Hey thrift users! If you are looking for a painless way to put thrift objects into a NoSQL database, check out a library I wrote called ZombieDB. It lets users efficiently store & retrieve thrift objects without requiring any database schema beyond the thrift definition itself. The main idea is that the field with tag number 1 is the primary key, and any fields with tag numbers in the range [2,16] are secondary indices with < O(N) lookup. If a field does not need a secondary index, assign it a tag number of 17 or greater.
As of version 1.2, the following databases are supported: HBase, Redis, Cassandra, JDBM, and BerkeleyDB. ZombieDB also has functions for writing/reading the database contents to a SequenceFile for backup or to create an offline copy for processing with hadoop. For more information, check out the github page here: https://github.com/MisterTea/ZombieDB Jason Gauci
