. One exception is when a CharField has both unique=True and blank=True set. In this situation, null=True is required to avoid unique constraint violations when saving multiple objects with blank values.

For both string-based and non-string-based fields, you will also need to set blank=True if you wish to permit empty values in forms, as the null parameter only affects database storage (see blank).


When using the Oracle database backend, the value NULL will be stored to denote the empty string regardless of this attribute.



If True, the field is allowed to be blank. Default is False.

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

Supplying missing values

blank=True can be used with fields having null=False, but this will require implementing clean() on the model in order to programmatically supply any missing values.



A mapping or iterable in the format described below to use as choices for this field. If choices are given, they’re enforced by model validation and the default form widget will be a select box with these choices instead of the standard text field.

If a mapping is given, the key element is the actual value to be set on the model, and the second element is the human readable name. For example:

    "FR": "Freshman",
    "SO": "Sophomore",
    "JR": "Junior",
    "SR": "Senior",
    "GR": "Graduate",

You can also pass a sequence consisting itself of iterables of exactly two items (e.g. [(A1, B1), (A2, B2), …]). The first element in each tuple is the actual value to be set on the model, and the second element is the human-readable name. For example:

    ("FR", "Freshman"),
    ("SO", "Sophomore"),
    ("JR", "Junior"),
    ("SR", "Senior"),
    ("GR", "Graduate"),

choices can also be defined as a callable that expects no arguments and returns any of the formats described above. For example:

def get_currencies():
    return {i: i for i in settings.CURRENCIES}

class Expense(models.Model):
    amount = models.DecimalField(max_digits=10, decimal_places=2)
    currency = models.CharField(max_length=3, choices=get_currencies)

Passing a callable for choices can be particularly handy when, for example, the choices are:

  • the result of I/O-bound operations (which could potentially be cached), such as querying a table in the same or an external database, or accessing the choices from a static file.
  • a list that is mostly stable but could vary from time to time or from project to project. Examples in this category are using third-party apps that provide a well-known inventory of values, such as currencies, countries, languages, time zones, etc.
Changed in Django 5.0:

Support for mappings and callables was added.

Generally, it’s best to define choices inside a model class, and to define a suitably-named constant for each value:

from django.db import models

class Student(models.Model):
    JUNIOR = "JR"
    SENIOR = "SR"
        FRESHMAN: "Freshman",
        SOPHOMORE: "Sophomore",
        JUNIOR: "Junior",
        SENIOR: "Senior",
        GRADUATE: "Graduate",
    year_in_school = models.CharField(

    def is_upperclass(self):
        return self.year_in_school in {self.JUNIOR, self.SENIOR}

Though you can define a choices list outside of a model class and then refer to it, defining the choices and names for each choice inside the model class keeps all of that information with the class that uses it, and helps reference the choices (e.g, Student.SOPHOMORE will work anywhere that the Student model has been imported).

You can also collect your available choices into named groups that can be used for organizational purposes:

    "Audio": {
        "vinyl": "Vinyl",
        "cd": "CD",
    "Video": {
        "vhs": "VHS Tape",
        "dvd": "DVD",
    "unknown": "Unknown",

The key of the mapping is the name to apply to the group and the value is the choices inside that group, consisting of the field value and a human-readable name for an option. Grouped options may be combined with ungrouped options within a single mapping (such as the "unknown" option in this example).

You can also use a sequence, e.g. a list of 2-tuples:

            ("vinyl", "Vinyl"),
            ("cd", "CD"),
            ("vhs", "VHS Tape"),
            ("dvd", "DVD"),
    ("unknown", "Unknown"),

Note that choices can be any sequence object – not necessarily a list or tuple. This lets you construct choices dynamically. But if you find yourself hacking choices to be dynamic, you’re probably better off using a proper database table with a ForeignKey. choices is meant for static data that doesn’t change much, if ever.


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

For each model field that has choices set, Django will normalize the choices to a list of 2-tuples and add a method to retrieve the human-readable name for the field’s current value. See get_FOO_display() in the database API documentation.

Unless blank=False is set on the field along with a default then a label containing "---------" will be rendered with the select box. To override this behavior, add a tuple to choices containing None; e.g. (None, 'Your String For Display'). Alternatively, you can use an empty string instead of None where this makes sense - such as on a CharField.

Enumeration types

In addition, Django provides enumeration types that you can subclass to define choices in a concise way:

from django.utils.translation import gettext_lazy as _

class Student(models.Model):
    class YearInSchool(models.TextChoices):
        FRESHMAN = "FR", _("Freshman")
        SOPHOMORE = "SO", _("Sophomore")
        JUNIOR = "JR", _("Junior")
        SENIOR = "SR", _("Senior")
        GRADUATE = "GR", _("Graduate")

    year_in_school = models.CharField(

    def is_upperclass(self):
        return self.year_in_school in {

These work similar to enum from Python’s standard library, but with some modifications:

  • Enum member values are a tuple of arguments to use when constructing the concrete data type. Django supports adding an extra string value to the end of this tuple to be used as the human-readable name, or label. The label can be a lazy translatable string. Thus, in most cases, the member value will be a (value, label) 2-tuple. See below for an example of subclassing choices using a more complex data type. If a tuple is not provided, or the last item is not a (lazy) string, the label is automatically generated from the member name.

  • A .label property is added on values, to return the human-readable name.

  • A number of custom properties are added to the enumeration classes – .choices, .labels, .values, and .names – to make it easier to access lists of those separate parts of the enumeration.


    These property names cannot be used as member names as they would conflict.

  • The use of enum.unique() is enforced to ensure that values cannot be defined multiple times. This is unlikely to be expected in choices for a field.

Note that using YearInSchool.SENIOR, YearInSchool['SENIOR'], or YearInSchool('SR') to access or lookup enum members work as expected, as do the .name and .value properties on the members.

If you don’t need to have the human-readable names translated, you can have them inferred from the member name (replacing underscores with spaces and using title-case):

>>> class Vehicle(models.TextChoices):
...     CAR = "C"
...     TRUCK = "T"
...     JET_SKI = "J"
>>> Vehicle.JET_SKI.label
'Jet Ski'

Since the case where the enum values need to be integers is extremely common, Django provides an IntegerChoices class. For example:

class Card(models.Model):
    class Suit(models.IntegerChoices):
        DIAMOND = 1
        SPADE = 2
        HEART = 3
        CLUB = 4

    suit = models.IntegerField(choices=Suit)

It is also possible to make use of the Enum Functional API with the caveat that labels are automatically generated as highlighted above:

>>> MedalType = models.TextChoices("MedalType", "GOLD SILVER BRONZE")
>>> MedalType.choices
[('GOLD', 'Gold'), ('SILVER', 'Silver'), ('BRONZE', 'Bronze')]
>>> Place = models.IntegerChoices("Place", "FIRST SECOND THIRD")
>>> Place.choices
[(1, 'First'), (2, 'Second'), (3, 'Third')]

If you require support for a concrete data type other than int or str, you can subclass Choices and the required concrete data type, e.g. date for use with DateField:

class MoonLandings(, models.Choices):
    APOLLO_11 = 1969, 7, 20, "Apollo 11 (Eagle)"
    APOLLO_12 = 1969, 11, 19, "Apollo 12 (Intrepid)"
    APOLLO_14 = 1971, 2, 5, "Apollo 14 (Antares)"
    APOLLO_15 = 1971, 7, 30, "Apollo 15 (Falcon)"
    APOLLO_16 = 1972, 4, 21, "Apollo 16 (Orion)"
    APOLLO_17 = 1972, 12, 11, "Apollo 17 (Challenger)"

There are some additional caveats to be aware of:

  • Enumeration types do not support named groups.

