Databases

Django attempts to support as many features as possible on all databasebackends. However, not all database backends are alike, and we’ve had to makedesign decisions on which features to support and which assumptions we can makesafely.

This file describes some of the features that might be relevant to Djangousage. Of course, it is not intended as a replacement for server-specificdocumentation or reference manuals.

General notes

Persistent connections

Persistent connections avoid the overhead of re-establishing a connection tothe database in each request. They’re controlled by theCONN_MAX_AGE parameter which defines the maximum lifetime of aconnection. It can be set independently for each database.

The default value is 0, preserving the historical behavior of closing thedatabase connection at the end of each request. To enable persistentconnections, set CONN_MAX_AGE to a positive number of seconds. Forunlimited persistent connections, set it to None.

Connection management

Django opens a connection to the database when it first makes a databasequery. It keeps this connection open and reuses it in subsequent requests.Django closes the connection once it exceeds the maximum age defined byCONN_MAX_AGE or when it isn’t usable any longer.

In detail, Django automatically opens a connection to the database whenever itneeds one and doesn’t have one already — either because this is the firstconnection, or because the previous connection was closed.

At the beginning of each request, Django closes the connection if it hasreached its maximum age. If your database terminates idle connections aftersome time, you should set CONN_MAX_AGE to a lower value, so thatDjango doesn’t attempt to use a connection that has been terminated by thedatabase server. (This problem may only affect very low traffic sites.)

At the end of each request, Django closes the connection if it has reached itsmaximum age or if it is in an unrecoverable error state. If any databaseerrors have occurred while processing the requests, Django checks whether theconnection still works, and closes it if it doesn’t. Thus, database errorsaffect at most one request; if the connection becomes unusable, the nextrequest gets a fresh connection.

Caveats

Since each thread maintains its own connection, your database must support atleast as many simultaneous connections as you have worker threads.

Sometimes a database won’t be accessed by the majority of your views, forexample because it’s the database of an external system, or thanks to caching.In such cases, you should set CONN_MAX_AGE to a low value or even0, because it doesn’t make sense to maintain a connection that’s unlikelyto be reused. This will help keep the number of simultaneous connections tothis database small.

The development server creates a new thread for each request it handles,negating the effect of persistent connections. Don’t enable them duringdevelopment.

When Django establishes a connection to the database, it sets up appropriateparameters, depending on the backend being used. If you enable persistentconnections, this setup is no longer repeated every request. If you modifyparameters such as the connection’s isolation level or time zone, you shouldeither restore Django’s defaults at the end of each request, force anappropriate value at the beginning of each request, or disable persistentconnections.

Encoding

Django assumes that all databases use UTF-8 encoding. Using other encodings mayresult in unexpected behavior such as “value too long” errors from yourdatabase for data that is valid in Django. See the database specific notesbelow for information on how to set up your database correctly.

PostgreSQL notes

Django supports PostgreSQL 9.5 and higher. psycopg2 2.5.4 or higher isrequired, though the latest release is recommended.

PostgreSQL connection settings

See HOST for details.

Optimizing PostgreSQL’s configuration

Django needs the following parameters for its database connections:

  • client_encoding: 'UTF8',
  • default_transaction_isolation: 'read committed' by default,or the value set in the connection options (see below),
  • timezone: 'UTC' when USE_TZ is True, value ofTIME_ZONE otherwise.If these parameters already have the correct values, Django won’t set them forevery new connection, which improves performance slightly. You can configurethem directly in postgresql.conf or more conveniently per databaseuser with ALTER ROLE.

Django will work just fine without this optimization, but each new connectionwill do some additional queries to set these parameters.

Isolation level

Like PostgreSQL itself, Django defaults to the READ COMMITTED isolationlevel. If you need a higher isolation level such as REPEATABLE READ orSERIALIZABLE, set it in the OPTIONS part of your databaseconfiguration in DATABASES:

  1. import psycopg2.extensions
  2.  
  3. DATABASES = {
  4.     # ...
  5.     'OPTIONS': {
  6.         'isolation_level': psycopg2.extensions.ISOLATION_LEVEL_SERIALIZABLE,
  7.     },
  8. }

Note

Under higher isolation levels, your application should be prepared tohandle exceptions raised on serialization failures. This option isdesigned for advanced uses.

