Django documentation

This document is for Django's SVN release, which can be significantly different from previous releases. Get old docs here: Django 1.0 and is Django’s command-line utility for administrative tasks. This document outlines all it can do.

In addition, is automatically created in each Django project. is a thin wrapper around that takes care of two things for you before delegating to

  • It puts your project’s package on sys.path.
  • It sets the DJANGO_SETTINGS_MODULE environment variable so that it points to your project’s file.

The script should be on your system path if you installed Django via its utility. If it’s not on your path, you can find it in site-packages/django/bin within your Python installation. Consider symlinking it from some place on your path, such as /usr/local/bin.

For Windows users, who do not have symlinking functionality available, you can copy to a location on your existing path or edit the PATH settings (under Settings - Control Panel - System - Advanced - Environment...) to point to its installed location.

Generally, when working on a single Django project, it’s easier to use Use with DJANGO_SETTINGS_MODULE, or the --settings command line option, if you need to switch between multiple Django settings files.

The command-line examples throughout this document use to be consistent, but any example can use just as well.

Usage <subcommand> [options] <subcommand> [options]

subcommand should be one of the subcommands listed in this document. options, which is optional, should be zero or more of the options available for the given subcommand.

Getting runtime help


Run help to display a list of all available subcommands. Run help <subcommand> to display a description of the given subcommand and a list of its available options.

App names

Many subcommands take a list of "app names." An "app name" is the basename of the package containing your models. For example, if your INSTALLED_APPS contains the string '', the app name is blog.

Determining the version


Run --version to display the current Django version.

Examples of output:


Displaying debug output

Use --verbosity to specify the amount of notification and debug information that should print to the console. For more details, see the documentation for the --verbosity option.

Available subcommands

cleanup cleanup
New in Django 1.0: Please, see the release notes

Can be run as a cronjob or directly to clean out old data from the database (only expired sessions at the moment).

compilemessages compilemessages
Changed in Django 1.0: Before 1.0 this was the "bin/" command.

Compiles .po files created with makemessages to .mo files for use with the builtin gettext support. See Internationalization.

Use the --locale` option to specify the locale to process. If not provided, all locales are processed.

Example usage: compilemessages --locale=br_PT

createcachetable createcachetable

Creates a cache table named tablename for use with the database cache backend. See Django's cache framework for more information.

createsuperuser createsuperuser
New in Django 1.0: Please, see the release notes

Creates a superuser account (a user who has all permissions). This is useful if you need to create an initial superuser account but did not do so during syncdb, or if you need to programmatically generate superuser accounts for your site(s).

When run interactively, this command will prompt for a password for the new superuser account. When run non-interactively, no password will be set, and the superuser account will not be able to log in until a password has been manually set for it.


The username and e-mail address for the new account can be supplied by using the --username and --email arguments on the command line. If either of those is not supplied, createsuperuser will prompt for it when running interactively.

This command is only available if Django's authentication system (django.contrib.auth) is installed.

dbshell dbshell

Runs the command-line client for the database engine specified in your DATABASE_ENGINE setting, with the connection parameters specified in your DATABASE_USER, DATABASE_PASSWORD, etc., settings.

  • For PostgreSQL, this runs the psql command-line client.
  • For MySQL, this runs the mysql command-line client.
  • For SQLite, this runs the sqlite3 command-line client.

This command assumes the programs are on your PATH so that a simple call to the program name (psql, mysql, sqlite3) will find the program in the right place. There's no way to specify the location of the program manually.

diffsettings diffsettings

Displays differences between the current settings file and Django's default settings.

Settings that don't appear in the defaults are followed by "###". For example, the default settings don't define ROOT_URLCONF, so ROOT_URLCONF is followed by "###" in the output of diffsettings.

Note that Django's default settings live in django/conf/, if you're ever curious to see the full list of defaults.

dumpdata <appname appname appname.Model ...> dumpdata

Outputs to standard output all data in the database associated with the named application(s).

If no application name is provided, all installed applications will be dumped.

The output of dumpdata can be used as input for loaddata.

Note that dumpdata uses the default manager on the model for selecting the records to dump. If you're using a custom manager as the default manager and it filters some of the available records, not all of the objects will be dumped.

New in Django 1.0: Please, see the release notes

Exclude a specific application from the applications whose contents is output. For example, to specifically exclude the auth application from the output, you would call: dumpdata --exclude=auth

If you want to exclude multiple applications, use multiple --exclude directives: dumpdata --exclude=auth --exclude=contenttypes
--format <fmt>

By default, dumpdata will format its output in JSON, but you can use the --format option to specify another format. Currently supported formats are listed in Serialization formats.

--indent <num>

By default, dumpdata will output all data on a single line. This isn't easy for humans to read, so you can use the --indent option to pretty-print the output with a number of indentation spaces.