  • Because an enumeration with a concrete data type requires all values to match the type, overriding the blank label cannot be achieved by creating a member with a value of None. Instead, set the __empty__ attribute on the class:

    class Answer(models.IntegerChoices):
        NO = 0, _("No")
        YES = 1, _("Yes")
        __empty__ = _("(Unknown)")
Changed in Django 5.0:

Support for using enumeration types directly in the choices was added.



The name of the database column to use for this field. If this isn’t given, Django will use the field’s name.

If your database column name is an SQL reserved word, or contains characters that aren’t allowed in Python variable names – notably, the hyphen – that’s OK. Django quotes column and table names behind the scenes.


New in Django 4.2.

The comment on the database column to use for this field. It is useful for documenting fields for individuals with direct database access who may not be looking at your Django code. For example:

pub_date = models.DateTimeField(
    db_comment="Date and time when the article was published",


New in Django 5.0.

The database-computed default value for this field. This can be a literal value or a database function, such as Now:

created = models.DateTimeField(db_default=Now())

More complex expressions can be used, as long as they are made from literals and database functions:

month_due = models.DateField(
        Now() + timedelta(days=90),

Database defaults cannot reference other fields or models. For example, this is invalid:

end = models.IntegerField(db_default=F("start") + 50)

If both db_default and Field.default are set, default will take precedence when creating instances in Python code. db_default will still be set at the database level and will be used when inserting rows outside of the ORM or when adding a new field in a migration.



If True, a database index will be created for this field.

Use the indexes option instead.

Where possible, use the Meta.indexes option instead. In nearly all cases, indexes provides more functionality than db_index. db_index may be deprecated in the future.



The name of the database tablespace to use for this field’s index, if this field is indexed. The default is the project’s DEFAULT_INDEX_TABLESPACE setting, if set, or the db_tablespace of the model, if any. If the backend doesn’t support tablespaces for indexes, this option is ignored.



The default value for the field. This can be a value or a callable object. If callable it will be called every time a new object is created.

The default can’t be a mutable object (model instance, list, set, etc.), as a reference to the same instance of that object would be used as the default value in all new model instances. Instead, wrap the desired default in a callable. For example, if you want to specify a default dict for JSONField, use a function:

def contact_default():
    return {"email": ""}

contact_info = JSONField("ContactInfo", default=contact_default)

lambdas can’t be used for field options like default because they can’t be serialized by migrations. See that documentation for other caveats.

For fields like ForeignKey that map to model instances, defaults should be the value of the field they reference (pk unless to_field is set) instead of model instances.

The default value is used when new model instances are created and a value isn’t provided for the field. When the field is a primary key, the default is also used when the field is set to None.

The default value can also be set at the database level with Field.db_default.



If False, the field will not be displayed in the admin or any other ModelForm. They are also skipped during model validation. Default is True.



The error_messages argument lets you override the default messages that the field will raise. Pass in a dictionary with keys matching the error messages you want to override.

Error message keys include null, blank, invalid, invalid_choice, unique, and unique_for_date. Additional error message keys are specified for each field in the Field types section below.

These error messages often don’t propagate to forms. See Considerations regarding model’s error_messages.



Extra “help” text to be displayed with the form widget. It’s useful for documentation even if your field isn’t used on a form.

Note that this value is not HTML-escaped in automatically-generated forms. This lets you include HTML in help_text if you so desire. For example:

help_text = "Please use the following format: <em>YYYY-MM-DD</em>."

Alternatively you can use plain text and django.utils.html.escape() to escape any HTML special characters. Ensure that you escape any help text that may come from untrusted users to avoid a cross-site scripting attack.



If True, this field is the primary key for the model.

If you don’t specify primary_key=True for any field in your model, Django will automatically add a field to hold the primary key, so you don’t need to set primary_key=True on any of your fields unless you want to override the default primary-key behavior. The type of auto-created primary key fields can be specified per app in AppConfig.default_auto_field or globally in the DEFAULT_AUTO_FIELD setting. For more, see Automatic primary key fields.

primary_key=True implies null=False and unique=True. Only one primary key is allowed on an object.

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

The primary key field is set to None when deleting an object.



If True, this field must be unique throughout the table.

This is enforced at the database level and by model validation. If you try to save a model with a duplicate value in a unique field, a django.db.IntegrityError will be raised by the model’s save() method.

This option is valid on all field types except ManyToManyField and OneToOneField.

Note that when unique is True, you don’t need to specify db_index, because unique implies the creation of an index.



Set this to the name of a DateField or DateTimeField to require that this field be unique for the value of the date field.

For example, if you have a field title that has unique_for_date="pub_date", then Django wouldn’t allow the entry of two records with the same title and pub_date.

Note that if you set this to point to a DateTimeField, only the date portion of the field will be considered. Besides, when USE_TZ is True, the check will be performed in the current time zone at the time the object gets saved.

This is enforced by Model.validate_unique() during model validation but not at the database level. If any unique_for_date constraint involves fields that are not part of a ModelForm (for example, if one of the fields is listed in exclude or has editable=False), Model.validate_unique() will skip validation for that particular constraint.



Like unique_for_date, but requires the field to be unique with respect to the month.



Like unique_for_date and unique_for_month.



A human-readable name for the field. If the verbose name isn’t given, Django will automatically create it using the field’s attribute name, converting underscores to spaces. See Verbose field names.



A list of validators to run for this field. See the validators documentation for more information.

Field types


class AutoField(**options)

An IntegerField that automatically increments according to available IDs. You usually won’t need to use this directly; a primary key field will automatically be added to your model if you don’t specify otherwise. See Automatic primary key fields.


class BigAutoField(**options)

A 64-bit integer, much like an AutoField except that it is guaranteed to fit numbers from 1 to 9223372036854775807.


class BigIntegerField(**options)

A 64-bit integer, much like an IntegerField except that it is guaranteed to fit numbers from -9223372036854775808 to 9223372036854775807. The default form widget for this field is a NumberInput.


class BinaryField(max_length=None, **options)

A field to store raw binary data. It can be assigned bytes, bytearray, or memoryview.

By default, BinaryField sets editable to False, in which case it can’t be included in a ModelForm.


