django-mysql-fuzzycount

Django MySQL Fuzzycount

When using MySQL, counting all rows_ is very expensive on large InnoDB tables. This is a replacement for QuerySet that returns an approximation if COUNT(*) is called with no additional constraints. In all other cases it should behave exactly as QuerySet.

This only works with MySQL and behaves normally for all other engines.

Installation

Install the package django-mysql-fuzzycount from PyPI using pip_::

$ pip install -U django-mysql-fuzzycount

Usage

There are a couple ways to use the FuzzyCountQuerySet.

You can import and use the provided FuzzyCountManager on your Django models::

from django.db import models

from mysql_fuzzycount.managers import FuzzyCountManager

class Choice(model.Model):
    objects = FuzzyCountManager()

    # ...

Then, doing a count on the Choice model, without any constraints, will approximate the total count::

>>> Choice.objects.count()  # approximation
100
>>> Choice.objects.filter(votes__gt=10).count()  # not an approximation
28

Another common issue is counts in the admin for a model. There is a base ModelAdmin class that you can subclass in your admin.py files to prevent expensive COUNT(*) queries upon loading the admin page. In an admin.py for one of your models::

from django.contrib import admin

from mysql_fuzzycount.admin import FuzzyCountModelAdmin

from myapp.models import Choice


class ChoiceAdmin(FuzzyCountModelAdmin):
    pass

admin.site.register(Choice, ChoiceAdmin)

Now, when you load the admin page for the Choice model, the count for pagination will be approximate.

Testing

It has been tested in production, but there are no unit or integration tests.

License

Copyright © 2013, Educreations, Inc under the MIT LICENSE.

.. _counting all rows: http://www.mysqlperformanceblog.com/2006/12/01/count-for-innodb-tables/ .. _pip: http://www.pip-installer.org/