django-admin and manage.pyLink to this heading#django-admin is Django’s command-line utility for administrative tasks.\nThis document outlines all it can do.
In addition, manage.py is automatically created in each Django project. It\ndoes the same thing as django-admin but also sets the\nDJANGO_SETTINGS_MODULE environment variable so that it points to your\nproject’s settings.py file.
The django-admin script should be on your system path if you installed\nDjango via pip. If it’s not in your path, ensure you have your virtual\nenvironment activated.
Generally, when working on a single Django project, it’s easier to use\nmanage.py than django-admin. If you need to switch between multiple\nDjango settings files, use django-admin with\nDJANGO_SETTINGS_MODULE or the --settings command line\noption.
The command-line examples throughout this document use django-admin to\nbe consistent, but any example can use manage.py or python -m django\njust as well.
Linux / macOS
$ django-admin <command> [options]\n$ manage.py <command> [options]\n$ python -m django <command> [options]\nWindows
...\\> django-admin <command> [options]\n...\\> manage.py <command> [options]\n...\\> py -m django <command> [options]\ncommand should be one of the commands listed in this document.\noptions, which is optional, should be zero or more of the options available\nfor the given command.
Run django-admin help to display usage information and a list of the\ncommands provided by each application.
Run django-admin help --commands to display a list of all available\ncommands.
Run django-admin help <command> to display a description of the given\ncommand and a list of its available options.
Many commands take a list of «app names.» An «app name» is the basename of the\npackage containing your models. For example, if your INSTALLED_APPS\ncontains the string 'mysite.blog', the app name is blog.
Run django-admin version to display the current Django version.
The output follows the schema described in PEP 440:
\n1.4.dev17026\n1.4a1\n1.4\nUse --verbosity, where it is supported, to specify the amount of\nnotification and debug information that django-admin prints to the console.
checkLink to this heading#Uses the system check framework to inspect the entire\nDjango project for common problems.
\nBy default, all apps will be checked. You can check a subset of apps by\nproviding a list of app labels as arguments:
\ndjango-admin check auth admin myapp\nThe system check framework performs many different types of checks that are\ncategorized with tags. You can use these\ntags to restrict the checks performed to just those in a particular category.\nFor example, to perform only models and compatibility checks, run:
\ndjango-admin check --tag models --tag compatibility\nSpecifies the database to run checks requiring database access:
\ndjango-admin check --database default --database other\nBy default, these checks will not be run.
\nLists all available tags.
\nActivates some additional checks that are only relevant in a deployment\nsetting.
\nYou can use this option in your local development environment, but since your\nlocal development settings module may not have many of your production\nsettings, you will probably want to point the check command at a different\nsettings module, either by setting the DJANGO_SETTINGS_MODULE\nenvironment variable, or by passing the --settings option:
django-admin check --deploy --settings=production_settings\nOr you could run it directly on a production or staging deployment to verify\nthat the correct settings are in use (omitting --settings). You could even\nmake it part of your integration test suite.
Specifies the message level that will cause the command to exit with a non-zero\nstatus. Default is ERROR.
compilemessagesLink to this heading#Compiles .po files created by makemessages to .mo files for\nuse with the built-in gettext support. See Internacionalización y localización.
Specifies the locale(s) to process. If not provided, all locales are processed.
\nSpecifies the locale(s) to exclude from processing. If not provided, no locales\nare excluded.
\nIncludes fuzzy translations into compiled files.
\nExample usage:
\ndjango-admin compilemessages --locale=pt_BR\ndjango-admin compilemessages --locale=pt_BR --locale=fr -f\ndjango-admin compilemessages -l pt_BR\ndjango-admin compilemessages -l pt_BR -l fr --use-fuzzy\ndjango-admin compilemessages --exclude=pt_BR\ndjango-admin compilemessages --exclude=pt_BR --exclude=fr\ndjango-admin compilemessages -x pt_BR\ndjango-admin compilemessages -x pt_BR -x fr\nIgnores directories matching the given glob-style pattern. Use\nmultiple times to ignore more.
Example usage:
\ndjango-admin compilemessages --ignore=cache --ignore=outdated/*/locale\ncreatecachetableLink to this heading#Creates the cache tables for use with the database cache backend using the\ninformation from your settings file. See Django’s cache framework for more\ninformation.
\nSpecifies the database in which the cache table(s) will be created. Defaults to\ndefault.
Prints the SQL that would be run without actually running it, so you can\ncustomize it or use the migrations framework.
\ndbshellLink to this heading#Runs the command-line client for the database engine specified in your\nENGINE setting, with the connection parameters\nspecified in your USER, 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.
For Oracle, this runs the sqlplus command-line client.
This command assumes the programs are on your PATH so that a call to\nthe program name (psql, mysql, sqlite3, sqlplus) will find the\nprogram in the right place. There’s no way to specify the location of the\nprogram manually.
Specifies the database onto which to open a shell. Defaults to default.
Any arguments following a -- divider will be passed on to the underlying\ncommand-line client. For example, with PostgreSQL you can use the psql\ncommand’s -c flag to execute a raw SQL query directly:
Linux / macOS
$ django-admin dbshell -- -c 'select current_user'\n current_user\n--------------\n postgres\n(1 row)\nWindows
...\\> django-admin dbshell -- -c 'select current_user'\n current_user\n--------------\n postgres\n(1 row)\nOn MySQL/MariaDB, you can do this with the mysql command’s -e flag:
Linux / macOS
$ django-admin dbshell -- -e "select user()"\n+----------------------+\n| user() |\n+----------------------+\n| djangonaut@localhost |\n+----------------------+\nWindows
...\\> django-admin dbshell -- -e "select user()"\n+----------------------+\n| user() |\n+----------------------+\n| djangonaut@localhost |\n+----------------------+\ndiffsettingsLink to this heading#Displays differences between the current settings file and Django’s default\nsettings (or another settings file specified by --default).
Settings that don’t appear in the defaults are followed by "###". For\nexample, the default settings don’t define ROOT_URLCONF, so\nROOT_URLCONF is followed by "###" in the output of\ndiffsettings.
