On Fri, Feb 25, 2011 at 9:25 PM, Simon Metson <simonmet...@googlemail.com> wrote: > And query it for a specific article with > > http://localhost:5984/voting/_design/vote/_view/vote?group=true&key=%22foo%22 > > Mike Miller did a nice summary of how you could do time binned data for your > historical info (using the timestamp field in the vote doc) on > http://nosqltapes.com/video/understanding-mapreduce-with-mike-miller > > HTH > Cheers > Simon
So each comment and rating is a single document. And grouping should be used to reduce. Thank you very much Simon. I got the idea now. -- Best regards, He Shiming