Indexes for varchar and text columns

When specifying db_index=True on your model fields, Django typicallyoutputs a single CREATE INDEX statement. However, if the database typefor the field is either varchar or text (e.g., used by CharField,FileField, and TextField), then Django will createan additional index that uses an appropriate PostgreSQL operator classfor the column. The extra index is necessary to correctly performlookups that use the LIKE operator in their SQL, as is done with thecontains and startswith lookup types.

Migration operation for adding extensions

If you need to add a PostgreSQL extension (like hstore, postgis, etc.)using a migration, use theCreateExtension operation.

Server-side cursors

When using QuerySet.iterator(), Django opens a server-sidecursor. By default, PostgreSQL assumes thatonly the first 10% of the results of cursor queries will be fetched. The queryplanner spends less time planning the query and starts returning resultsfaster, but this could diminish performance if more than 10% of the results areretrieved. PostgreSQL’s assumptions on the number of rows retrieved for acursor query is controlled with the cursor_tuple_fraction option.

Transaction pooling and server-side cursors

Using a connection pooler in transaction pooling mode (e.g. pgBouncer)requires disabling server-side cursors for that connection.

Server-side cursors are local to a connection and remain open at the end of atransaction when AUTOCOMMIT is True. Asubsequent transaction may attempt to fetch more results from a server-sidecursor. In transaction pooling mode, there’s no guarantee that subsequenttransactions will use the same connection. If a different connection is used,an error is raised when the transaction references the server-side cursor,because server-side cursors are only accessible in the connection in which theywere created.

One solution is to disable server-side cursors for a connection inDATABASES by setting DISABLE_SERVER_SIDE_CURSORS to True.

To benefit from server-side cursors in transaction pooling mode, you could setup another connection to the database in order toperform queries that use server-side cursors. This connection needs to eitherbe directly to the database or to a connection pooler in session pooling mode.

Another option is to wrap each QuerySet using server-side cursors in anatomic() block, because it disables autocommitfor the duration of the transaction. This way, the server-side cursor will onlylive for the duration of the transaction.

Manually-specifying values of auto-incrementing primary keys

Django uses PostgreSQL’s SERIAL data type to store auto-incrementing primarykeys. A SERIAL column is populated with values from a sequence thatkeeps track of the next available value. Manually assigning a value to anauto-incrementing field doesn’t update the field’s sequence, which might latercause a conflict. For example:

  1. >>> from django.contrib.auth.models import User
  2. >>> User.objects.create(username='alice', pk=1)
  3. <User: alice>
  4. >>> # The sequence hasn't been updated; its next value is 1.
  5. >>> User.objects.create(username='bob')
  6. ...
  7. IntegrityError: duplicate key value violates unique constraint
  8. "auth_user_pkey" DETAIL:  Key (id)=(1) already exists.

If you need to specify such values, reset the sequence afterwards to avoidreusing a value that’s already in the table. The sqlsequenceresetmanagement command generates the SQL statements to do that.

Test database templates

You can use the TEST['TEMPLATE'] setting to specifya template (e.g. 'template0') from which to create a test database.

Speeding up test execution with non-durable settings

You can speed up test execution times by configuring PostgreSQL to benon-durable.

Warning

This is dangerous: it will make your database more susceptible to data lossor corruption in the case of a server crash or power loss. Only use this ona development machine where you can easily restore the entire contents ofall databases in the cluster.

MariaDB notes

New in Django 3.0:

Django supports MariaDB 10.1 and higher.

To use MariaDB, use the MySQL backend, which is shared between the two. See theMySQL notes for more details.

MySQL notes

Version support

Django supports MySQL 5.6 and higher.

Django’s inspectdb feature uses the information_schema database, whichcontains detailed data on all database schemas.

