Why select related field doesn't have any effect to invalidate queries?

[angryfoxx]

today i realized that situattion on my project, but why? why doesn't django-cacheops invalidate the query?

[Suor]

This is a design choice.

Making queries when cache is used and when not should follow different strategies. The most obvious thing is that you usually want smaller number of queries when not using cache but more granular ones and hence bigger number of them when you do. The reason is a query becomes stale and should be invalidated as fast as ANY object in its result changes, so splitting queries into smaller parts so that only one/some of them will be invalidated on any particular change is a good idea.

So using .select_related() doesn't make much sense when a query is cached, one should use .prefetch_related() instead. To return some sense to it cacheops does not invalidate on those related objects, this way by caching a query with .select_related() you say that you don't care about those related object being a bit stale.

[angryfoxx]

is there any way to activate invalidation on select_related? can i do it ? and is there any bad efffect to that?

ngl i couldnt understand correclty the problem, could you elaborate the design choice and why did you choose that?

[Suor]

The simplest way it's using .prefetch_related() instead of .select_related(). If you want things to be invalidated for select related objects then you'll need to create and maintain a modified cacheops version. This is doable, you will need to supplement conditions tree calculation somehow.

On not understanding - not sure what I can add to the above

[angryfoxx]

in actually i wonder why you choose that design choice is there any foundation on computer science