I always wondered if using SQLite for the back end for a distributed map/reduce jobs was efficient. Each machine holds part of the data in an SQLite file. It w

I always wondered if using SQLite for the back end for a distributed map/reduce jobs was efficient. Each machine holds part of the data in an SQLite file. It would not solve the usual sort/group by problems that require cross-machine communication, but would take full advantage of SQLite's optimizations for other problems.