[PYTHON] High Performance Django --Model

I learned a lot from reading High Perfomance Django, so a memorandum.

Reduce the number of queries

• select_related
• prefetch_ralated

Please use this properly. During development, check the published query in the Django Debug Toolbar.

Shorter query execution time

• Set index properly
• Avoid big joins
• Stop all () on big tables and apply LIMIT
• Existence check is not spam.count ()> 0, butspam.exists ()

Foreign Keys Generic Foreign Keys are useful, but they sometimes throw queries with poor performance behind the scenes. Sometimes you should put your own foreign key.

cached_propety Cache is enabled for properties that are costly to execute & are called many times. https://docs.djangoproject.com/ja/1.9/ref/utils/

Handles small objects

• defer
• only
• values
• values_list

If the results you get are huge, use them to improve your performance.

Query cache

• Johnny Cache
• Cache Machine

These two were introduced in the book, but Jonny Cache has stopped development. (It seems that it can not be used with Django 1.5 or higher ...)

If you make your own, the Abstruct Cached Model introduced on gumi's blog will be helpful. Model with cache function in Django --gumi Engineer ’s Blog

Other

• Repli DB
• raw query
• Denormalization
• NoSQL
• Sharding

Other materials that I learned

• Pycon Jp 2014 --Django Application, Performance Tuning
• DjangoCon 2014- A Nice Problem to Have: Django Under Heavy Load
• DjangoCon 2014- Performant Django