改善Django Rest框架中的分页和序列化时间

Question

I am using ListAPIView to render a list of objects. The response returns paginated results with limit=50 . The queryset size varies from 0 to a few hundred thousands.

What I have noticed is that when the queryset size is small, the API response time is small. But as the queryset size increases, the API response time becomes very large.

Is pagination actually the culprit here?
Is there a way to optimize the response time for ListAPIView s using pagination?

Answer 1

As stated in the official django documentation, querysets are lazy .

QuerySets are lazy – the act of creating a QuerySet doesn't involve any database activity. You can stack filters together all day long, and Django won't actually run the query until the QuerySet is evaluated.

Also, pagination compliments queryset's laziness .

Therefore, the culprit for your delay IS NOT pagination.
Case closed...

---

But what is it then?

• Processing the queryset before pagination . Hits the database multiple times, slowing the app down.
• Ordering as @sha256 suggests.
• Very costly queries perhaps (you can check that by following the docs ).

What to do to get faster?

• Data access optimization
• Solving performance problems in the Django ORM