On-demand sync + joins in TanStack DB: how to efficiently fetch subsets per collection?
Hey folks 
I couldn’t find a dedicated channel for TanStack DB questions, so posting this in No Access — hope that’s okay.
I’m playing around with TanStack DB, and I’m struggling to understand the proper way to fetch data inside a collection’s
From the docs and examples, it looks like the intended approach is to fetch all data for a collection in a single request. Most examples do exactly that. That obviously doesn’t scale well for large datasets, so I tried switching to
This works fine in simple cases, but I’m hitting a wall once joins are involved.
What works
If I have a liveQuery that reads from a single collection, everything makes sense:
- props.meta.loadSubsetOptions is populated
- I can parse it and forward it to my API
- I can efficiently fetch only the required subset
What breaks with joins
Once I introduce joins:
- loadSubsetOptions becomes empty
- There’s no metadata passed to collections about joins
- I have no way to know what subset of data is actually needed for that live query
- This makes it impossible to efficiently fetch a small subset from the server
Any guidance or pointers would be hugely appreciated
I couldn’t find a dedicated channel for TanStack DB questions, so posting this in No Access — hope that’s okay.
I’m playing around with TanStack DB, and I’m struggling to understand the proper way to fetch data inside a collection’s
queryFnFrom the docs and examples, it looks like the intended approach is to fetch all data for a collection in a single request. Most examples do exactly that. That obviously doesn’t scale well for large datasets, so I tried switching to
syncMode: "on-demand"This works fine in simple cases, but I’m hitting a wall once joins are involved.
What works
If I have a liveQuery that reads from a single collection, everything makes sense:
- props.meta.loadSubsetOptions is populated
- I can parse it and forward it to my API
- I can efficiently fetch only the required subset
What breaks with joins
Once I introduce joins:
- loadSubsetOptions becomes empty
- There’s no metadata passed to collections about joins
- I have no way to know what subset of data is actually needed for that live query
- This makes it impossible to efficiently fetch a small subset from the server
Any guidance or pointers would be hugely appreciated
