Offline storage SQLite React Native?

[mrflip]

Hi, I'd like to use the offline caching in React Native. I may do this on top of expo's SQLite support. It seems fairly straightforward, but before I dive in I'm wondering if anyone has already implemented a storage engine for SQLite or otherwise that will work in react native? I'd rather not use the async storage as that has size limitations.

Also, I'd be doing this primarily for speed up: our app is pretty read-heavy, it will be carrying around static catalogs with 10,000's of rarely changing objects and a small number of invoices. Would graphcache with offline support on top of a SQLite db let me read large parts of the catalog in, or is it more just for patching over small amounts of data while awaiting a cell signal?

Lastly, if I did implement this, is it something you'd consider pulling in to mainline URQL, or should I create a separate project for it? I'm assuming it would be about as long and about as straightforward as the single-file default storage engine is.

[kitten]

We haven't run a lot of experiments yet, so @StevenLangbroek may know more as he works on a project that uses more offline data than what needs be personally tested.

This comes down to our default storage engine. This engine keeps a limited set of data around for a limited set of time. There's no problem with persisting the data forever however as long as you take that into account. You can implement patterns like the requestPolicyExchange (which we're still working on) to let views automatically update their data.
The garbage collection will also give you appropriate signals to clear out unused values.

The only limitations that you'll encounter are:

• Will all used values (taking GC into account) fit into memory. I suppose yours must given a standard app use case.
• Will the data load in a reasonable amount of time from the SQLite engine? This may depend a lot on how you implement it.

So overall you should be ready to go, but I'd advise you to run some tests on how the storage behaves and what the load times look like, so you can optimise for those first.

[smontlouis]

@mrflip May I ask what you finally did ?