Displays all settings, even if they have Django’s default value. Such settings\nare prefixed by "###".
The settings module to compare the current settings against. Leave empty to\ncompare against Django’s default settings.
\nSpecifies the output format. Available values are hash and unified.\nhash is the default mode that displays the output that’s described above.\nunified displays the output similar to diff -u. Default settings are\nprefixed with a minus sign, followed by the changed setting prefixed with a\nplus sign.
dumpdataLink to this heading#Outputs to standard output all data in the database associated with the named\napplication(s).
\nIf no application name is provided, all installed applications will be dumped.
\nThe output of dumpdata can be used as input for loaddata.
When result of dumpdata is saved as a file, it can serve as a\nfixture for\ntests or as an\ninitial data.
Note that dumpdata uses the default manager on the model for selecting the\nrecords to dump. If you’re using a custom manager as\nthe default manager and it filters some of the available records, not all of\nthe objects will be dumped.
Uses Django’s base manager, dumping records which might otherwise be filtered\nor modified by a custom manager.
\nSpecifies the serialization format of the output. Defaults to JSON. Supported\nformats are listed in Serialization formats.
\nSpecifies the number of indentation spaces to use in the output. Defaults to\nNone which displays all data on single line.
Prevents specific applications or models (specified in the form of\napp_label.ModelName) from being dumped. If you specify a model name, then\nonly that model will be excluded, rather than the entire application. You can\nalso mix application names and model names.
If you want to exclude multiple applications, pass --exclude more than\nonce:
django-admin dumpdata --exclude=auth --exclude=contenttypes\nSpecifies the database from which data will be dumped. Defaults to default.
Uses the natural_key() model method to serialize any foreign key and\nmany-to-many relationship to objects of the type that defines the method. If\nyou’re dumping contrib.auth Permission objects or\ncontrib.contenttypes ContentType objects, you should probably use this\nflag. See the natural keys\ndocumentation for more details on this and the next option.
Omits the primary key in the serialized data of this object since it can be\ncalculated during deserialization.
\nOutputs only the objects specified by a comma separated list of primary keys.\nThis is only available when dumping one model. By default, all the records of\nthe model are output.
\nSpecifies a file to write the serialized data to. By default, the data goes to\nstandard output.
\nWhen this option is set and --verbosity is greater than 0 (the default), a\nprogress bar is shown in the terminal.
The output file can be compressed with one of the bz2, gz, lzma, or\nxz formats by ending the filename with the corresponding extension.\nFor example, to output the data as a compressed JSON file:
django-admin dumpdata -o mydata.json.gz\nflushLink to this heading#Removes all data from the database and re-executes any post-synchronization\nhandlers. The table of which migrations have been applied is not cleared.
\nIf you would rather start from an empty database and rerun all migrations, you\nshould drop and recreate the database and then run migrate instead.
Suppresses all user prompts.
\nSpecifies the database to flush. Defaults to default.
inspectdbLink to this heading#Introspects the database tables in the database pointed-to by the\nNAME setting and outputs a Django model module (a models.py\nfile) to standard output.
You may choose what tables or views to inspect by passing their names as\narguments. If no arguments are provided, models are created for views only if\nthe --include-views option is used. Models for partition tables are\ncreated on PostgreSQL if the --include-partitions option is used.
Use this if you have a legacy database with which you’d like to use Django.\nThe script will inspect the database and create a model for each table within\nit.
\nAs you might expect, the created models will have an attribute for every field\nin the table. Note that inspectdb has a few special cases in its field-name\noutput:
If inspectdb cannot map a column’s type to a model field type, it’ll\nuse TextField and will insert the Python comment\n'This field type is a guess.' next to the field in the generated\nmodel. The recognized fields may depend on apps listed in\nINSTALLED_APPS. For example, django.contrib.postgres adds\nrecognition for several PostgreSQL-specific field types.
If the database column name is a Python reserved word (such as\n'pass', 'class' or 'for'), inspectdb will append\n'_field' to the attribute name. For example, if a table has a column\n'for', the generated model will have a field 'for_field', with\nthe db_column attribute set to 'for'. inspectdb will insert\nthe Python comment\n'Field renamed because it was a Python reserved word.' next to the\nfield.
This feature is meant as a shortcut, not as definitive model generation. After\nyou run it, you’ll want to look over the generated models yourself to make\ncustomizations. In particular, you’ll need to rearrange models” order, so that\nmodels that refer to other models are ordered properly.
\nDjango doesn’t create database defaults when a\ndefault is specified on a model field.\nSimilarly, database defaults aren’t translated to model field defaults or\ndetected in any fashion by inspectdb.
By default, inspectdb creates unmanaged models. That is, managed =\nFalse in the model’s Meta class tells Django not to manage each table’s\ncreation, modification, and deletion. If you do want to allow Django to manage\nthe table’s lifecycle, you’ll need to change the\nmanaged option to True (or remove it\nbecause True is its default value).
Models are created for materialized views if --include-views is\nused.
Models are created for foreign tables.
Models are created for materialized views if\n--include-views is used.
Models are created for partition tables if\n--include-partitions is used.
Specifies the database to introspect. Defaults to default.
If this option is provided, models are also created for partitions.
\nOnly support for PostgreSQL is implemented.
\nIf this option is provided, models are also created for database views.
\nloaddataLink to this heading#Searches for and loads the contents of the named\nfixture into the database.
\nSpecifies the database into which the data will be loaded. Defaults to\ndefault.
Ignores fields and models that may have been removed since the fixture was\noriginally generated.
\nSpecifies a single app to look for fixtures in rather than looking in all apps.
\nSpecifies the serialization format (e.g.,\njson or xml) for fixtures read from stdin.
Excludes loading the fixtures from the given applications and/or models (in the\nform of app_label or app_label.ModelName). Use the option multiple\ntimes to exclude more than one app or model.
stdinLink to this heading#You can use a dash as the fixture name to load input from sys.stdin. For\nexample:
django-admin loaddata --format=json -\nWhen reading from stdin, the --format option\nis required to specify the serialization format\nof the input (e.g., json or xml).
