Are partial results already possible with Houdini?

Hi @AlecAivazis,

I just discovered this wonderful project. Endless congratulations.

I’m currently using Urql on all projects but as you may know Urql and Svelte don’t get along very well (no SSR, broken (Svelte) API and others).

I haven’t used Houdini yet (I’ll try in the next few hours), but one thing that keeps me from changing is the following and from a quick read it doesn’t seem to me that with Houdini you can get it today: I HOPE I AM WRONG.

This is what the world calls: Partial results which Urql obtains with Schema awareness.

This is one of the biggest points in favor of solutions like urql or apollo for me. And that’s one of the reasons I really enjoy working with Graphcache (normalized data) and schema awareness. Although it makes everything much heavier (in terms of development and runtime).

The UX show in the hands of the user is that the application looks native and almost offline (if the internet is fast enough). The loading spinners disappear completely and everything seems immediate.

Ultimately that’s why I chose to complicate my life by choosing to develop single page apps (SPAs) instead of multi-page (server rendered) apps (MPAs).

A very simple example is the list > detail. As soon as I enter the details I can show the data that I have already shown in the list and download in the background (almost without the person noticing, especially if I do it in the prefetch) the data that did not appear in the list.

I know that you understand what I mean and I also know that it is difficult to understand well how to organize this thing.

Is it already possible today with Houdini?

If you answer me yes, I could pass out with joy… 😄

[AlecAivazis]

Yea, for now it’ll be all data but we can get smarter about that with time.

I’m not against having a CacheOnly policy for completeness. It’s one of those things that I probably wouldn’t use very often but it doesn’t hurt anything to have it.

[AlecAivazis]

@frederikhors, I just opened #240 which adds support for partial results by extending the cache policies as I described above. There is still some more testing to do but it should be ready relatively soon.