Django expects the database to support Unicode (UTF-8 encoding) and delegates toit the task of enforcing transactions and referential integrity. It is importantto be aware of the fact that the two latter ones aren’t actually enforced byMySQL when using the MyISAM storage engine, see the next section.

Storage engines

MySQL has several storage engines. You can change the default storage enginein the server configuration.

MySQL’s default storage engine is InnoDB. This engine is fully transactionaland supports foreign key references. It’s the recommended choice. However, theInnoDB autoincrement counter is lost on a MySQL restart because it does notremember the AUTO_INCREMENT value, instead recreating it as “max(id)+1”.This may result in an inadvertent reuse of AutoFieldvalues.

The main drawbacks of MyISAM are that it doesn’t support transactions orenforce foreign-key constraints.

MySQL DB API Drivers

MySQL has a couple drivers that implement the Python Database API described inPEP 249:

  • mysqlclient is a native driver. It’s the recommended choice.
  • MySQL Connector/Python is a pure Python driver from Oracle that does notrequire the MySQL client library or any Python modules outside the standardlibrary.These drivers are thread-safe and provide connection pooling.

In addition to a DB API driver, Django needs an adapter to access the databasedrivers from its ORM. Django provides an adapter for mysqlclient while MySQLConnector/Python includes its own.

mysqlclient

Django requires mysqlclient 1.3.13 or later.

MySQL Connector/Python

MySQL Connector/Python is available from the download page.The Django adapter is available in versions 1.1.X and later. It may notsupport the most recent releases of Django.

Time zone definitions

If you plan on using Django’s timezone support,use mysql_tzinfo_to_sql to load time zone tables into the MySQL database.This needs to be done just once for your MySQL server, not per database.

Creating your database

You can create your database using the command-line tools and this SQL:

  1. CREATE DATABASE <dbname> CHARACTER SET utf8;

This ensures all tables and columns will use UTF-8 by default.

Collation settings

The collation setting for a column controls the order in which data is sortedas well as what strings compare as equal. It can be set on a database-widelevel and also per-table and per-column. This is documented thoroughly inthe MySQL documentation. In all cases, you set the collation by directlymanipulating the database tables; Django doesn’t provide a way to set this onthe model definition.

By default, with a UTF-8 database, MySQL will use theutf8general_ci collation. This results in all string equalitycomparisons being done in a _case-insensitive manner. That is, "Fred" and"freD" are considered equal at the database level. If you have a uniqueconstraint on a field, it would be illegal to try to insert both "aa" and"AA" into the same column, since they compare as equal (and, hence,non-unique) with the default collation. If you want case-sensitive comparisonson a particular column or table, change the column or table to use theutf8_bin collation.

Please note that according to MySQL Unicode Character Sets, comparisons forthe utf8_general_ci collation are faster, but slightly less correct, thancomparisons for utf8_unicode_ci. If this is acceptable for your application,you should use utf8_general_ci because it is faster. If this is not acceptable(for example, if you require German dictionary order), use utf8_unicode_cibecause it is more accurate.

Warning

Model formsets validate unique fields in a case-sensitive manner. Thus whenusing a case-insensitive collation, a formset with unique field values thatdiffer only by case will pass validation, but upon calling save(), anIntegrityError will be raised.

Connecting to the database

Refer to the settings documentation.

Connection settings are used in this order:

Here’s a sample configuration which uses a MySQL option file:

  1. # settings.py
  2. DATABASES = {
  3.     'default': {
  4.         'ENGINE': 'django.db.backends.mysql',
  5.         'OPTIONS': {
  6.             'read_default_file': '/path/to/my.cnf',
  7.         },
  8.     }
  9. }
  10.  
  11.  
  12. # my.cnf
  13. [client]
  14. database = NAME
  15. user = USER
  16. password = PASSWORD
  17. default-character-set = utf8

Several other MySQLdb connection options may be useful, such as ssl,init_command, and sql_mode.