Optional. The maximum length (in bytes) of the field. The maximum length is enforced in Django’s validation using MaxLengthValidator.

Abusing BinaryField

Although you might think about storing files in the database, consider that it is bad design in 99% of the cases. This field is not a replacement for proper static files handling.


class BooleanField(**options)

A true/false field.

The default form widget for this field is CheckboxInput, or NullBooleanSelect if null=True.

The default value of BooleanField is None when Field.default isn’t defined.


class CharField(max_length=None, **options)

A string field, for small- to large-sized strings.

For large amounts of text, use TextField.

The default form widget for this field is a TextInput.

CharField has the following extra arguments:


The maximum length (in characters) of the field. The max_length is enforced at the database level and in Django’s validation using MaxLengthValidator. It’s required for all database backends included with Django except PostgreSQL, which supports unlimited VARCHAR columns.


If you are writing an application that must be portable to multiple database backends, you should be aware that there are restrictions on max_length for some backends. Refer to the database backend notes for details.

Changed in Django 4.2:

Support for unlimited VARCHAR columns was added on PostgreSQL.


Optional. The database collation name of the field.


Collation names are not standardized. As such, this will not be portable across multiple database backends.


Oracle supports collations only when the MAX_STRING_SIZE database initialization parameter is set to EXTENDED.


class DateField(auto_now=False, auto_now_add=False, **options)

A date, represented in Python by a instance. Has a few extra, optional arguments:


Automatically set the field to now every time the object is saved. Useful for “last-modified” timestamps. Note that the current date is always used; it’s not just a default value that you can override.

The field is only automatically updated when calling The field isn’t updated when making updates to other fields in other ways such as QuerySet.update(), though you can specify a custom value for the field in an update like that.


Automatically set the field to now when the object is first created. Useful for creation of timestamps. Note that the current date is always used; it’s not just a default value that you can override. So even if you set a value for this field when creating the object, it will be ignored. If you want to be able to modify this field, set the following instead of auto_now_add=True:

The default form widget for this field is a DateInput. The admin adds a JavaScript calendar, and a shortcut for “Today”. Includes an additional invalid_date error message key.

The options auto_now_add, auto_now, and default are mutually exclusive. Any combination of these options will result in an error.


As currently implemented, setting auto_now or auto_now_add to True will cause the field to have editable=False and blank=True set.


The auto_now and auto_now_add options will always use the date in the default timezone at the moment of creation or update. If you need something different, you may want to consider using your own callable default or overriding save() instead of using auto_now or auto_now_add; or using a DateTimeField instead of a DateField and deciding how to handle the conversion from datetime to date at display time.


class DateTimeField(auto_now=False, auto_now_add=False, **options)

A date and time, represented in Python by a datetime.datetime instance. Takes the same extra arguments as DateField.

The default form widget for this field is a single DateTimeInput. The admin uses two separate TextInput widgets with JavaScript shortcuts.


class DecimalField(max_digits=None, decimal_places=None, **options)

A fixed-precision decimal number, represented in Python by a Decimal instance. It validates the input using DecimalValidator.

Has the following required arguments:


The maximum number of digits allowed in the number. Note that this number must be greater than or equal to decimal_places.


The number of decimal places to store with the number.

For example, to store numbers up to 999.99 with a resolution of 2 decimal places, you’d use:

models.DecimalField(..., max_digits=5, decimal_places=2)

And to store numbers up to approximately one billion with a resolution of 10 decimal places:

models.DecimalField(..., max_digits=19, decimal_places=10)

The default form widget for this field is a NumberInput when localize is False or TextInput otherwise.


For more information about the differences between the FloatField and DecimalField classes, please see FloatField vs. DecimalField. You should also be aware of SQLite limitations of decimal fields.


class DurationField(**options)

A field for storing periods of time - modeled in Python by timedelta. When used on PostgreSQL, the data type used is an interval and on Oracle the data type is INTERVAL DAY(9) TO SECOND(6). Otherwise a bigint of microseconds is used.


Arithmetic with DurationField works in most cases. However on all databases other than PostgreSQL, comparing the value of a DurationField to arithmetic on DateTimeField instances will not work as expected.


class EmailField(max_length=254, **options)

A CharField that checks that the value is a valid email address using EmailValidator.


class FileField(upload_to='', storage=None, max_length=100, **options)

A file-upload field.


The primary_key argument isn’t supported and will raise an error if used.

Has the following optional arguments:


This attribute provides a way of setting the upload directory and file name, and can be set in two ways. In both cases, the value is passed to the method.

If you specify a string value or a Path, it may contain strftime() formatting, which will be replaced by the date/time of the file upload (so that uploaded files don’t fill up the given directory). For example:

class MyModel(models.Model):
    # file will be uploaded to MEDIA_ROOT/uploads
    upload = models.FileField(upload_to="uploads/")
    # or...
    # file will be saved to MEDIA_ROOT/uploads/2015/01/30
    upload = models.FileField(upload_to="uploads/%Y/%m/%d/")

If you are using the default FileSystemStorage, the string value will be appended to your MEDIA_ROOT path to form the location on the local filesystem where uploaded files will be stored. If you are using a different storage, check that storage’s documentation to see how it handles upload_to.

upload_to may also be a callable, such as a function. This will be called to obtain the upload path, including the filename. This callable must accept two arguments and return a Unix-style path (with forward slashes) to be passed along to the storage system. The two arguments are:

Argument Description

An instance of the model where the FileField is defined. More specifically, this is the particular instance where the current file is being attached.

In most cases, this object will not have been saved to the database yet, so if it uses the default AutoField, it might not yet have a value for its primary key field.

filename The filename that was originally given to the file. This may or may not be taken into account when determining the final destination path.

For example:

def user_directory_path(instance, filename):
    # file will be uploaded to MEDIA_ROOT/user_<id>/<filename>
    return "user_{0}/{1}".format(, filename)

class MyModel(models.Model):
    upload = models.FileField(upload_to=user_directory_path)

A storage object, or a callable which returns a storage object. This handles the storage and retrieval of your files. See Managing files for details on how to provide this object.

The default form widget for this field is a ClearableFileInput.

Using a FileField or an ImageField (see below) in a model takes a few steps:

  1. In your settings file, you’ll need to define MEDIA_ROOT as the full path to a directory where you’d like Django to store uploaded files. (For performance, these files are not stored in the database.) Define MEDIA_URL as the base public URL of that directory. Make sure that this directory is writable by the web server’s user account.
  2. Add the FileField or ImageField to your model, defining the upload_to option to specify a subdirectory of MEDIA_ROOT to use for uploaded files.
  3. All that will be stored in your database is a path to the file (relative to MEDIA_ROOT). You’ll most likely want to use the convenience url attribute provided by Django. For example, if your ImageField is called mug_shot, you can get the absolute path to your image in a template with {{ object.mug_shot.url }}.

