On Fri, Mar 30, 2012 at 6:17 PM, idmartin <[hidden email]> wrote:
> Is there anyway to analyze Riak objectes to make sure there isnt duplicates
> or near duplicates?
A M/R Job? Identity map, followed by a reduce that does the
comparison? Yeah, it won't be super-efficient, but then a comparison
like that never is, and you probably want to define what "near
The leveldb back end is not as fast as the default (bitcask), but leveldb does not keep all keys in memory. So for a very large or unbounded set of keys leveldb is superior. Leveldb stores values sorted by key, which also lets riak speed up certain operations such as listing keys in a bucket.