3 ms·
It seems like a pretty straightforward MapReduce, though I agree that if you're doing searches repeatedly or your data is small enough you should use a database
by iskander 14y ago
It seems like a pretty straightforward MapReduce, though I agree that if you're doing searches repeatedly or your data is small enough you should use a database.
(map = search a partition and return top k results, reduce = combine multiple n*k result lists into a single result list)