I have a Django 1.7 project that uses Postgres 9.3. I have a table that will have rather high volume. The table will have anywhere from 13million to 40million new rows a month.
You can use Architect application for Postgresql Table Partitioning Django Project
PostgreSQL’s partitioning implementation in Architect is done purely at the database level. That means that Architect creates several triggers and functions and inserts them directly into the database, so even if you issue direct insert statement from database console and not from the ORM, everything will work as expected and record will be inserted into the correct partition, if partition doesn’t exist, it will be created for you automatically. Also partitions may be created in any order and not only from lower to higher.
It's new version of old Django DB Parti application
As long as you use inheritance, and then only connect the parent table to your Django model, the partitions should be entirely transparent to Django. That is, a SELECT
on the parent table will cascade down to the partitions, unless the ONLY
keyword is explicitly used (where applicable).
Note that partitioning does add complexity in terms of needing to implement a programmatic method of determining when new partitions need to be created, and then creating them -- or doing this manually at certain intervals. Depending on your exact data and business logic, it is quite likely you may also need to implement triggers and rules to determine which partition to, say, INSERT
something into (since you wouldn't want to INSERT
into the parent table). These, too, should be abstracted from Django, however.
I have found that, depending on the exact circumstances, this may need to be done with your main apps shutdown, lest the new partition creation cause a deadlock.
Also worth considering is whether you need true partitions which get created over time, or if an inheritance model of, say, tables foo
and foo_archive
would suffice, where foo_archive
inherits from foo
, and periodically something (e.g. a script) moves older data into foo_archive
to keep foo
smaller.
If you are using newer version of PostgreSQL, you can try this
https://github.com/chaitin/django-pg-timepart
A Django extension that implements PostgreSQL tables for partitioning and management based on dates.