Loading from stdin is useful with standard input and output redirections.\nFor example:
django-admin dumpdata --format=json --database=test app_label.ModelName | django-admin loaddata --format=json --database=prod -\nThe dumpdata command can be used to generate input for loaddata.
makemessagesLink to this heading#Runs over the entire source tree of the current directory and pulls out all\nstrings marked for translation. It creates (or updates) a message file in the\nconf/locale (in the Django tree) or locale (for project and application)\ndirectory. After making changes to the messages files you need to compile them\nwith compilemessages for use with the builtin gettext support. See\nthe i18n documentation for details.
This command doesn’t require configured settings. However, when settings aren’t\nconfigured, the command can’t ignore the MEDIA_ROOT and\nSTATIC_ROOT directories or include LOCALE_PATHS.
Updates the message files for all available languages.
\nSpecifies a list of file extensions to examine (default: html, txt,\npy or js if --domain is djangojs).
Example usage:
\ndjango-admin makemessages --locale=de --extension xhtml\nSeparate multiple extensions with commas or use -e or --extension\nmultiple times:
django-admin makemessages --locale=de --extension=html,txt --extension xml\nSpecifies the locale(s) to process.
\nSpecifies the locale(s) to exclude from processing. If not provided, no locales\nare excluded.
\nExample usage:
\ndjango-admin makemessages --locale=pt_BR\ndjango-admin makemessages --locale=pt_BR --locale=fr\ndjango-admin makemessages -l pt_BR\ndjango-admin makemessages -l pt_BR -l fr\ndjango-admin makemessages --exclude=pt_BR\ndjango-admin makemessages --exclude=pt_BR --exclude=fr\ndjango-admin makemessages -x pt_BR\ndjango-admin makemessages -x pt_BR -x fr\nSpecifies the domain of the messages files. Supported options are:
\ndjango for all *.py, *.html and *.txt files (default)
djangojs for *.js files
Follows symlinks to directories when looking for new translation strings.
\nExample usage:
\ndjango-admin makemessages --locale=de --symlinks\nIgnores files or directories matching the given glob-style pattern. Use\nmultiple times to ignore more.
These patterns are used by default: 'CVS', '.*', '*~', '*.pyc'.
Example usage:
\ndjango-admin makemessages --locale=en_US --ignore=apps/* --ignore=secret/*.html\nDisables the default values of --ignore.
Disables breaking long message lines into several lines in language files.
\nSuppresses writing “#: filename:line’ comment lines in language files.\nUsing this option makes it harder for technically skilled translators to\nunderstand each message’s context.
Controls #: filename:line comment lines in language files. If the option\nis:
full (the default if not given): the lines include both file name and\nline number.
file: the line number is omitted.
never: the lines are suppressed (same as --no-location).
Requires gettext 0.19 or newer.
Removes obsolete message strings from the .po files.
Prevents deleting the temporary .pot files generated before creating the\n.po file. This is useful for debugging errors which may prevent the final\nlanguage files from being created.
makemigrationsLink to this heading#Creates new migrations based on the changes detected to your models.\nMigrations, their relationship with apps and more are covered in depth in\nthe migrations documentation.
\nProviding one or more app names as arguments will limit the migrations created\nto the app(s) specified and any dependencies needed (the table at the other end\nof a ForeignKey, for example).
To add migrations to an app that doesn’t have a migrations directory, run\nmakemigrations with the app’s app_label.
Suppresses all user prompts. If a suppressed prompt cannot be resolved\nautomatically, the command will exit with error code 3.
\nOutputs an empty migration for the specified apps, for manual editing. This is\nfor advanced users and should not be used unless you are familiar with the\nmigration format, migration operations, and the dependencies between your\nmigrations.
\nShows what migrations would be made without actually writing any migrations\nfiles to disk. Using this option along with --verbosity 3 will also show\nthe complete migrations files that would be written.
Enables fixing of migration conflicts.
\nAllows naming the generated migration(s) instead of using a generated name. The\nname must be a valid Python identifier.
\nGenerate migration files without Django version and timestamp header.
\nMakes makemigrations exit with a non-zero status when model changes without\nmigrations are detected. Implies --dry-run.
Diverts log output and input prompts to stderr, writing only paths of\ngenerated migration files to stdout.
Merges model changes into the latest migration and optimize the resulting\noperations.
\nThe updated migration will have a generated name. In order to preserve the\nprevious name, set it using --name.
migrateLink to this heading#Synchronizes the database state with the current set of models and migrations.\nMigrations, their relationship with apps and more are covered in depth in\nthe migrations documentation.
\nThe behavior of this command changes depending on the arguments provided:
\nNo arguments: All apps have all of their migrations run.
<app_label>: The specified app has its migrations run, up to the most\nrecent migration. This may involve running other apps” migrations too, due\nto dependencies.
<app_label> <migrationname>: Brings the database schema to a state where\nthe named migration is applied, but no later migrations in the same app are\napplied. This may involve unapplying migrations if you have previously\nmigrated past the named migration. You can use a prefix of the migration\nname, e.g. 0001, as long as it’s unique for the given app name. Use the\nname zero to migrate all the way back i.e. to revert all applied\nmigrations for an app.
Specifies the database to migrate. Defaults to default.
Marks the migrations up to the target one (following the rules above) as\napplied, but without actually running the SQL to change your database schema.
\nThis is intended for advanced users to manipulate the\ncurrent migration state directly if they’re manually applying changes;\nbe warned that using --fake runs the risk of putting the migration state\ntable into a state where manual recovery will be needed to make migrations\nrun correctly.
Allows Django to skip an app’s initial migration if all database tables with\nthe names of all models created by all\nCreateModel operations in that\nmigration already exist. This option is intended for use when first running\nmigrations against a database that preexisted the use of migrations. This\noption does not, however, check for matching database schema beyond matching\ntable names and so is only safe to use if you are confident that your existing\nschema matches what is recorded in your initial migration.