For example, say your MEDIA_ROOT is set to '/home/media', and upload_to is set to 'photos/%Y/%m/%d'. The '%Y/%m/%d' part of upload_to is strftime() formatting; '%Y' is the four-digit year, '%m' is the two-digit month and '%d' is the two-digit day. If you upload a file on Jan. 15, 2007, it will be saved in the directory /home/media/photos/2007/01/15.

If you wanted to retrieve the uploaded file’s on-disk filename, or the file’s size, you could use the name and size attributes respectively; for more information on the available attributes and methods, see the File class reference and the Managing files topic guide.


The file is saved as part of saving the model in the database, so the actual file name used on disk cannot be relied on until after the model has been saved.

The uploaded file’s relative URL can be obtained using the url attribute. Internally, this calls the url() method of the underlying Storage class.

Note that whenever you deal with uploaded files, you should pay close attention to where you’re uploading them and what type of files they are, to avoid security holes. Validate all uploaded files so that you’re sure the files are what you think they are. For example, if you blindly let somebody upload files, without validation, to a directory that’s within your web server’s document root, then somebody could upload a CGI or PHP script and execute that script by visiting its URL on your site. Don’t allow that.

Also note that even an uploaded HTML file, since it can be executed by the browser (though not by the server), can pose security threats that are equivalent to XSS or CSRF attacks.

FileField instances are created in your database as varchar columns with a default max length of 100 characters. As with other fields, you can change the maximum length using the max_length argument.

FileField and FieldFile

class FieldFile

When you access a FileField on a model, you are given an instance of FieldFile as a proxy for accessing the underlying file.

The API of FieldFile mirrors that of File, with one key difference: The object wrapped by the class is not necessarily a wrapper around Python’s built-in file object. Instead, it is a wrapper around the result of the method, which may be a File object, or it may be a custom storage’s implementation of the File API.

In addition to the API inherited from File such as read() and write(), FieldFile includes several methods that can be used to interact with the underlying file:


Two methods of this class, save() and delete(), default to saving the model object of the associated FieldFile in the database.

The name of the file including the relative path from the root of the Storage of the associated FileField.


A read-only property to access the file’s local filesystem path by calling the path() method of the underlying Storage class.


The result of the underlying Storage.size() method.


A read-only property to access the file’s relative URL by calling the url() method of the underlying Storage class.'rb')

Opens or reopens the file associated with this instance in the specified mode. Unlike the standard Python open() method, it doesn’t return a file descriptor.

Since the underlying file is opened implicitly when accessing it, it may be unnecessary to call this method except to reset the pointer to the underlying file or to change the mode.


Behaves like the standard Python file.close() method and closes the file associated with this instance., content, save=True)

This method takes a filename and file contents and passes them to the storage class for the field, then associates the stored file with the model field. If you want to manually associate file data with FileField instances on your model, the save() method is used to persist that file data.

Takes two required arguments: name which is the name of the file, and content which is an object containing the file’s contents. The optional save argument controls whether or not the model instance is saved after the file associated with this field has been altered. Defaults to True.

Note that the content argument should be an instance of django.core.files.File, not Python’s built-in file object. You can construct a File from an existing Python file object like this:

from django.core.files import File

# Open an existing file using Python's built-in open()
f = open("/path/to/")
myfile = File(f)

Or you can construct one from a Python string like this:

from django.core.files.base import ContentFile

myfile = ContentFile("hello world")

For more information, see Managing files.


Deletes the file associated with this instance and clears all attributes on the field. Note: This method will close the file if it happens to be open when delete() is called.

The optional save argument controls whether or not the model instance is saved after the file associated with this field has been deleted. Defaults to True.

Note that when a model is deleted, related files are not deleted. If you need to cleanup orphaned files, you’ll need to handle it yourself (for instance, with a custom management command that can be run manually or scheduled to run periodically via e.g. cron).


class FilePathField(path='', match=None, recursive=False, allow_files=True, allow_folders=False, max_length=100, **options)

A CharField whose choices are limited to the filenames in a certain directory on the filesystem. Has some special arguments, of which the first is required:


Required. The absolute filesystem path to a directory from which this FilePathField should get its choices. Example: "/home/images".

path may also be a callable, such as a function to dynamically set the path at runtime. Example:

import os
from django.conf import settings
from django.db import models

def images_path():
    return os.path.join(settings.LOCAL_FILE_DIR, "images")

class MyModel(models.Model):
    file = models.FilePathField(path=images_path)

Optional. A regular expression, as a string, that FilePathField will use to filter filenames. Note that the regex will be applied to the base filename, not the full path. Example: "foo.*\.txt$", which will match a file called foo23.txt but not bar.txt or foo23.png.


Optional. Either True or False. Default is False. Specifies whether all subdirectories of path should be included


Optional. Either True or False. Default is True. Specifies whether files in the specified location should be included. Either this or allow_folders must be True.


Optional. Either True or False. Default is False. Specifies whether folders in the specified location should be included. Either this or allow_files must be True.

The one potential gotcha is that match applies to the base filename, not the full path. So, this example:

FilePathField(path="/home/images", match="foo.*", recursive=True)

…will match /home/images/foo.png but not /home/images/foo/bar.png because the match applies to the base filename (foo.png and bar.png).

FilePathField instances are created in your database as varchar columns with a default max length of 100 characters. As with other fields, you can change the maximum length using the max_length argument.


class FloatField(**options)

A floating-point number represented in Python by a float instance.

The default form widget for this field is a NumberInput when localize is False or TextInput otherwise.

FloatField vs. DecimalField

The FloatField class is sometimes mixed up with the DecimalField class. Although they both represent real numbers, they represent those numbers differently. FloatField uses Python’s float type internally, while DecimalField uses Python’s Decimal type. For information on the difference between the two, see Python’s documentation for the decimal module.


New in Django 5.0.
class GeneratedField(expression, output_field, db_persist=None, **kwargs)

A field that is always computed based on other fields in the model. This field is managed and updated by the database itself. Uses the GENERATED ALWAYS SQL syntax.

There are two kinds of generated columns: stored and virtual. A stored generated column is computed when it is written (inserted or updated) and occupies storage as if it were a regular column. A virtual generated column occupies no storage and is computed when it is read. Thus, a virtual generated column is similar to a view and a stored generated column is similar to a materialized view.


An Expression used by the database to automatically set the field value each time the model is changed.

The expressions should be deterministic and only reference fields within the model (in the same database table). Generated fields cannot reference other generated fields. Database backends can impose further restrictions.


A model field instance to define the field’s data type.


Determines if the database column should occupy storage as if it were a real column. If False, the column acts as a virtual column and does not occupy database storage space.

PostgreSQL only supports persisted columns. Oracle only supports virtual columns.

Refresh the data

Since the database always computed the value, the object must be reloaded to access the new value after save(), for example, by using refresh_from_db().

Database limitations

There are many database-specific restrictions on generated fields that Django doesn’t validate and the database may raise an error e.g. PostgreSQL requires functions and operators referenced in a generated column to be marked as IMMUTABLE.

