This is normally useful for lot's of web apps. Sort in Hbase is done at insert time not when scanning. Using a reversed timestamp you ensure the most recent activity of the user will be shown first.
-- View this message in context: http://lucene.472066.n3.nabble.com/Why-use-Reverse-Timestamp-as-the-Row-Key-tp3190719p3190906.html Sent from the Hadoop lucene-users mailing list archive at Nabble.com.