> > The rows could be named and partitioned by date/time, which can be known in > advance. For example, '!public!20100318' could contain the public timeline > for that day. >
Yes, I thought of doing this. Then I realized there'd be boundary cases, on the start of a new day, where it'd be best to query for both the new day and the old one. In the end, since I wanted the project to be simpler to understand, I decided to keep the simpler approach. Perhaps I will add a comment noting that this isn't a particularly scalable solution, and suggest the day barrier. Also if my understanding is correct, CASSANDRA-674 will help alleviate (although it won't solve) the problem. -Eric