You should always check that expression is supported on your database. Check out MariaDB, MySQL, Oracle, PostgreSQL, or SQLite docs.


class GenericIPAddressField(protocol='both', unpack_ipv4=False, **options)

An IPv4 or IPv6 address, in string format (e.g. or 2a02:42fe::4). The default form widget for this field is a TextInput.

The IPv6 address normalization follows RFC 4291#section-2.2 section 2.2, including using the IPv4 format suggested in paragraph 3 of that section, like ::ffff: For example, 2001:0::0:01 would be normalized to 2001::1, and ::ffff:0a0a:0a0a to ::ffff: All characters are converted to lowercase.


Limits valid inputs to the specified protocol. Accepted values are 'both' (default), 'IPv4' or 'IPv6'. Matching is case insensitive.


Unpacks IPv4 mapped addresses like ::ffff: If this option is enabled that address would be unpacked to Default is disabled. Can only be used when protocol is set to 'both'.

If you allow for blank values, you have to allow for null values since blank values are stored as null.


class ImageField(upload_to=None, height_field=None, width_field=None, max_length=100, **options)

Inherits all attributes and methods from FileField, but also validates that the uploaded object is a valid image.

In addition to the special attributes that are available for FileField, an ImageField also has height and width attributes.

To facilitate querying on those attributes, ImageField has the following optional arguments:


Name of a model field which will be auto-populated with the height of the image each time the model instance is saved.


Name of a model field which will be auto-populated with the width of the image each time the model instance is saved.

Requires the Pillow library.

ImageField instances are created in your database as varchar columns with a default max length of 100 characters. As with other fields, you can change the maximum length using the max_length argument.

The default form widget for this field is a ClearableFileInput.


class IntegerField(**options)

An integer. Values from -2147483648 to 2147483647 are safe in all databases supported by Django.

It uses MinValueValidator and MaxValueValidator to validate the input based on the values that the default database supports.

The default form widget for this field is a NumberInput when localize is False or TextInput otherwise.


class JSONField(encoder=None, decoder=None, **options)

A field for storing JSON encoded data. In Python the data is represented in its Python native format: dictionaries, lists, strings, numbers, booleans and None.

JSONField is supported on MariaDB, MySQL, Oracle, PostgreSQL, and SQLite (with the JSON1 extension enabled).


An optional json.JSONEncoder subclass to serialize data types not supported by the standard JSON serializer (e.g. datetime.datetime or UUID). For example, you can use the DjangoJSONEncoder class.

Defaults to json.JSONEncoder.


An optional json.JSONDecoder subclass to deserialize the value retrieved from the database. The value will be in the format chosen by the custom encoder (most often a string). Your deserialization may need to account for the fact that you can’t be certain of the input type. For example, you run the risk of returning a datetime that was actually a string that just happened to be in the same format chosen for datetimes.

Defaults to json.JSONDecoder.

To query JSONField in the database, see Querying JSONField.

Default value

If you give the field a default, ensure it’s a callable such as the dict class or a function that returns a fresh object each time. Incorrectly using a mutable object like default={} or default=[] creates a mutable default that is shared between all instances.


Index and Field.db_index both create a B-tree index, which isn’t particularly helpful when querying JSONField. On PostgreSQL only, you can use GinIndex that is better suited.

PostgreSQL users

PostgreSQL has two native JSON based data types: json and jsonb. The main difference between them is how they are stored and how they can be queried. PostgreSQL’s json field is stored as the original string representation of the JSON and must be decoded on the fly when queried based on keys. The jsonb field is stored based on the actual structure of the JSON which allows indexing. The trade-off is a small additional cost on writing to the jsonb field. JSONField uses jsonb.

Oracle users

Oracle Database does not support storing JSON scalar values. Only JSON objects and arrays (represented in Python using dict and list) are supported.


class PositiveBigIntegerField(**options)

Like a PositiveIntegerField, but only allows values under a certain (database-dependent) point. Values from 0 to 9223372036854775807 are safe in all databases supported by Django.


class PositiveIntegerField(**options)

Like an IntegerField, but must be either positive or zero (0). Values from 0 to 2147483647 are safe in all databases supported by Django. The value 0 is accepted for backward compatibility reasons.


class PositiveSmallIntegerField(**options)

Like a PositiveIntegerField, but only allows values under a certain (database-dependent) point. Values from 0 to 32767 are safe in all databases supported by Django.


class SlugField(max_length=50, **options)

Slug is a newspaper term. A slug is a short label for something, containing only letters, numbers, underscores or hyphens. They’re generally used in URLs.

Like a CharField, you can specify max_length (read the note about database portability and max_length in that section, too). If max_length is not specified, Django will use a default length of 50.

Implies setting Field.db_index to True.

It is often useful to automatically prepopulate a SlugField based on the value of some other value. You can do this automatically in the admin using prepopulated_fields.

It uses validate_slug or validate_unicode_slug for validation.


If True, the field accepts Unicode letters in addition to ASCII letters. Defaults to False.


class SmallAutoField(**options)

Like an AutoField, but only allows values under a certain (database-dependent) limit. Values from 1 to 32767 are safe in all databases supported by Django.


class SmallIntegerField(**options)

Like an IntegerField, but only allows values under a certain (database-dependent) point. Values from -32768 to 32767 are safe in all databases supported by Django.


class TextField(**options)

A large text field. The default form widget for this field is a Textarea.

If you specify a max_length attribute, it will be reflected in the Textarea widget of the auto-generated form field. However it is not enforced at the model or database level. Use a CharField for that.


Optional. The database collation name of the field.


Collation names are not standardized. As such, this will not be portable across multiple database backends.


Oracle does not support collations for a TextField.


class TimeField(auto_now=False, auto_now_add=False, **options)

A time, represented in Python by a datetime.time instance. Accepts the same auto-population options as DateField.

The default form widget for this field is a TimeInput. The admin adds some JavaScript shortcuts.


class URLField(max_length=200, **options)

A CharField for a URL, validated by URLValidator.

The default form widget for this field is a URLInput.

Like all CharField subclasses, URLField takes the optional max_length argument. If you don’t specify max_length, a default of 200 is used.


class UUIDField(**options)

A field for storing universally unique identifiers. Uses Python’s UUID class. When used on PostgreSQL and MariaDB 10.7+, this stores in a uuid datatype, otherwise in a char(32).

Universally unique identifiers are a good alternative to AutoField for primary_key. The database will not generate the UUID for you, so it is recommended to use default:

import uuid
from django.db import models

class MyUUIDModel(models.Model):
    id = models.UUIDField(primary_key=True, default=uuid.uuid4, editable=False)
    # other fields

Note that a callable (with the parentheses omitted) is passed to default, not an instance of UUID.

Lookups on PostgreSQL and MariaDB 10.7+

Using iexact, contains, icontains, startswith, istartswith, endswith, or iendswith lookups on PostgreSQL don’t work for values without hyphens, because PostgreSQL and MariaDB 10.7+ store them in a hyphenated uuid datatype type.