Setting sql_mode

From MySQL 5.7 onwards and on fresh installs of MySQL 5.6, the default value ofthe sql_mode option contains STRICT_TRANS_TABLES. That option escalateswarnings into errors when data are truncated upon insertion, so Django highlyrecommends activating a strict mode for MySQL to prevent data loss (eitherSTRICT_TRANS_TABLES or STRICT_ALL_TABLES).

If you need to customize the SQL mode, you can set the sql_mode variablelike other MySQL options: either in a config file or with the entry'init_command': "SET sql_mode='STRICT_TRANS_TABLES'" in theOPTIONS part of your database configuration in DATABASES.

Isolation level

When running concurrent loads, database transactions from different sessions(say, separate threads handling different requests) may interact with eachother. These interactions are affected by each session’s transaction isolationlevel. You can set a connection’s isolation level with an'isolation_level' entry in the OPTIONS part of your databaseconfiguration in DATABASES. Valid values forthis entry are the four standard isolation levels:

  • 'read uncommitted'
  • 'read committed'
  • 'repeatable read'
  • 'serializable'or None to use the server’s configured isolation level. However, Djangoworks best with and defaults to read committed rather than MySQL’s default,repeatable read. Data loss is possible with repeatable read. In particular,you may see cases where get_or_create()will raise an IntegrityError but the object won’t appear ina subsequent get() call.

Creating your tables

When Django generates the schema, it doesn’t specify a storage engine, sotables will be created with whatever default storage engine your databaseserver is configured for. The easiest solution is to set your database server’sdefault storage engine to the desired engine.

If you’re using a hosting service and can’t change your server’s defaultstorage engine, you have a couple of options.

  • After the tables are created, execute an ALTER TABLE statement toconvert a table to a new storage engine (such as InnoDB):
  1. ALTER TABLE <tablename> ENGINE=INNODB;

This can be tedious if you have a lot of tables.

  • Another option is to use the init_command option for MySQLdb prior tocreating your tables:
  1. 'OPTIONS': {
  2.    'init_command': 'SET default_storage_engine=INNODB',
  3. }

This sets the default storage engine upon connecting to the database.After your tables have been created, you should remove this option as itadds a query that is only needed during table creation to each databaseconnection.

Table names

There are known issues in even the latest versions of MySQL that can cause thecase of a table name to be altered when certain SQL statements are executedunder certain conditions. It is recommended that you use lowercase tablenames, if possible, to avoid any problems that might arise from this behavior.Django uses lowercase table names when it auto-generates table names frommodels, so this is mainly a consideration if you are overriding the table namevia the db_table parameter.

Savepoints

Both the Django ORM and MySQL (when using the InnoDB storage engine) support database savepoints.

If you use the MyISAM storage engine please be aware of the fact that you willreceive database-generated errors if you try to use the savepoint-relatedmethods of the transactions API. The reasonfor this is that detecting the storage engine of a MySQL database/table is anexpensive operation so it was decided it isn’t worth to dynamically convertthese methods in no-op’s based in the results of such detection.

Notes on specific fields

Character fields

Any fields that are stored with VARCHAR column types have theirmax_length restricted to 255 characters if you are using unique=Truefor the field. This affects CharField,SlugField.

TextField limitations

MySQL can index only the first N chars of a BLOB or TEXT column. SinceTextField doesn’t have a defined length, you can’t mark it asunique=True. MySQL will report: “BLOB/TEXT column ‘<db_column>’ used in keyspecification without a key length”.

Fractional seconds support for Time and DateTime fields

MySQL 5.6.4 and later can store fractional seconds, provided that thecolumn definition includes a fractional indication (e.g. DATETIME(6)).Earlier versions do not support them at all.

Django will not upgrade existing columns to include fractional seconds if thedatabase server supports it. If you want to enable them on an existing database,it’s up to you to either manually update the column on the target database, byexecuting a command like:

  1. ALTER TABLE `your_table` MODIFY `your_datetime_column` DATETIME(6)

or using a RunSQL operation in adata migration.

TIMESTAMP columns

If you are using a legacy database that contains TIMESTAMP columns, you mustset USE_TZ = False to avoid data corruption.inspectdb maps these columns toDateTimeField and if you enable timezone support,both MySQL and Django will attempt to convert the values from UTC to local time.