New in Django 1.1: Please, see the release notes

In addition to specifying application names, you can provide a list of individual models, in the form of appname.Model. If you specify a model name to dumpdata, the dumped output will be restricted to that model, rather than the entire application. You can also mix application names and model names.

flush flush

Returns the database to the state it was in immediately after syncdb was executed. This means that all data will be removed from the database, any post-synchronization handlers will be re-executed, and the initial_data fixture will be re-installed.

The --noinput option may be provided to suppress all user prompts.

inspectdb inspectdb

Introspects the database tables in the database pointed-to by the DATABASE_NAME setting and outputs a Django model module (a file) to standard output.

Use this if you have a legacy database with which you'd like to use Django. The script will inspect the database and create a model for each table within it.

As you might expect, the created models will have an attribute for every field in the table. Note that inspectdb has a few special cases in its field-name output:

  • If inspectdb cannot map a column's type to a model field type, it'll use TextField and will insert the Python comment 'This field type is a guess.' next to the field in the generated model.
  • If the database column name is a Python reserved word (such as 'pass', 'class' or 'for'), inspectdb will append '_field' to the attribute name. For example, if a table has a column 'for', the generated model will have a field 'for_field', with the db_column attribute set to 'for'. inspectdb will insert the Python comment 'Field renamed because it was a Python reserved word.' next to the field.

This feature is meant as a shortcut, not as definitive model generation. After you run it, you'll want to look over the generated models yourself to make customizations. In particular, you'll need to rearrange models' order, so that models that refer to other models are ordered properly.

Primary keys are automatically introspected for PostgreSQL, MySQL and SQLite, in which case Django puts in the primary_key=True where needed.

inspectdb works with PostgreSQL, MySQL and SQLite. Foreign-key detection only works in PostgreSQL and with certain types of MySQL tables.

loaddata <fixture fixture ...> loaddata

Searches for and loads the contents of the named fixture into the database.

What's a "fixture"?

A fixture is a collection of files that contain the serialized contents of the database. Each fixture has a unique name, and the files that comprise the fixture can be distributed over multiple directories, in multiple applications.

Django will search in three locations for fixtures:

  1. In the fixtures directory of every installed application
  2. In any directory named in the FIXTURE_DIRS setting
  3. In the literal path named by the fixture

Django will load any and all fixtures it finds in these locations that match the provided fixture names.

If the named fixture has a file extension, only fixtures of that type will be loaded. For example: loaddata mydata.json

would only load JSON fixtures called mydata. The fixture extension must correspond to the registered name of a serializer (e.g., json or xml).

If you omit the extensions, Django will search all available fixture types for a matching fixture. For example: loaddata mydata

would look for any fixture of any fixture type called mydata. If a fixture directory contained mydata.json, that fixture would be loaded as a JSON fixture.

The fixtures that are named can include directory components. These directories will be included in the search path. For example: loaddata foo/bar/mydata.json

would search <appname>/fixtures/foo/bar/mydata.json for each installed application, <dirname>/foo/bar/mydata.json for each directory in FIXTURE_DIRS, and the literal path foo/bar/mydata.json.

When fixture files are processed, the data is saved to the database as is. Model defined save methods and pre_save signals are not called.

Note that the order in which fixture files are processed is undefined. However, all fixture data is installed as a single transaction, so data in one fixture can reference data in another fixture. If the database backend supports row-level constraints, these constraints will be checked at the end of the transaction.

The dumpdata command can be used to generate input for loaddata.

Compressed fixtures

Fixtures may be compressed in zip, gz, or bz2 format. For example: loaddata mydata.json

would look for any of mydata.json,, mydata.json.gz, or mydata.json.bz2. The first file contained within a zip-compressed archive is used.

Note that if two fixtures with the same name but different fixture type are discovered (for example, if mydata.json and mydata.xml.gz were found in the same fixture directory), fixture installation will be aborted, and any data installed in the call to loaddata will be removed from the database.

MySQL and Fixtures

Unfortunately, MySQL isn't capable of completely supporting all the features of Django fixtures. If you use MyISAM tables, MySQL doesn't support transactions or constraints, so you won't get a rollback if multiple transaction files are found, or validation of fixture data. If you use InnoDB tables, you won't be able to have any forward references in your data files - MySQL doesn't provide a mechanism to defer checking of row constraints until a transaction is committed.

makemessages makemessages
Changed in Django 1.0: Before 1.0 this was the bin/ command.

Runs over the entire source tree of the current directory and pulls out all strings marked for translation. It creates (or updates) a message file in the conf/locale (in the django tree) or locale (for project and application) directory. After making changes to the messages files you need to compile them with compilemessages for use with the builtin gettext support. See the i18n documentation for details.


Use the --all or -a option to update the message files for all available languages.

Example usage: makemessages --all

Use the --extension or -e option to specify a list of file extensions to examine (default: ".html").

