On Thu, 2011-02-03 at 15:35 -0800, Mike Malone wrote:
In my dealings with the Cassandra code, super columns end up making a mess all over the place when algorithms need to be special cased and branch based on the column/supercolumn distinction.
I won't even mention what it does to the thrift interface.
My observation is similar, in that they (SCFs) make the "type system" in Cassandra disjoint. This makes me doubt that moving to Avro would simplify anything for Cassandra users. It also means knock-on effects such as no common supertype in APIs for languages like Java (so the surface area of clients like Hector blow up badly when you compare it the HBase client). I can't wait to see how CQL fares with SCFs; a sane query language will be closed under its operations and I doubt it can be done atm.
That said, I keep finding uses for them, which is irksome; but maybe I'm being lazy when it comes to modelling and now that secondary indexes are in, I should pretend SCFs don't exist.