Why large Django projects need a data (prefetching) layer with Flávio Juvenal - DjangoCon US 2022

Django

Learn how to improve performance in large Django projects by introducing a data layer, utilizing prefetching, annotations, and virtual models with Flávio Juvenal at DjangoCon US 2022.

Key takeaways

Django REST Framework‘s “don’t repeat yourself” principle can lead to complex code and maintenance issues.
Prefetching and annotations are crucial for performance and preventing n+1 queries.
Virtual models can help decouple query sets from serializers, making code more maintainable.
Django virtual models can be used to prefetch and annotate data, and even support custom functions.
The getOptimizedQuerySet method can be used to optimize query sets and prevent unnecessary prefetches.
Using mixins can help reuse code and reduce the need for nested serializers.
The serializer method field can be used to nest serializers and prefetch data.
Django readers can be used to project complex data and prevent n+1 queries.
The prefetch_related method can be used to prefetch related objects and prevent n+1 queries.
The annotate method can be used to add custom annotations to query sets.
The select_related method can be used to select related objects and prevent n+1 queries.
The deferred fields feature can be used to defer loading of related objects and prevent n+1 queries.
The Strawberry Django Plus project provides a more explicit way of handling prefetching and annotations.
The Django virtual models library provides a way to define virtual models and use them to prefetch and annotate data.
The prefetchers library provides a way to define prefetchers and use them to prefetch data.
The annotations library provides a way to define annotations and use them to annotate data.
The query_set method can be used to define a query set and use it to prefetch and annotate data.
The serializer method can be used to define a serializer and use it to serialize data.
The view_set method can be used to define a view set and use it to handle requests and responses.

Why large Django projects need a data (prefetching) layer with Flávio Juvenal - DjangoCon US 2022

More talks