django-dbarray 0.2

Requirements

• PostgreSQL

• psycopg2

• Django >= 1.2

The ARRAY db type is PostgreSQL specific, so these model fields currently work only on PostgreSQL with psycopg2. A FieldError will be raised if the field is used with another database.

This package has been tested with Django versions 1.2 through 1.6, inclusive.

Fields

The field types defined in the dbarray module are listed in the table below, along with each field’s parent class (from django.db.models), and the data type of the postgresql column it will create.

Each field takes the same arguments as its parent class.

Custom Fields

To define a new array field for a base field type FooField:

import dbarray

class FooArrayField(dbarray.ArrayFieldBase, FooField):
    __metaclass__ = dbarray.ArrayFieldMetaclass

This may or may not work depending on a few factors. You might, for example, need to override the db_type method to make it put the [] in the right spot in the column type used in generated SQL.

Another issue that may arise when performing lookups with array fields is that PostgreSQL may get the query parameter as a data type that isn’t compatible with the type of the db column (for example text[] instead of varchar[]). Then you will get an error like the following:

DatabaseError: operator does not exist: character varying[] = text[]
LINE 1: ... FROM "dbarray_chars" WHERE "dbarray_chars"."arr" = ARRAY['A...
                                                             ^
HINT:  No operator matches the given name and argument type(s). You might need to add explicit type casts.

You can rectify this by setting cast_lookups = True on your ArrayFieldBase subclass, which instructs it to add an explicit type cast to the SQL query.

Look in the source code for more examples of how to handle issues with new ArrayField types.

Version History