Shows the migration operations that will be performed for the given migrate\ncommand.
Allows creating tables for apps without migrations. While this isn’t\nrecommended, the migrations framework is sometimes too slow on large projects\nwith hundreds of models.
\nSuppresses all user prompts. An example prompt is asking about removing stale\ncontent types.
\nMakes migrate exit with a non-zero status when unapplied migrations are\ndetected.
Deletes nonexistent migrations from the django_migrations table. This is\nuseful when migration files replaced by a squashed migration have been removed.\nSee Squashing migrations for more details.
optimizemigrationLink to this heading#Optimizes the operations for the named migration and overrides the existing\nfile. If the migration contains functions that must be manually copied, the\ncommand creates a new migration file suffixed with _optimized that is meant\nto replace the named migration.
Makes optimizemigration exit with a non-zero status when a migration can be\noptimized.
runserverLink to this heading#Starts a lightweight development web server on the local machine. By default,\nthe server runs on port 8000 on the IP address 127.0.0.1. You can pass in\nan IP address and port number explicitly.
If you run this script as a user with normal privileges (recommended), you\nmight not have access to start a port on a low port number. Low port numbers\nare reserved for the superuser (root).
\nThis server uses the WSGI application object specified by the\nWSGI_APPLICATION setting.
The development server automatically reloads Python code for each request, as\nneeded. You don’t need to restart the server for code changes to take effect.\nHowever, some actions like adding files don’t trigger a restart, so you’ll\nhave to restart the server in these cases.
\nIf you’re using Linux or MacOS and install both pywatchman and the\nWatchman service, kernel signals will be used to autoreload the server\n(rather than polling file modification timestamps each second). This offers\nbetter performance on large projects, reduced response time after code changes,\nmore robust change detection, and a reduction in power usage. Django supports\npywatchman 1.2.0 and higher.
When you start the server, and each time you change Python code while the\nserver is running, the system check framework will check your entire Django\nproject for some common errors (see the check command). If any\nerrors are found, they will be printed to standard output. You can use the\n--skip-checks option to skip running system checks.
You can run as many concurrent servers as you want, as long as they’re on\nseparate ports by executing django-admin runserver more than once.
Note that the default IP address, 127.0.0.1, is not accessible from other\nmachines on your network. To make your development server viewable to other\nmachines on the network, use its own IP address (e.g. 192.168.2.1), 0\n(shortcut for 0.0.0.0), 0.0.0.0, or :: (with IPv6 enabled).
You can provide an IPv6 address surrounded by brackets\n(e.g. [200a::1]:8000). This will automatically enable IPv6 support.
A hostname containing ASCII-only characters can also be used.
\nIf the staticfiles contrib app is enabled\n(default in new projects) the runserver command will be overridden\nwith its own runserver command.
Logging of each request and response of the server is sent to the\ndjango.server logger.
\nDisables the auto-reloader. This means any Python code changes you make while\nthe server is running will not take effect if the particular Python modules\nhave already been loaded into memory.
\nDisables use of threading in the development server. The server is\nmultithreaded by default.
\nUses IPv6 for the development server. This changes the default IP address from\n127.0.0.1 to ::1.
By default, a warning is printed to the console that runserver is not\nsuitable for production:
WARNING: This is a development server. Do not use it in a production setting. Use a production WSGI or ASGI server instead.\nFor more information on production servers see: https://docs.djangoproject.com/en/|version|/howto/deployment/\nSet this environment variable to "true" to hide this warning.
Port 8000 on IP address 127.0.0.1:
django-admin runserver\nPort 8000 on IP address 1.2.3.4:
django-admin runserver 1.2.3.4:8000\nPort 7000 on IP address 127.0.0.1:
django-admin runserver 7000\nPort 7000 on IP address 1.2.3.4:
django-admin runserver 1.2.3.4:7000\nPort 8000 on IPv6 address ::1:
django-admin runserver -6\nPort 7000 on IPv6 address ::1:
django-admin runserver -6 7000\nPort 7000 on IPv6 address 2001:0db8:1234:5678::9:
django-admin runserver [2001:0db8:1234:5678::9]:7000\nPort 8000 on IPv4 address of host localhost:
django-admin runserver localhost:8000\nPort 8000 on IPv6 address of host localhost:
django-admin runserver -6 localhost:8000\nBy default, the development server doesn’t serve any static files for your site\n(such as CSS files, images, things under MEDIA_URL and so forth). If\nyou want to configure Django to serve static media, read\nHow to manage static files (e.g. images, JavaScript, CSS).
Django’s runserver command provides a WSGI server. In order to run under\nASGI you will need to use an ASGI server.\nThe Django Daphne project provides Integration with runserver that you can use.
sendtestemailLink to this heading#Sends a test email (to confirm email sending through Django is working) to the\nrecipient(s) specified. For example:
\ndjango-admin sendtestemail foo@example.com bar@example.com\nThere are a couple of options, and you may use any combination of them\ntogether:
\nMails the email addresses specified in MANAGERS using\nmail_managers().
Mails the email addresses specified in ADMINS using\nmail_admins().
shellLink to this heading#Starts the Python interactive interpreter.
\nAll models from installed apps are automatically imported into the shell\nenvironment. Models from apps listed earlier in INSTALLED_APPS take\nprecedence. The following common utilities are also imported:
from django.db import connection, reset_queries, models\nfrom django.conf import settings\nfrom django.utils import timezone\nFor a --verbosity of 2 or higher, the automatically imported objects will\nbe listed. To disable automatic importing entirely, use the --no-imports\nflag.
See the guide on customizing this behavior to add or remove automatic imports.
\n\n\nSpecifies the shell to use. By default, Django will use IPython or bpython if\neither is installed. If both are installed, specify which one you want like so:
\nIPython:
\ndjango-admin shell -i ipython\nbpython:
\ndjango-admin shell -i bpython\nIf you have a «rich» shell installed but want to force use of the «plain»\nPython interpreter, use python as the interface name, like so:
django-admin shell -i python\nDisables reading the startup script for the «plain» Python interpreter. By\ndefault, the script pointed to by the PYTHONSTARTUP environment\nvariable or the ~/.pythonrc.py script is read.
