First of all, thanks for the excellent documentation.
Secondly, I have a doubt at http://www.datastax.com/docs/1.1/ddl/column_family
At the "Timeline Table" part, it says:
"The storage engine guarantees that the columns are clustered according to the partition key (tweet_id). ... "
but, isn't tweet_id the remaining key instead of the partition key?
Thanks for your help