Hacker News Reader

What's the SQL for the second optimized tree?

It's the same. That's the point :). You need to plan your queries much differently when your data is not in a single node.

I don't think it's the same. The filters have to be done in separate queries on both the tables, and then the results the joined.