Relationship fields

Django also defines a set of fields that represent relations.


class ForeignKey(to, on_delete, **options)

A many-to-one relationship. Requires two positional arguments: the class to which the model is related and the on_delete option.

To create a recursive relationship – an object that has a many-to-one relationship with itself – use models.ForeignKey('self', on_delete=models.CASCADE).

If you need to create a relationship on a model that has not yet been defined, you can use the name of the model, rather than the model object itself:

from django.db import models

class Car(models.Model):
    manufacturer = models.ForeignKey(
    # ...

class Manufacturer(models.Model):
    # ...

Relationships defined this way on abstract models are resolved when the model is subclassed as a concrete model and are not relative to the abstract model’s app_label:

from django.db import models

class AbstractCar(models.Model):
    manufacturer = models.ForeignKey("Manufacturer", on_delete=models.CASCADE)

    class Meta:
        abstract = True
from django.db import models
from products.models import AbstractCar

class Manufacturer(models.Model):

class Car(AbstractCar):

# Car.manufacturer will point to `production.Manufacturer` here.

To refer to models defined in another application, you can explicitly specify a model with the full application label. For example, if the Manufacturer model above is defined in another application called production, you’d need to use:

class Car(models.Model):
    manufacturer = models.ForeignKey(

This sort of reference, called a lazy relationship, can be useful when resolving circular import dependencies between two applications.

A database index is automatically created on the ForeignKey. You can disable this by setting db_index to False. You may want to avoid the overhead of an index if you are creating a foreign key for consistency rather than joins, or if you will be creating an alternative index like a partial or multiple column index.

Database Representation

Behind the scenes, Django appends "_id" to the field name to create its database column name. In the above example, the database table for the Car model will have a manufacturer_id column. (You can change this explicitly by specifying db_column) However, your code should never have to deal with the database column name, unless you write custom SQL. You’ll always deal with the field names of your model object.


ForeignKey accepts other arguments that define the details of how the relation works.


When an object referenced by a ForeignKey is deleted, Django will emulate the behavior of the SQL constraint specified by the on_delete argument. For example, if you have a nullable ForeignKey and you want it to be set null when the referenced object is deleted:

user = models.ForeignKey(

on_delete doesn’t create an SQL constraint in the database. Support for database-level cascade options may be implemented later.

The possible values for on_delete are found in django.db.models:


    Cascade deletes. Django emulates the behavior of the SQL constraint ON DELETE CASCADE and also deletes the object containing the ForeignKey.

    Model.delete() isn’t called on related models, but the pre_delete and post_delete signals are sent for all deleted objects.


    Prevent deletion of the referenced object by raising ProtectedError, a subclass of django.db.IntegrityError.


    Prevent deletion of the referenced object by raising RestrictedError (a subclass of django.db.IntegrityError). Unlike PROTECT, deletion of the referenced object is allowed if it also references a different object that is being deleted in the same operation, but via a CASCADE relationship.

    Consider this set of models:

    class Artist(models.Model):
        name = models.CharField(max_length=10)
    class Album(models.Model):
        artist = models.ForeignKey(Artist, on_delete=models.CASCADE)
    class Song(models.Model):
        artist = models.ForeignKey(Artist, on_delete=models.CASCADE)
        album = models.ForeignKey(Album, on_delete=models.RESTRICT)

    Artist can be deleted even if that implies deleting an Album which is referenced by a Song, because Song also references Artist itself through a cascading relationship. For example:

    >>> artist_one = Artist.objects.create(name="artist one")
    >>> artist_two = Artist.objects.create(name="artist two")
    >>> album_one = Album.objects.create(artist=artist_one)
    >>> album_two = Album.objects.create(artist=artist_two)
    >>> song_one = Song.objects.create(artist=artist_one, album=album_one)
    >>> song_two = Song.objects.create(artist=artist_one, album=album_two)
    >>> album_one.delete()
    # Raises RestrictedError.
    >>> artist_two.delete()
    # Raises RestrictedError.
    >>> artist_one.delete()
    (4, {'Song': 2, 'Album': 1, 'Artist': 1})

    Set the ForeignKey null; this is only possible if null is True.


    Set the ForeignKey to its default value; a default for the ForeignKey must be set.

  • SET()

    Set the ForeignKey to the value passed to SET(), or if a callable is passed in, the result of calling it. In most cases, passing a callable will be necessary to avoid executing queries at the time your is imported:

    from django.conf import settings
    from django.contrib.auth import get_user_model
    from django.db import models
    def get_sentinel_user():
        return get_user_model().objects.get_or_create(username="deleted")[0]
    class MyModel(models.Model):
        user = models.ForeignKey(

    Take no action. If your database backend enforces referential integrity, this will cause an IntegrityError unless you manually add an SQL ON DELETE constraint to the database field.


Sets a limit to the available choices for this field when this field is rendered using a ModelForm or the admin (by default, all objects in the queryset are available to choose). Either a dictionary, a Q object, or a callable returning a dictionary or Q object can be used.

For example:

staff_member = models.ForeignKey(
    limit_choices_to={"is_staff": True},

causes the corresponding field on the ModelForm to list only Users that have is_staff=True. This may be helpful in the Django admin.

The callable form can be helpful, for instance, when used in conjunction with the Python datetime module to limit selections by date range. For example:

def limit_pub_date_choices():
    return {"pub_date__lte":}

limit_choices_to = limit_pub_date_choices

If limit_choices_to is or returns a Q object, which is useful for complex queries, then it will only have an effect on the choices available in the admin when the field is not listed in raw_id_fields in the ModelAdmin for the model.


If a callable is used for limit_choices_to, it will be invoked every time a new form is instantiated. It may also be invoked when a model is validated, for example by management commands or the admin. The admin constructs querysets to validate its form inputs in various edge cases multiple times, so there is a possibility your callable may be invoked several times.


The name to use for the relation from the related object back to this one. It’s also the default value for related_query_name (the name to use for the reverse filter name from the target model). See the related objects documentation for a full explanation and example. Note that you must set this value when defining relations on abstract models; and when you do so some special syntax is available.

If you’d prefer Django not to create a backwards relation, set related_name to '+' or end it with '+'. For example, this will ensure that the User model won’t have a backwards relation to this model:

user = models.ForeignKey(

The name to use for the reverse filter name from the target model. It defaults to the value of related_name or default_related_name if set, otherwise it defaults to the name of the model:

# Declare the ForeignKey with related_query_name
class Tag(models.Model):
    article = models.ForeignKey(
    name = models.CharField(max_length=255)

# That's now the name of the reverse filter

Like related_name, related_query_name supports app label and class interpolation via some special syntax.


The field on the related object that the relation is to. By default, Django uses the primary key of the related object. If you reference a different field, that field must have unique=True.


Controls whether or not a constraint should be created in the database for this foreign key. The default is True, and that’s almost certainly what you want; setting this to False can be very bad for data integrity. That said, here are some scenarios where you might want to do this:

  • You have legacy data that is not valid.
  • You’re sharding your database.

