It has been talked about in the past, see https://issues.apache.org/jira/browse/CASSANDRA-4175 for example. However with https://issues.apache.org/jira/browse/CASSANDRA-8099, the duplication of column names is gone. So once you’re on Cassandra 3+, this optimization is a lot less valuable.
> On Feb 11, 2016, at 4:19 PM, Bhuvan Rawal <bhu1ra...@gmail.com> wrote: > > Hi, > > We are modelling schema for database revamp from mysql to Cassandra. It has > been recommended in several places that column names must be kept as small > as possible to optimise disk storage. > > I have a doubt here, why can't we map column names and store it as an > index, say in memory. I mean, make column name really small human > unreadable and store it in disk but map it with real column while > querying. That way one can go ahead with readable column names . > > Let me know if I can go ahead and create a jira for the same > > Regards, > Bhuvan
signature.asc
Description: Message signed with OpenPGP using GPGMail