Not using cache / fetching on every query

Thanks for this really cool project, I'm having a lot of fun playing around with it. I might be misunderstanding something completely, but from the readme I assumed that once pages were downloaded, they'd be stored in a (shared) cache, allowing fast access afterwards. So I would expect that if you once would fetch the full data set `select * from data` and after that just order it `select * from data order by id`, the second query wouldn't result in any (significant) new data fetching. But in my simple tests almost every query resulted in lots of data fetching. Why is that the case?