Disables the automatic import of models from INSTALLED_APPS.
Lets you pass a command as a string to execute it as Django, like so:
\ndjango-admin shell --command="import django; print(django.__version__)"\nYou can also pass code in on standard input to execute it. For example:
\n$ django-admin shell <<EOF\n> import django\n> print(django.__version__)\n> EOF\nOn Windows, the REPL is output due to implementation limits of\nselect.select() on that platform.
showmigrationsLink to this heading#Shows all migrations in a project. You can choose from one of two formats:
\nLists all of the apps Django knows about, the migrations available for each\napp, and whether or not each migration is applied (marked by an [X] next to\nthe migration name). For a --verbosity of 2 and above, the applied\ndatetimes are also shown.
Apps without migrations are also listed, but have (no migrations) printed\nunder them.
This is the default output format.
\nShows the migration plan Django will follow to apply migrations. Like\n--list, applied migrations are marked by an [X]. For a --verbosity\nof 2 and above, all dependencies of a migration will also be shown.
app_labels arguments limit the output, however, dependencies of provided\napps may also be included.
Specifies the database to examine. Defaults to default.
sqlflushLink to this heading#Prints the SQL statements that would be executed for the flush\ncommand.
Specifies the database for which to print the SQL. Defaults to default.
sqlmigrateLink to this heading#Prints the SQL for the named migration. This requires an active database\nconnection, which it will use to resolve constraint names; this means you must\ngenerate the SQL against a copy of the database you wish to later apply it on.
\nNote that sqlmigrate doesn’t colorize its output.
Generates the SQL for unapplying the migration. By default, the SQL created is\nfor running the migration in the forwards direction.
\nSpecifies the database for which to generate the SQL. Defaults to default.
sqlsequenceresetLink to this heading#Prints the SQL statements for resetting sequences for the given app name(s).
\nSequences are indexes used by some database engines to track the next available\nnumber for automatically incremented fields.
\nUse this command to generate SQL which will fix cases where a sequence is out\nof sync with its automatically incremented field data.
\nSpecifies the database for which to print the SQL. Defaults to default.
squashmigrationsLink to this heading#Squashes the migrations for app_label up to and including\nmigration_name down into fewer migrations, if possible. The resulting\nsquashed migrations can live alongside the unsquashed ones safely. For more\ninformation, please read Squashing migrations.
When start_migration_name is given, Django will only include migrations\nstarting from and including this migration. This helps to mitigate the\nsquashing limitation of RunPython and\ndjango.db.migrations.operations.RunSQL migration operations.
Disables the optimizer when generating a squashed migration. By default, Django\nwill try to optimize the operations in your migrations to reduce the size of\nthe resulting file. Use this option if this process is failing or creating\nincorrect migrations, though please also file a Django bug report about the\nbehavior, as optimization is meant to be safe.
\nSuppresses all user prompts.
\nSets the name of the squashed migration. When omitted, the name is based on the\nfirst and last migration, with _squashed_ in between.
Generate squashed migration file without Django version and timestamp header.
\nstartappLink to this heading#Creates a Django app directory structure for the given app name in the current\ndirectory or the given destination.
\nBy default, the new directory contains a\nmodels.py file and other app template files. If only the app name is given,\nthe app directory will be created in the current working directory.
If the optional destination is provided, Django will use that name instead. If\nthe directory with the given name doesn’t exist, it will be created. You can\nuse “.” to denote the current working directory.
\n\nFor example:
\ndjango-admin startapp myapp /Users/jezdez/Code/myapp\nProvides the path to a directory with a custom app template file, or a path to\nan uncompressed archive (.tar) or a compressed archive (.tar.gz,\n.tar.bz2, .tar.xz, .tar.lzma, .tgz, .tbz2, .txz,\n.tlz, .zip) containing the app template files.
For example, this would look for an app template in the given directory when\ncreating the myapp app:
django-admin startapp --template=/Users/jezdez/Code/my_app_template myapp\nDjango will also accept URLs (http, https, ftp) to compressed\narchives with the app template files, downloading and extracting them on the\nfly.
For example, taking advantage of GitHub’s feature to expose repositories as\nzip files, you can use a URL like:
\ndjango-admin startapp --template=https://github.com/githubuser/django-app-template/archive/main.zip myapp\nSpecifies which file extensions in the app template should be rendered with the\ntemplate engine. Defaults to py.
Specifies which files in the app template (in addition to those matching\n--extension) should be rendered with the template engine. Defaults to an\nempty list.
Specifies which directories in the app template should be excluded, in addition\nto .git and __pycache__. If this option is not provided, directories\nnamed __pycache__ or starting with . will be excluded.
The template context used for all matching\nfiles is:
Any option passed to the startapp command (among the command’s supported\noptions)
app_name – the app name as passed to the command
app_directory – the full path of the newly created app
camel_case_app_name – the app name in camel case format
docs_version – the version of the documentation: 'dev' or '1.x'
django_version – the version of Django, e.g. '2.0.3'
startprojectLink to this heading#Creates a Django project directory structure for the given project name in\nthe current directory or the given destination.
\nBy default, the new directory contains\nmanage.py and a project package (containing a settings.py and other\nfiles).
If only the project name is given, both the project directory and project\npackage will be named <projectname> and the project directory\nwill be created in the current working directory.
If the optional destination is provided, Django will use that name as the\nproject directory, and create manage.py and the project package within it.\nIf the directory with the given name doesn’t exist, it will be created. Use “.”\nto denote the current working directory.
For example:
\ndjango-admin startproject myproject /Users/jezdez/Code/myproject_repo\nSpecifies a directory, file path, or URL of a custom project template. See the\nstartapp --template documentation for examples and usage. The same\nsecurity considerations described for startapp templates apply here:\nmalicious or poorly constructed templates may introduce weaknesses or consume\nexcessive resources, and templates should be carefully inspected before use.
Specifies which file extensions in the project template should be rendered with\nthe template engine. Defaults to py.
