django-greenplum-updated 0.1.2

No project description provided

Django-Greenplum

Django ORM backend for Greenplum — a massively parallel PostgreSQL-based database.

Originally authored by souno.

---

Requirements

• Python 3.8+
• Django >= 3.0
• psycopg2 (or psycopg2-binary on Linux)

---

Installation

You can install the package via pip or uv:

Using uv (recommended)

uv add django-greenplum-updated

DATABASES = {
    'default': {
        'ENGINE'  : 'django-greenplum',
        'NAME'    : 'mydb',
        'USER'    : 'gpadmin',
        'PASSWORD': 'secret',
        'HOST'    : '127.0.0.1',
        'PORT'    : '5432',
    }
}

---

Partitioning

Greenplum supports table partitioning at the storage level. This backend lets you declare the partition strategy directly on your Django model as a class attribute.

Import the partition class and assign it to greenplum_partition:

from partition import RangePartition, ListPartition

Range partition — automatic (EVERY)

Splits the table into equal sub-partitions over a continuous range.

class Orders(models.Model):
    order_date = models.DateField()
    amount = models.DecimalField(max_digits=12, decimal_places=2)

    greenplum_partition = RangePartition(
        column='order_date',
        start='2020-01-01',
        end='2026-01-01',
        every='1 month',  # string → INTERVAL '1 month'
        # integer → plain numeric step
    )

Generated SQL suffix:

PARTITION BY RANGE ("order_date")
(START ('2020-01-01') INCLUSIVE END ('2026-01-01') EXCLUSIVE EVERY (INTERVAL '1 month'))

Range partition — manual named partitions

class Events(models.Model):
    event_date = models.DateField()
    description = models.TextField()

    greenplum_partition = RangePartition(
        column='event_date',
        partitions=[
            ('p2023', '2023-01-01', '2024-01-01'),
            ('p2024', '2024-01-01', '2025-01-01'),
            ('p2025', '2025-01-01', '2026-01-01'),
        ],
    )

List partition

class Sales(models.Model):
    region = models.CharField(max_length=20)
    revenue = models.DecimalField(max_digits=14, decimal_places=2)

    greenplum_partition = ListPartition(
        column='region',
        partitions=[
            ('north', ['north', 'northeast']),
            ('south', ['south', 'southeast']),
            ('other', ['west', 'east']),
        ],
    )

RangePartition parameters

Parameter	Type	Description
column	str	Model field name to partition by
start	value	Inclusive start bound (required with every)
end	value	Exclusive end bound (required with every)
every	str or int	Step for automatic partitions ("1 month", "1 year", 10, …)
partitions	list of 3-tuples	Manual named partitions (name, start, end)
start_inclusive	bool (default True)	Whether the start bound is INCLUSIVE
end_exclusive	bool (default True)	Whether the end bound is EXCLUSIVE

every and partitions are mutually exclusive.

ListPartition parameters

Parameter	Type	Description
column	str	Model field name to partition by
partitions	list of 2-tuples	(partition_name, [value, value, …]) for each partition

---

Notices

1. The primary key is used as the distribution key by default. Greenplum does not support multiple unique keys on the same table — avoid defining unique=True on non-primary-key fields.

2. Because of the above limitation, the default Django admin and auth applications may not work as expected.

3. Partition child tables are automatically excluded from Django's model introspection — only the parent (root) table is visible.