Row locking with QuerySet.select_for_update()

MySQL does not support the NOWAIT, SKIP LOCKED, and OF options tothe SELECT … FOR UPDATE statement. If select_for_update() is usedwith nowait=True, skip_locked=True, or of then aNotSupportedError is raised.

Automatic typecasting can cause unexpected results

When performing a query on a string type, but with an integer value, MySQL willcoerce the types of all values in the table to an integer before performing thecomparison. If your table contains the values 'abc', 'def' and youquery for WHERE mycolumn=0, both rows will match. Similarly, WHERE mycolumn=1will match the value 'abc1'. Therefore, string type fields included in Djangowill always cast the value to a string before using it in a query.

If you implement custom model fields that inherit fromField directly, are overridingget_prep_value(), or useRawSQL,extra(), orraw(), you should ensure that you performappropriate typecasting.

SQLite notes

Django supports SQLite 3.8.3 and later.

SQLite provides an excellent development alternative for applications thatare predominantly read-only or require a smaller installation footprint. Aswith all database servers, though, there are some differences that arespecific to SQLite that you should be aware of.

Substring matching and case sensitivity

For all SQLite versions, there is some slightly counter-intuitive behavior whenattempting to match some types of strings. These are triggered when using theiexact or contains filters in Querysets. The behaviorsplits into two cases:

  1. For substring matching, all matches are done case-insensitively. That is afilter such as filter(name__contains="aa") will match a name of "Aabb".

  2. For strings containing characters outside the ASCII range, all exact stringmatches are performed case-sensitively, even when the case-insensitive optionsare passed into the query. So the iexact filter will behave exactlythe same as the exact filter in these cases.

Some possible workarounds for this are documented at sqlite.org, but theyaren’t utilized by the default SQLite backend in Django, as incorporating themwould be fairly difficult to do robustly. Thus, Django exposes the defaultSQLite behavior and you should be aware of this when doing case-insensitive orsubstring filtering.

“Database is locked” errors

SQLite is meant to be a lightweight database, and thus can’t support a highlevel of concurrency. OperationalError: database is locked errors indicatethat your application is experiencing more concurrency than sqlite canhandle in default configuration. This error means that one thread or process hasan exclusive lock on the database connection and another thread timed outwaiting for the lock the be released.

Python’s SQLite wrapper hasa default timeout value that determines how long the second thread is allowed towait on the lock before it times out and raises the OperationalError: databaseis locked error.

If you’re getting this error, you can solve it by:

  • Switching to another database backend. At a certain point SQLite becomestoo “lite” for real-world applications, and these sorts of concurrencyerrors indicate you’ve reached that point.

  • Rewriting your code to reduce concurrency and ensure that databasetransactions are short-lived.

  • Increase the default timeout value by setting the timeout databaseoption:

  1. 'OPTIONS': {
  2.     # ...
  3.     'timeout': 20,
  4.     # ...
  5. }

This will make SQLite wait a bit longer before throwing “database is locked”errors; it won’t really do anything to solve them.

QuerySet.select_for_update() not supported

SQLite does not support the SELECT … FOR UPDATE syntax. Calling it willhave no effect.

“pyformat” parameter style in raw queries not supported

For most backends, raw queries (Manager.raw() or cursor.execute())can use the “pyformat” parameter style, where placeholders in the queryare given as '%(name)s' and the parameters are passed as a dictionaryrather than a list. SQLite does not support this.

Isolation when using QuerySet.iterator()

There are special considerations described in Isolation In SQLite whenmodifying a table while iterating over it using QuerySet.iterator(). Ifa row is added, changed, or deleted within the loop, then that row may or maynot appear, or may appear twice, in subsequent results fetched from theiterator. Your code must handle this.

Oracle notes

Django supports Oracle Database Server versions 12.2 and higher. Version6.0 or higher of the cx_Oracle Python driver is required.

