Django from queryset to serialization with Iuri de Silvio - DjangoCon US 2022

Django Testing

Django expert Iuri de Silvio shares techniques for optimizing queryset to serialization, including custom serializers, `select_related` and `prefetch_related`, and testing query counts to achieve scalable solutions.

Key takeaways

The speaker recommends creating a custom serializer to encapsulate the serialization process and avoid repetitive and error-prone code.
The goal is to have a single query that loads all the related objects, rather than having multiple queries for each object.
The speaker uses the select_related and prefetch_related methods to solve this problem.
A custom manager can be created to encapsulate the serialization process and make it easier to test and maintain.
The speaker recommends testing the number of queries being executed and iterating until the desired number is achieved.
Django Outprefetch is recommended but does not solve the problem, it just hides it.
The speaker discusses the importance of encapsulation and how it can make the code more manageable.
The speaker also mentions that the problem of n+1 queries is common and that many projects have it.
Pydantic is not directly implemented in the library but is discussed as a potential tool to format the data.
The speaker’s approach is to create a single query that loads all the related objects and then serialize the data.
The speaker’s custom serializer looks like a class and has a method to serialize the data.
The speaker hopes to link loaded data with serialized data in the future.
The speaker also discusses the importance of testing and iterating on the code to achieve the desired results.
The speaker mentions that some of his previous projects were struggling with scalability and that this solution may be helpful.
The speaker also recommends considering the context-dependent and enterprise-level solution to serialization.

Django from queryset to serialization with Iuri de Silvio - DjangoCon US 2022

More talks