Specifies which files in the project template (in addition to those matching\n--extension) should be rendered with the template engine. Defaults to an\nempty list.
Specifies which directories in the project template should be excluded, in\naddition to .git and __pycache__. If this option is not provided,\ndirectories named __pycache__ or starting with . will be excluded.
The template context used is:
Any option passed to the startproject command (among the command’s\nsupported options)
project_name – the project name as passed to the command
project_directory – the full path of the newly created project
secret_key – a random key for the SECRET_KEY setting
docs_version – the version of the documentation: 'dev' or '1.x'
django_version – the version of Django, e.g. '2.0.3'
Please also see the rendering warning and\ntrusted code warning as mentioned for\nstartapp.
testLink to this heading#Runs tests for all installed apps. See Testing in Django for more\ninformation.
\nStops running tests and reports the failure immediately after a test fails.
\nControls the test runner class that is used to execute tests. This value\noverrides the value provided by the TEST_RUNNER setting.
Suppresses all user prompts. A typical prompt is a warning about deleting an\nexisting test database.
\nThe test command receives options on behalf of the specified\n--testrunner. These are the options of the default test runner:\nDiscoverRunner.
Preserves the test database between test runs. This has the advantage of\nskipping both the create and destroy actions which can greatly decrease the\ntime to run tests, especially those in a large test suite. If the test database\ndoes not exist, it will be created on the first run and then preserved for each\nsubsequent run. Unless the MIGRATE test setting is\nFalse, any unapplied migrations will also be applied to the test database\nbefore running the test suite.
Randomizes the order of tests before running them. This can help detect tests\nthat aren’t properly isolated. The test order generated by this option is a\ndeterministic function of the integer seed given. When no seed is passed, a\nseed is chosen randomly and printed to the console. To repeat a particular test\norder, pass a seed. The test orders generated by this option preserve Django’s\nguarantees on test order. They also keep tests grouped\nby test case class.
\nThe shuffled orderings also have a special consistency property useful when\nnarrowing down isolation issues. Namely, for a given seed and when running a\nsubset of tests, the new order will be the original shuffling restricted to the\nsmaller set. Similarly, when adding tests while keeping the seed the same, the\norder of the original tests will be the same in the new order.
\nSorts test cases in the opposite execution order. This may help in debugging\nthe side effects of tests that aren’t properly isolated. Grouping by test\nclass is preserved when using this option. This can be used\nin conjunction with --shuffle to reverse the order for a particular seed.
Sets the DEBUG setting to True prior to running tests. This may\nhelp troubleshoot test failures.
Enables SQL logging for failing tests. If\n--verbosity is 2, then queries in passing tests are also output.
Runs tests in separate parallel processes. Since modern processors have\nmultiple cores, this allows running tests significantly faster.
\nUsing --parallel without a value, or with the value auto, runs one test\nprocess per core according to multiprocessing.cpu_count(). You can\noverride this by passing the desired number of processes, e.g.\n--parallel 4, or by setting the DJANGO_TEST_PROCESSES environment\nvariable.
Django distributes test cases — unittest.TestCase subclasses — to\nsubprocesses. If there are fewer test case classes than configured processes,\nDjango will reduce the number of processes accordingly.
Each process gets its own database. You must ensure that different test case\nclasses don’t access the same resources. For instance, test case classes that\ntouch the filesystem should create a temporary directory for their own use.
\n\nThis option requires the third-party tblib package to display tracebacks\ncorrectly:
$ python -m pip install tblib\nThis feature isn’t available on Windows. It doesn’t work with the Oracle\ndatabase backend either.
\nIf you want to use pdb while debugging tests, you must disable parallel\nexecution (--parallel=1). You’ll see something like bdb.BdbQuit if you\ndon’t.
Runs only tests marked with the specified tags.\nMay be specified multiple times and combined with test --exclude-tag.
Tests that fail to load are always considered matching.
\nExcludes tests marked with the specified tags.\nMay be specified multiple times and combined with test --tag.
Runs test methods and classes matching test name patterns, in the same way as\nunittest's -k option. Can be specified multiple times.
Spawns a pdb debugger at each test error or failure. If you have it\ninstalled, ipdb is used instead.
Discards output (stdout and stderr) for passing tests, in the same way\nas unittest's --buffer option.
Django automatically calls faulthandler.enable() when starting the\ntests, which allows it to print a traceback if the interpreter crashes. Pass\n--no-faulthandler to disable this behavior.
Outputs timings, including database setup and total run time.
\nShows the N slowest test cases (N=0 for all).
\ntestserverLink to this heading#Runs a Django development server (as in runserver) using data from\nthe given fixture(s).
For example, this command:
\ndjango-admin testserver mydata.json\n…would perform the following steps:
\nCreate a test database, as described in The test database.
Populate the test database with fixture data from the given fixtures.\n(For more on fixtures, see the documentation for loaddata above.)
Runs the Django development server (as in runserver), pointed at\nthis newly created test database instead of your production database.
This is useful in a number of ways:
\nWhen you’re writing unit tests of how your\nviews act with certain fixture data, you can use testserver to interact\nwith the views in a web browser, manually.
Let’s say you’re developing your Django application and have a «pristine»\ncopy of a database that you’d like to interact with. You can dump your\ndatabase to a fixture (using the\ndumpdata command, explained above), then use testserver to run\nyour web application with that data. With this arrangement, you have the\nflexibility of messing up your data in any way, knowing that whatever data\nchanges you’re making are only being made to a test database.
Note that this server does not automatically detect changes to your Python\nsource code (as runserver does). It does, however, detect changes to\ntemplates.
Specifies a different port, or IP address and port, from the default of\n127.0.0.1:8000. This value follows exactly the same format and serves\nexactly the same function as the argument to the runserver command.
Examples:
\nTo run the test server on port 7000 with fixture1 and fixture2:
django-admin testserver --addrport 7000 fixture1 fixture2\ndjango-admin testserver fixture1 fixture2 --addrport 7000\n(The above statements are equivalent. We include both of them to demonstrate\nthat it doesn’t matter whether the options come before or after the fixture\narguments.)