In order for the python manage.py migrate command to work, your Oracledatabase user must have privileges to run the following commands:

  • CREATE TABLE
  • CREATE SEQUENCE
  • CREATE PROCEDURE

  • CREATE TRIGGERTo run a project’s test suite, the user usually needs these _additional_privileges:

  • CREATE USER

  • ALTER USER
  • DROP USER
  • CREATE TABLESPACE
  • DROP TABLESPACE
  • CREATE SESSION WITH ADMIN OPTION
  • CREATE TABLE WITH ADMIN OPTION
  • CREATE SEQUENCE WITH ADMIN OPTION
  • CREATE PROCEDURE WITH ADMIN OPTION
  • CREATE TRIGGER WITH ADMIN OPTIONWhile the RESOURCE role has the required CREATE TABLE,CREATE SEQUENCE, CREATE PROCEDURE, and CREATE TRIGGER privileges,and a user granted RESOURCE WITH ADMIN OPTION can grant RESOURCE, sucha user cannot grant the individual privileges (e.g. CREATE TABLE), and thusRESOURCE WITH ADMIN OPTION is not usually sufficient for running tests.

Some test suites also create views or materialized views; to run these, theuser also needs CREATE VIEW WITH ADMIN OPTION andCREATE MATERIALIZED VIEW WITH ADMIN OPTION privileges. In particular, thisis needed for Django’s own test suite.

All of these privileges are included in the DBA role, which is appropriatefor use on a private developer’s database.

The Oracle database backend uses the SYS.DBMS_LOB and SYS.DBMS_RANDOMpackages, so your user will require execute permissions on it. It’s normallyaccessible to all users by default, but in case it is not, you’ll need to grantpermissions like so:

  1. GRANT EXECUTE ON SYS.DBMS_LOB TO user;
  2. GRANT EXECUTE ON SYS.DBMS_RANDOM TO user;

Connecting to the database

To connect using the service name of your Oracle database, your settings.pyfile should look something like this:

  1. DATABASES = {
  2.     'default': {
  3.         'ENGINE': 'django.db.backends.oracle',
  4.         'NAME': 'xe',
  5.         'USER': 'a_user',
  6.         'PASSWORD': 'a_password',
  7.         'HOST': '',
  8.         'PORT': '',
  9.     }
  10. }

In this case, you should leave both HOST and PORT empty.However, if you don’t use a tnsnames.ora file or a similar naming methodand want to connect using the SID (“xe” in this example), then fill in bothHOST and PORT like so:

  1. DATABASES = {
  2.     'default': {
  3.         'ENGINE': 'django.db.backends.oracle',
  4.         'NAME': 'xe',
  5.         'USER': 'a_user',
  6.         'PASSWORD': 'a_password',
  7.         'HOST': 'dbprod01ned.mycompany.com',
  8.         'PORT': '1540',
  9.     }
  10. }

You should either supply both HOST and PORT, or leaveboth as empty strings. Django will use a different connect descriptor dependingon that choice.

Full DSN and Easy Connect

A Full DSN or Easy Connect string can be used in NAME if bothHOST and PORT are empty. This format is required whenusing RAC or pluggable databases without tnsnames.ora, for example.

Example of an Easy Connect string:

  1. 'NAME': 'localhost:1521/orclpdb1',

Example of a full DSN string:

  1. 'NAME': (
  2.     '(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=localhost)(PORT=1521))'
  3.     '(CONNECT_DATA=(SERVICE_NAME=orclpdb1)))'
  4. ),

Threaded option

If you plan to run Django in a multithreaded environment (e.g. Apache using thedefault MPM module on any modern operating system), then you must setthe threaded option of your Oracle database configuration to True:

  1. 'OPTIONS': {
  2.     'threaded': True,
  3. },

Failure to do this may result in crashes and other odd behavior.

