Why doesn't packages/powersync-db-collection implement a synchronization mode like on-demand?

[JinJieBeWater]

I've been wanting to integrate TanStack DB with Powersync, but I've always had a question: are there any issues with large table sets?
The current strategy of powersync-db-collection seems to be to fully synchronize all data from the tables into the collection, without implementing on-demand. I think this may have performance issues in scenarios with large amounts of data.
Although TanStack Query also integrates with Powersync, I think the user experience of TanStack DB is excellent, and I don't want to mix the two mechanisms in one project.

[cahofmeyr]

Thanks for the feedback! This is something the PowerSync team is currently working on: powersync-ja#6