If this is set to False, accessing a related object that doesn’t exist will raise its DoesNotExist exception.


Controls the migration framework’s reaction if this ForeignKey is pointing at a swappable model. If it is True - the default - then if the ForeignKey is pointing at a model which matches the current value of settings.AUTH_USER_MODEL (or another swappable model setting) the relationship will be stored in the migration using a reference to the setting, not to the model directly.

You only want to override this to be False if you are sure your model should always point toward the swapped-in model - for example, if it is a profile model designed specifically for your custom user model.

Setting it to False does not mean you can reference a swappable model even if it is swapped out - False means that the migrations made with this ForeignKey will always reference the exact model you specify (so it will fail hard if the user tries to run with a User model you don’t support, for example).

If in doubt, leave it to its default of True.


class ManyToManyField(to, **options)

A many-to-many relationship. Requires a positional argument: the class to which the model is related, which works exactly the same as it does for ForeignKey, including recursive and lazy relationships.

Related objects can be added, removed, or created with the field’s RelatedManager.

Database Representation

Behind the scenes, Django creates an intermediary join table to represent the many-to-many relationship. By default, this table name is generated using the name of the many-to-many field and the name of the table for the model that contains it. Since some databases don’t support table names above a certain length, these table names will be automatically truncated and a uniqueness hash will be used, e.g. author_books_9cdf. You can manually provide the name of the join table using the db_table option.


ManyToManyField accepts an extra set of arguments – all optional – that control how the relationship functions.


Same as ForeignKey.related_name.


Same as ForeignKey.related_query_name.


Same as ForeignKey.limit_choices_to.


Only used in the definition of ManyToManyFields on self. Consider the following model:

from django.db import models

class Person(models.Model):
    friends = models.ManyToManyField("self")

When Django processes this model, it identifies that it has a ManyToManyField on itself, and as a result, it doesn’t add a person_set attribute to the Person class. Instead, the ManyToManyField is assumed to be symmetrical – that is, if I am your friend, then you are my friend.

If you do not want symmetry in many-to-many relationships with self, set symmetrical to False. This will force Django to add the descriptor for the reverse relationship, allowing ManyToManyField relationships to be non-symmetrical.


Django will automatically generate a table to manage many-to-many relationships. However, if you want to manually specify the intermediary table, you can use the through option to specify the Django model that represents the intermediate table that you want to use.

The most common use for this option is when you want to associate extra data with a many-to-many relationship.


If you don’t want multiple associations between the same instances, add a UniqueConstraint including the from and to fields. Django’s automatically generated many-to-many tables include such a constraint.


Recursive relationships using an intermediary model can’t determine the reverse accessors names, as they would be the same. You need to set a related_name to at least one of them. If you’d prefer Django not to create a backwards relation, set related_name to '+'.

If you don’t specify an explicit through model, there is still an implicit through model class you can use to directly access the table created to hold the association. It has three fields to link the models.

If the source and target models differ, the following fields are generated:

  • id: the primary key of the relation.
  • <containing_model>_id: the id of the model that declares the ManyToManyField.
  • <other_model>_id: the id of the model that the ManyToManyField points to.

If the ManyToManyField points from and to the same model, the following fields are generated:

  • id: the primary key of the relation.
  • from_<model>_id: the id of the instance which points at the model (i.e. the source instance).
  • to_<model>_id: the id of the instance to which the relationship points (i.e. the target model instance).

This class can be used to query associated records for a given model instance like a normal model:


Only used when a custom intermediary model is specified. Django will normally determine which fields of the intermediary model to use in order to establish a many-to-many relationship automatically. However, consider the following models:

from django.db import models

class Person(models.Model):
    name = models.CharField(max_length=50)

class Group(models.Model):
    name = models.CharField(max_length=128)
    members = models.ManyToManyField(
        through_fields=("group", "person"),

class Membership(models.Model):
    group = models.ForeignKey(Group, on_delete=models.CASCADE)
    person = models.ForeignKey(Person, on_delete=models.CASCADE)
    inviter = models.ForeignKey(
    invite_reason = models.CharField(max_length=64)

Membership has two foreign keys to Person (person and inviter), which makes the relationship ambiguous and Django can’t know which one to use. In this case, you must explicitly specify which foreign keys Django should use using through_fields, as in the example above.

through_fields accepts a 2-tuple ('field1', 'field2'), where field1 is the name of the foreign key to the model the ManyToManyField is defined on (group in this case), and field2 the name of the foreign key to the target model (person in this case).

When you have more than one foreign key on an intermediary model to any (or even both) of the models participating in a many-to-many relationship, you must specify through_fields. This also applies to recursive relationships when an intermediary model is used and there are more than two foreign keys to the model, or you want to explicitly specify which two Django should use.


The name of the table to create for storing the many-to-many data. If this is not provided, Django will assume a default name based upon the names of: the table for the model defining the relationship and the name of the field itself.


Controls whether or not constraints should be created in the database for the foreign keys in the intermediary table. The default is True, and that’s almost certainly what you want; setting this to False can be very bad for data integrity. That said, here are some scenarios where you might want to do this:

  • You have legacy data that is not valid.
  • You’re sharding your database.

It is an error to pass both db_constraint and through.


Controls the migration framework’s reaction if this ManyToManyField is pointing at a swappable model. If it is True - the default - then if the ManyToManyField is pointing at a model which matches the current value of settings.AUTH_USER_MODEL (or another swappable model setting) the relationship will be stored in the migration using a reference to the setting, not to the model directly.

You only want to override this to be False if you are sure your model should always point toward the swapped-in model - for example, if it is a profile model designed specifically for your custom user model.

If in doubt, leave it to its default of True.

ManyToManyField does not support validators.

null has no effect since there is no way to require a relationship at the database level.


class OneToOneField(to, on_delete, parent_link=False, **options)

A one-to-one relationship. Conceptually, this is similar to a ForeignKey with unique=True, but the “reverse” side of the relation will directly return a single object.

This is most useful as the primary key of a model which “extends” another model in some way; Multi-table inheritance is implemented by adding an implicit one-to-one relation from the child model to the parent model, for example.

One positional argument is required: the class to which the model will be related. This works exactly the same as it does for ForeignKey, including all the options regarding recursive and lazy relationships.

If you do not specify the related_name argument for the OneToOneField, Django will use the lowercase name of the current model as default value.