\nTo run on 1.2.3.4:7000 with a test fixture:
django-admin testserver --addrport 1.2.3.4:7000 test\nSuppresses all user prompts. A typical prompt is a warning about deleting an\nexisting test database.
\nSome commands are only available when the django.contrib application that\nimplements them has been\nenabled. This section describes them grouped by\ntheir application.
django.contrib.authLink to this heading#changepasswordLink to this heading#This command is only available if Django’s authentication system (django.contrib.auth) is installed.
Allows changing a user’s password. It prompts you to enter a new password twice\nfor the given user. If the entries are identical, this immediately becomes the\nnew password. If you do not supply a user, the command will attempt to change\nthe password whose username matches the current user.
\nSpecifies the database to query for the user. Defaults to default.
Example usage:
\ndjango-admin changepassword ringo\ncreatesuperuserLink to this heading#This command is only available if Django’s authentication system (django.contrib.auth) is installed.
Creates a superuser account (a user who has all permissions). This is\nuseful if you need to create an initial superuser account or if you need to\nprogrammatically generate superuser accounts for your site(s).
\nWhen run interactively, this command will prompt for a password for\nthe new superuser account. When run non-interactively, you can provide\na password by setting the DJANGO_SUPERUSER_PASSWORD environment\nvariable. Otherwise, no password will be set, and the superuser account will\nnot be able to log in until a password has been manually set for it.
In non-interactive mode, the\nUSERNAME_FIELD and required\nfields (listed in\nREQUIRED_FIELDS) fall back to\nDJANGO_SUPERUSER_<uppercase_field_name> environment variables, unless they\nare overridden by a command line argument. For example, to provide an email\nfield, you can use DJANGO_SUPERUSER_EMAIL environment variable.
Suppresses all user prompts. If a suppressed prompt cannot be resolved\nautomatically, the command will exit with error code 1.
\nThe username and email address for the new account can be supplied by\nusing the --username and --email arguments on the command\nline. If either of those is not supplied, createsuperuser will prompt for\nit when running interactively.
Specifies the database into which the superuser object will be saved.
\nYou can subclass the management command and override get_input_data() if\nyou want to customize data input and validation. Consult the source code for\ndetails on the existing implementation and the method’s parameters. For\nexample, it could be useful if you have a ForeignKey in\nREQUIRED_FIELDS and want to\nallow creating an instance instead of entering the primary key of an existing\ninstance.
django.contrib.contenttypesLink to this heading#remove_stale_contenttypesLink to this heading#This command is only available if Django’s contenttypes app (django.contrib.contenttypes) is installed.
Deletes stale content types (from deleted models) in your database. Any objects\nthat depend on the deleted content types will also be deleted. A list of\ndeleted objects will be displayed before you confirm it’s okay to proceed with\nthe deletion.
\nSpecifies the database to use. Defaults to default.
Deletes stale content types including ones from previously installed apps that\nhave been removed from INSTALLED_APPS. Defaults to False.
django.contrib.gisLink to this heading#ogrinspectLink to this heading#This command is only available if GeoDjango\n(django.contrib.gis) is installed.
Please refer to its description in the GeoDjango\ndocumentation.
django.contrib.sessionsLink to this heading#clearsessionsLink to this heading#Can be run as a cron job or directly to clean out expired sessions.
\ndjango.contrib.staticfilesLink to this heading#collectstaticLink to this heading#This command is only available if the static files application (django.contrib.staticfiles) is installed.
Please refer to its description in the\nstaticfiles documentation.
findstaticLink to this heading#This command is only available if the static files application (django.contrib.staticfiles) is installed.
Please refer to its description in the\nstaticfiles documentation.
Although some commands may allow their own custom options, every command\nallows for the following options by default:
\nAdds the given filesystem path to the Python sys.path module\nattribute. If this isn’t provided, django-admin will use the\nPYTHONPATH environment variable.
This option is unnecessary in manage.py, because it takes care of setting\nthe Python path for you.
Example usage:
\ndjango-admin migrate --pythonpath='/home/djangoprojects/myproject'\nSpecifies the settings module to use. The settings module should be in Python\npackage syntax, e.g. mysite.settings. If this isn’t provided,\ndjango-admin will use the DJANGO_SETTINGS_MODULE environment\nvariable.
This option is unnecessary in manage.py, because it uses\nsettings.py from the current project by default.
Example usage:
\ndjango-admin migrate --settings=mysite.settings\nDisplays a full stack trace when a CommandError\nis raised. By default, django-admin will show an error message when a\nCommandError occurs and a full stack trace for any other exception.
This option is ignored by runserver.
Example usage:
\ndjango-admin migrate --traceback\nSpecifies the amount of notification and debug information that a command\nshould print to the console.
\n0 means no output.
1 means normal output (default).
2 means verbose output.
3 means very verbose output.
This option is ignored by runserver.
Example usage:
\ndjango-admin migrate --verbosity 2\nDisables colorized command output. Some commands format their output to be\ncolorized. For example, errors will be printed to the console in red and SQL\nstatements will be syntax highlighted.
\nExample usage:
\ndjango-admin runserver --no-color\nForces colorization of the command output if it would otherwise be disabled\nas discussed in Syntax coloring. For example, you may want to pipe\ncolored output to another command.
\nSkips running system checks prior to running the command. This option is only\navailable if the\nrequires_system_checks command\nattribute is not an empty list or tuple.
Example usage:
\ndjango-admin migrate --skip-checks\nThe django-admin / manage.py commands will use pretty\ncolor-coded output if your terminal supports ANSI-colored output. It\nwon’t use the color codes if you’re piping the command’s output to\nanother program unless the --force-color option is used.
On Windows 10, the Windows Terminal application, VS Code, and PowerShell\n(where virtual terminal processing is enabled) allow colored output, and are\nsupported by default.