Example usage: makemessages --locale=de --extension xhtml

Separate multiple extensions with commas or use -e or --extension multiple times: makemessages --locale=de --extension=html,txt --extension xml

Use the --locale option to specify the locale to process.

Example usage: makemessages --locale=br_PT

Use the --domain or -d option to change the domain of the messages files. Currently supported:

  • django for all *.py and *.html files (default)
  • djangojs for *.js files

reset <appname appname ...> reset

Executes the equivalent of sqlreset for the given app name(s).

The --noinput option may be provided to suppress all user prompts.

runfcgi [options] runfcgi

Starts a set of FastCGI processes suitable for use with any Web server that supports the FastCGI protocol. See the FastCGI deployment documentation for details. Requires the Python FastCGI module from flup.

runserver [port or ipaddr:port] runserver

Starts a lightweight development Web server on the local machine. By default, the server runs on port 8000 on the IP address You can pass in an IP address and port number explicitly.

If you run this script as a user with normal privileges (recommended), you might not have access to start a port on a low port number. Low port numbers are reserved for the superuser (root).

DO NOT USE THIS SERVER IN A PRODUCTION SETTING. It has not gone through security audits or performance tests. (And that's how it's gonna stay. We're in the business of making Web frameworks, not Web servers, so improving this server to be able to handle a production environment is outside the scope of Django.)

The development server automatically reloads Python code for each request, as needed. You don't need to restart the server for code changes to take effect.

When you start the server, and each time you change Python code while the server is running, the server will validate all of your installed models. (See the validate command below.) If the validator finds errors, it will print them to standard output, but it won't stop the server.

You can run as many servers as you want, as long as they're on separate ports. Just execute runserver more than once.

Note that the default IP address,, is not accessible from other machines on your network. To make your development server viewable to other machines on the network, use its own IP address (e.g. or


Use the --adminmedia option to tell Django where to find the various CSS and JavaScript files for the Django admin interface. Normally, the development server serves these files out of the Django source tree magically, but you'd want to use this if you made any changes to those files for your own site.

Example usage: runserver --adminmedia=/tmp/new-admin-style/

Use the --noreload option to disable the use of the auto-reloader. This means any Python code changes you make while the server is running will not take effect if the particular Python modules have already been loaded into memory.

Example usage: runserver --noreload

Examples of using different ports and addresses

Port 8000 on IP address runserver

Port 8000 on IP address runserver

Port 7000 on IP address runserver 7000

Port 7000 on IP address runserver

Serving static files with the development server

By default, the development server doesn't serve any static files for your site (such as CSS files, images, things under MEDIA_URL and so forth). If you want to configure Django to serve static media, read How to serve static files.

shell shell

Starts the Python interactive interpreter.

Django will use IPython, if it's installed. If you have IPython installed and want to force use of the "plain" Python interpreter, use the --plain option, like so: shell --plain

sql <appname appname ...> sql

Prints the CREATE TABLE SQL statements for the given app name(s).

sqlall <appname appname ...> sqlall

Prints the CREATE TABLE and initial-data SQL statements for the given app name(s).

Refer to the description of sqlcustom for an explanation of how to specify initial data.

sqlclear <appname appname ...> sqlclear

Prints the DROP TABLE SQL statements for the given app name(s).

sqlcustom <appname appname ...> sqlcustom

Prints the custom SQL statements for the given app name(s).

For each model in each specified app, this command looks for the file <appname>/sql/<modelname>.sql, where <appname> is the given app name and <modelname> is the model's name in lowercase. For example, if you have an app news that includes a Story model, sqlcustom will attempt to read a file news/sql/story.sql and append it to the output of this command.

Each of the SQL files, if given, is expected to contain valid SQL. The SQL files are piped directly into the database after all of the models' table-creation statements have been executed. Use this SQL hook to make any table modifications, or insert any SQL functions into the database.

Note that the order in which the SQL files are processed is undefined.

sqlflush sqlflush

Prints the SQL statements that would be executed for the flush command.

sqlindexes <appname appname ...> sqlindexes

Prints the CREATE INDEX SQL statements for the given app name(s).

sqlreset <appname appname ...> sqlreset

Prints the DROP TABLE SQL, then the CREATE TABLE SQL, for the given app name(s).

sqlsequencereset <appname appname ...> sqlsequencereset

Prints the SQL statements for resetting sequences for the given app name(s).

Sequences are indexes used by some database engines to track the next available number for automatically incremented fields.

Use this command to generate SQL which will fix cases where a sequence is out of sync with its automatically incremented field data.

startapp <appname> startapp

Creates a Django app directory structure for the given app name in the current directory.

startproject <projectname> startproject

Creates a Django project directory structure for the given project name in the current directory.