With the following example:

from django.conf import settings
from django.db import models

class MySpecialUser(models.Model):
    user = models.OneToOneField(
    supervisor = models.OneToOneField(

your resulting User model will have the following attributes:

>>> user = User.objects.get(pk=1)
>>> hasattr(user, "myspecialuser")
>>> hasattr(user, "supervisor_of")

A RelatedObjectDoesNotExist exception is raised when accessing the reverse relationship if an entry in the related table doesn’t exist. This is a subclass of the target model’s Model.DoesNotExist exception and can be accessed as an attribute of the reverse accessor. For example, if a user doesn’t have a supervisor designated by MySpecialUser:

except User.supervisor_of.RelatedObjectDoesNotExist:

Additionally, OneToOneField accepts all of the extra arguments accepted by ForeignKey, plus one extra argument:

When True and used in a model which inherits from another concrete model, indicates that this field should be used as the link back to the parent class, rather than the extra OneToOneField which would normally be implicitly created by subclassing.

See One-to-one relationships for usage examples of OneToOneField.

Field API reference

class Field

Field is an abstract class that represents a database table column. Django uses fields to create the database table (db_type()), to map Python types to database (get_prep_value()) and vice-versa (from_db_value()).

A field is thus a fundamental piece in different Django APIs, notably, models and querysets.

In models, a field is instantiated as a class attribute and represents a particular table column, see Models. It has attributes such as null and unique, and methods that Django uses to map the field value to database-specific values.

A Field is a subclass of RegisterLookupMixin and thus both Transform and Lookup can be registered on it to be used in QuerySets (e.g. field_name__exact="foo"). All built-in lookups are registered by default.

All of Django’s built-in fields, such as CharField, are particular implementations of Field. If you need a custom field, you can either subclass any of the built-in fields or write a Field from scratch. In either case, see How to create custom model fields.


A verbose description of the field, e.g. for the django.contrib.admindocs application.

The description can be of the form:

description = _("String (up to %(max_length)s)")

where the arguments are interpolated from the field’s __dict__.


A class implementing the descriptor protocol that is instantiated and assigned to the model instance attribute. The constructor must accept a single argument, the Field instance. Overriding this class attribute allows for customizing the get and set behavior.

To map a Field to a database-specific type, Django exposes several methods:


Returns a string naming this field for backend specific purposes. By default, it returns the class name.

See Emulating built-in field types for usage in custom fields.


Returns the database column data type for the Field, taking into account the connection.

See Custom database types for usage in custom fields.


Returns the database column data type for fields such as ForeignKey and OneToOneField that point to the Field, taking into account the connection.

See Custom database types for usage in custom fields.

There are three main situations where Django needs to interact with the database backend and fields:

  • when it queries the database (Python value -> database backend value)
  • when it loads data from the database (database backend value -> Python value)
  • when it saves to the database (Python value -> database backend value)

When querying, get_db_prep_value() and get_prep_value() are used:


value is the current value of the model’s attribute, and the method should return data in a format that has been prepared for use as a parameter in a query.

See Converting Python objects to query values for usage.

get_db_prep_value(value, connection, prepared=False)

Converts value to a backend-specific value. By default it returns value if prepared=True and get_prep_value() if is False.

See Converting query values to database values for usage.

When loading data, from_db_value() is used:

from_db_value(value, expression, connection)

Converts a value as returned by the database to a Python object. It is the reverse of get_prep_value().

This method is not used for most built-in fields as the database backend already returns the correct Python type, or the backend itself does the conversion.

expression is the same as self.

See Converting values to Python objects for usage.


For performance reasons, from_db_value is not implemented as a no-op on fields which do not require it (all Django fields). Consequently you may not call super in your definition.

When saving, pre_save() and get_db_prep_save() are used:

get_db_prep_save(value, connection)

Same as the get_db_prep_value(), but called when the field value must be saved to the database. By default returns get_db_prep_value().

pre_save(model_instance, add)

Method called prior to get_db_prep_save() to prepare the value before being saved (e.g. for DateField.auto_now).

model_instance is the instance this field belongs to and add is whether the instance is being saved to the database for the first time.

It should return the value of the appropriate attribute from model_instance for this field. The attribute name is in self.attname (this is set up by Field).

See Preprocessing values before saving for usage.

Fields often receive their values as a different type, either from serialization or from forms.


Converts the value into the correct Python object. It acts as the reverse of value_to_string(), and is also called in clean().

See Converting values to Python objects for usage.

Besides saving to the database, the field also needs to know how to serialize its value:


Returns the field’s value for the given model instance.

This method is often used by value_to_string().


Converts obj to a string. Used to serialize the value of the field.

See Converting field data for serialization for usage.

When using model forms, the Field needs to know which form field it should be represented by:

formfield(form_class=None, choices_form_class=None, **kwargs)

Returns the default django.forms.Field of this field for ModelForm.

By default, if both form_class and choices_form_class are None, it uses CharField. If the field has choices and choices_form_class isn’t specified, it uses TypedChoiceField.

See Specifying the form field for a model field for usage.


Returns a 4-tuple with enough information to recreate the field:

  1. The name of the field on the model.
  2. The import path of the field (e.g. "django.db.models.IntegerField"). This should be the most portable version, so less specific may be better.
  3. A list of positional arguments.
  4. A dict of keyword arguments.

This method must be added to fields prior to 1.7 to migrate its data using Migrations.

Registering and fetching lookups

Field implements the lookup registration API. The API can be used to customize which lookups are available for a field class and its instances, and how lookups are fetched from a field.

Changed in Django 4.2:

Support for registering lookups on Field instances was added.

Field attribute reference

Every Field instance contains several attributes that allow introspecting its behavior. Use these attributes instead of isinstance checks when you need to write code that depends on a field’s functionality. These attributes can be used together with the Model._meta API to narrow down a search for specific field types. Custom model fields should implement these flags.

Attributes for fields


Boolean flag that indicates if the field was automatically created, such as the OneToOneField used by model inheritance.


Boolean flag that indicates if the field has a database column associated with it.


Boolean flag that indicates if a field is used to back another non-hidden field’s functionality (e.g. the content_type and object_id fields that make up a GenericForeignKey). The hidden flag is used to distinguish what constitutes the public subset of fields on the model from all the fields on the model.


Options.get_fields() excludes hidden fields by default. Pass in include_hidden=True to return hidden fields in the results.


Boolean flag that indicates if a field contains references to one or more other models for its functionality (e.g. ForeignKey, ManyToManyField, OneToOneField, etc.).


Returns the model on which the field is defined. If a field is defined on a superclass of a model, model will refer to the superclass, not the class of the instance.

Attributes for fields with relations

These attributes are used to query for the cardinality and other details of a relation. These attribute are present on all fields; however, they will only have boolean values (rather than None) if the field is a relation type (Field.is_relation=True).


Boolean flag that is True if the field has a many-to-many relation; False otherwise. The only field included with Django where this is True is ManyToManyField.


Boolean flag that is True if the field has a many-to-one relation, such as a ForeignKey; False otherwise.


Boolean flag that is True if the field has a one-to-many relation, such as a GenericRelation or the reverse of a ForeignKey; False otherwise.


Boolean flag that is True if the field has a one-to-one relation, such as a OneToOneField; False otherwise.


Points to the model the field relates to. For example, Author in ForeignKey(Author, on_delete=models.CASCADE). The related_model for a GenericForeignKey is always None.