INSERT … RETURNING INTO

By default, the Oracle backend uses a RETURNING INTO clause to efficientlyretrieve the value of an AutoField when inserting new rows. This behaviormay result in a DatabaseError in certain unusual setups, such as wheninserting into a remote table, or into a view with an INSTEAD OF trigger.The RETURNING INTO clause can be disabled by setting theuse_returning_into option of the database configuration to False:

  1. 'OPTIONS': {
  2.     'use_returning_into': False,
  3. },

In this case, the Oracle backend will use a separate SELECT query toretrieve AutoField values.

Naming issues

Oracle imposes a name length limit of 30 characters. To accommodate this, thebackend truncates database identifiers to fit, replacing the final fourcharacters of the truncated name with a repeatable MD5 hash value.Additionally, the backend turns database identifiers to all-uppercase.

To prevent these transformations (this is usually required only when dealingwith legacy databases or accessing tables which belong to other users), usea quoted name as the value for db_table:

  1. class LegacyModel(models.Model):
  2.     class Meta:
  3.         db_table = '"name_left_in_lowercase"'
  4.  
  5. class ForeignModel(models.Model):
  6.     class Meta:
  7.         db_table = '"OTHER_USER"."NAME_ONLY_SEEMS_OVER_30"'

Quoted names can also be used with Django’s other supported databasebackends; except for Oracle, however, the quotes have no effect.

When running migrate, an ORA-06552 error may be encountered ifcertain Oracle keywords are used as the name of a model field or thevalue of a db_column option. Django quotes all identifiers usedin queries to prevent most such problems, but this error can stilloccur when an Oracle datatype is used as a column name. Inparticular, take care to avoid using the names date,timestamp, number or float as a field name.

NULL and empty strings

Django generally prefers to use the empty string ('') rather thanNULL, but Oracle treats both identically. To get around this, theOracle backend ignores an explicit null option on fields thathave the empty string as a possible value and generates DDL as ifnull=True. When fetching from the database, it is assumed thata NULL value in one of these fields really means the emptystring, and the data is silently converted to reflect this assumption.

TextField limitations

The Oracle backend stores TextFields as NCLOB columns. Oracle imposessome limitations on the usage of such LOB columns in general:

  • LOB columns may not be used as primary keys.
  • LOB columns may not be used in indexes.
  • LOB columns may not be used in a SELECT DISTINCT list. This means thatattempting to use the QuerySet.distinct method on a model thatincludes TextField columns will result in an ORA-00932 error whenrun against Oracle. As a workaround, use the QuerySet.defer method inconjunction with distinct() to prevent TextField columns from beingincluded in the SELECT DISTINCT list.

Subclassing the built-in database backends

Django comes with built-in database backends. You may subclass an existingdatabase backends to modify its behavior, features, or configuration.

Consider, for example, that you need to change a single database feature.First, you have to create a new directory with a base module in it. Forexample:

  1. mysite/
  2.     ...
  3.     mydbengine/
  4.         __init__.py
  5.         base.py

The base.py module must contain a class named DatabaseWrapper thatsubclasses an existing engine from the django.db.backends module. Here’s anexample of subclassing the PostgreSQL engine to change a feature classallows_group_by_selected_pks_on_model:

mysite/mydbengine/base.py

  1. from django.db.backends.postgresql import base, features
  2.  
  3. class DatabaseFeatures(features.DatabaseFeatures):
  4.     def allows_group_by_selected_pks_on_model(self, model):
  5.         return True
  6.  
  7. class DatabaseWrapper(base.DatabaseWrapper):
  8.     features_class = DatabaseFeatures

Finally, you must specify a DATABASE-ENGINE in your settings.pyfile:

  1. DATABASES = {
  2.     'default': {
  3.         'ENGINE': 'mydbengine',
  4.         ...
  5.     },
  6. }

You can see the current list of database engines by looking indjango/db/backends.

Using a 3rd-party database backend

In addition to the officially supported databases, there are backends providedby 3rd parties that allow you to use other databases with Django:

  • IBM DB2
  • Microsoft SQL Server
  • Firebird
  • ODBCThe Django versions and ORM features supported by these unofficial backendsvary considerably. Queries regarding the specific capabilities of theseunofficial backends, along with any support queries, should be directed tothe support channels provided by each 3rd party project.