\nUnder Windows, the legacy cmd.exe native console doesn’t support ANSI\nescape sequences so by default there is no color output. In this case either of\ntwo third-party libraries are needed:
Install colorama, a Python package that translates ANSI color codes\ninto Windows API calls. Django commands will detect its presence and will\nmake use of its services to color output just like on Unix-based platforms.\ncolorama can be installed via pip:
...\\> py -m pip install "colorama >= 0.4.6"\nInstall ANSICON, a third-party tool that allows cmd.exe to process\nANSI color codes. Django commands will detect its presence and will make use\nof its services to color output just like on Unix-based platforms.
Other modern terminal environments on Windows, that support terminal colors,\nbut which are not automatically detected as supported by Django, may «fake» the\ninstallation of ANSICON by setting the appropriate environmental variable,\nANSICON="on".
The colors used for syntax highlighting can be customized. Django\nships with three color palettes:
\ndark, suited to terminals that show white text on a black\nbackground. This is the default palette.
light, suited to terminals that show black text on a white\nbackground.
nocolor, which disables syntax highlighting.
You select a palette by setting a DJANGO_COLORS environment\nvariable to specify the palette you want to use. For example, to\nspecify the light palette under a Unix or OS/X BASH shell, you\nwould run the following at a command prompt:
export DJANGO_COLORS="light"\nYou can also customize the colors that are used. Django specifies a\nnumber of roles in which color is used:
\nerror - A major error.
notice - A minor error.
success - A success.
warning - A warning.
sql_field - The name of a model field in SQL.
sql_coltype - The type of a model field in SQL.
sql_keyword - An SQL keyword.
sql_table - The name of a model in SQL.
http_info - A 1XX HTTP Informational server response.
http_success - A 2XX HTTP Success server response.
http_not_modified - A 304 HTTP Not Modified server response.
http_redirect - A 3XX HTTP Redirect server response other than 304.
http_not_found - A 404 HTTP Not Found server response.
http_bad_request - A 4XX HTTP Bad Request server response other than 404.
http_server_error - A 5XX HTTP Server Error response.
migrate_heading - A heading in a migrations management command.
migrate_label - A migration name.
Each of these roles can be assigned a specific foreground and\nbackground color, from the following list:
\nblack
red
green
yellow
blue
magenta
cyan
white
Each of these colors can then be modified by using the following\ndisplay options:
\nbold
underscore
blink
reverse
conceal
A color specification follows one of the following patterns:
\nrole=fg
role=fg/bg
role=fg,option,option
role=fg/bg,option,option
where role is the name of a valid color role, fg is the\nforeground color, bg is the background color and each option\nis one of the color modifying options. Multiple color specifications\nare then separated by a semicolon. For example:
export DJANGO_COLORS="error=yellow/blue,blink;notice=magenta"\nwould specify that errors be displayed using blinking yellow on blue,\nand notices displayed using magenta. All other color roles would be\nleft uncolored.
\nColors can also be specified by extending a base palette. If you put\na palette name in a color specification, all the colors implied by that\npalette will be loaded. So:
\nexport DJANGO_COLORS="light;error=yellow/blue,blink;notice=magenta"\nwould specify the use of all the colors in the light color palette,\nexcept for the colors for errors and notices which would be\noverridden as specified.
\nIf you use the Bash shell, consider installing the Django bash completion\nscript, which lives in extras/django_bash_completion in the Django\nsource distribution. It enables tab-completion of django-admin and\nmanage.py commands, so you can, for instance…
Type django-admin.
Press [TAB] to see all available options.
Type sql, then [TAB], to see all available options whose names start\nwith sql.
See How to create custom django-admin commands for how to add customized actions.
\nThe Python files created by startproject, startapp,\noptimizemigration, makemigrations, and\nsquashmigrations are formatted using the black command if it is\npresent on your PATH.
If you have black globally installed, but do not wish it used for the\ncurrent project, you can set the PATH explicitly:
PATH=path/to/venv/bin django-admin makemigrations\nFor commands using stdout you can pipe the output to black if needed:
django-admin inspectdb | black -\nTo call a management command from code use call_command().
namethe name of the command to call or a command object. Passing the name is\npreferred unless the object is required for testing.
\n*argsa list of arguments accepted by the command. Arguments are passed to the\nargument parser, so you can use the same style as you would on the command\nline. For example, call_command('flush', '--verbosity=0').
**optionsnamed options accepted on the command-line. Options are passed to the command\nwithout triggering the argument parser, which means you’ll need to pass the\ncorrect type. For example, call_command('flush', verbosity=0) (zero must\nbe an integer rather than a string).
Examples:
\nfrom django.core import management\nfrom django.core.management.commands import loaddata\n\nmanagement.call_command("flush", verbosity=0, interactive=False)\nmanagement.call_command("loaddata", "test_data", verbosity=0)\nmanagement.call_command(loaddata.Command(), "test_data", verbosity=0)\nNote that command options that take no arguments are passed as keywords\nwith True or False, as you can see with the interactive option\nabove.
Named arguments can be passed by using either one of the following syntaxes:
\n# Similar to the command line\nmanagement.call_command("dumpdata", "--natural-foreign")\n\n# Named argument similar to the command line minus the initial dashes and\n# with internal dashes replaced by underscores\nmanagement.call_command("dumpdata", natural_foreign=True)\n\n# `use_natural_foreign_keys` is the option destination variable\nmanagement.call_command("dumpdata", use_natural_foreign_keys=True)\nSome command options have different names when using call_command() instead\nof django-admin or manage.py. For example, django-admin\ncreatesuperuser --no-input translates to call_command('createsuperuser',\ninteractive=False). To find what keyword argument name to use for\ncall_command(), check the command’s source code for the dest argument\npassed to parser.add_argument().
Command options which take multiple options are passed a list:
\nmanagement.call_command("dumpdata", exclude=["contenttypes", "auth"])\nThe return value of the call_command() function is the same as the return\nvalue of the handle() method of the command.
Note that you can redirect standard output and error streams as all commands\nsupport the stdout and stderr options. For example, you could write:
with open("/path/to/command_output", "w") as f:\n management.call_command("dumpdata", stdout=f)\n