MapReduce Views performance

Each view is backed by it’s own file per node, and there’s a lot of processing that can be amortized if you can consolidate it to a single view. What I might recommend is process it all in the same view logic and use compound keys to have different ranges in the same view. Then you’ll be addressing different parts of the same B+Tree, but not have many of them.

I do have to say, 10,000 is a lot though, so there might be a better solution depending on what you’re trying to do. Maybe you could describe why you think you’ll have so many?