This command is disabled when the --settings option to is used, or when the environment variable DJANGO_SETTINGS_MODULE has been set. To re-enable it in these situations, either omit the --settings option or unset DJANGO_SETTINGS_MODULE.

syncdb syncdb

Creates the database tables for all apps in INSTALLED_APPS whose tables have not already been created.

Use this command when you've added new applications to your project and want to install them in the database. This includes any apps shipped with Django that might be in INSTALLED_APPS by default. When you start a new project, run this command to install the default apps.

Syncdb will not alter existing tables

syncdb will only create tables for models which have not yet been installed. It will never issue ALTER TABLE statements to match changes made to a model class after installation. Changes to model classes and database schemas often involve some form of ambiguity and, in those cases, Django would have to guess at the correct changes to make. There is a risk that critical data would be lost in the process.

If you have made changes to a model and wish to alter the database tables to match, use the sql command to display the new SQL structure and compare that to your existing table schema to work out the changes.

If you're installing the django.contrib.auth application, syncdb will give you the option of creating a superuser immediately.

syncdb will also search for and install any fixture named initial_data with an appropriate extension (e.g. json or xml). See the documentation for loaddata for details on the specification of fixture data files.

The --noinput option may be provided to suppress all user prompts.

test <app or test identifier> test

Runs tests for all installed models. See Testing Django applications for more information.

testserver <fixture fixture ...> testserver
New in Django 1.0: Please, see the release notes

Runs a Django development server (as in runserver) using data from the given fixture(s).

For example, this command: testserver mydata.json

...would perform the following steps:

  1. Create a test database, as described in Testing Django applications.
  2. Populate the test database with fixture data from the given fixtures. (For more on fixtures, see the documentation for loaddata above.)
  3. Runs the Django development server (as in runserver), pointed at this newly created test database instead of your production database.

This is useful in a number of ways:

  • When you're writing unit tests of how your views act with certain fixture data, you can use testserver to interact with the views in a Web browser, manually.
  • Let's say you're developing your Django application and have a "pristine" copy of a database that you'd like to interact with. You can dump your database to a fixture (using the dumpdata command, explained above), then use testserver to run your Web application with that data. With this arrangement, you have the flexibility of messing up your data in any way, knowing that whatever data changes you're making are only being made to a test database.

Note that this server does not automatically detect changes to your Python source code (as runserver does). It does, however, detect changes to templates.

--addrport [port number or ipaddr:port]

Use --addrport to specify a different port, or IP address and port, from the default of This value follows exactly the same format and serves exactly the same function as the argument to the runserver subcommand.


To run the test server on port 7000 with fixture1 and fixture2: testserver --addrport 7000 fixture1 fixture2 testserver fixture1 fixture2 --addrport 7000

(The above statements are equivalent. We include both of them to demonstrate that it doesn't matter whether the options come before or after the fixture arguments.)

To run on with a test fixture: testserver --addrport test

validate validate

Validates all installed models (according to the INSTALLED_APPS setting) and prints validation errors to standard output.

Default options

Although some subcommands may allow their own custom options, every subcommand allows for the following options:


Example usage: syncdb --pythonpath='/home/djangoprojects/myproject'

Adds the given filesystem path to the Python import search path. If this isn't provided, will use the PYTHONPATH environment variable.

Note that this option is unnecessary in, because it takes care of setting the Python path for you.


Example usage: syncdb --settings=mysite.settings

Explicitly specifies the settings module to use. The settings module should be in Python package syntax, e.g. mysite.settings. If this isn't provided, will use the DJANGO_SETTINGS_MODULE environment variable.

Note that this option is unnecessary in, because it uses from the current project by default.


Example usage: syncdb --traceback

By default, will show a simple error message whenever an error occurs. If you specify --traceback, will output a full stack trace whenever an exception is raised.


Example usage: syncdb --verbosity 2

Use --verbosity to specify the amount of notification and debug information that should print to the console.

  • 0 means no output.
  • 1 means normal output (default).
  • 2 means verbose output.

Common options

The following options are not available on every commands, but they are common to a number of commands.


Use the --locale or -l option to specify the locale to process. If not provided all locales are processed.


Use the --noinput option to suppress all user prompting, such as "Are you sure?" confirmation messages. This is useful if is being executed as an unattended, automated script.

Extra niceties

Syntax coloring

The / commands that output SQL to standard output will use pretty color-coded output if your terminal supports ANSI-colored output. It won't use the color codes if you're piping the command's output to another program.

Bash completion

If you use the Bash shell, consider installing the Django bash completion script, which lives in extras/django_bash_completion in the Django distribution. It enables tab-completion of and commands, so you can, for instance...

  • Type
  • Press [TAB] to see all available options.
  • Type sql, then [TAB], to see all available options whose names start with sql.

See Writing custom django-admin commands for how to add customized actions.


Having trouble? We'd like to help!