Indexing JSON in Cassandra
The Java Zone is brought to you in partnership with ZeroTurnaround. Check out this 8-step guide to see how you can increase your productivity by skipping slow application redeploys and by implementing application profiling, as you code!
Cassandra has native indexing capabilities, but those capabilities only work if the values stored in your columns are the values you want indexed. If the data is structured in some way (e.g. JSON or Protobuf), it is impossible to leverage Cassandra's indexing capabilities out of the box.
A while back we had a discussion on the dev list regarding indexing attributes stored in columns in Cassandra. I think Jermiah Jordan had the best description of the problem. Lately, we found ourselves with the same problem so we ended up building it into our cassandra-indexing extension.
Our cassandra-indexing module, is an AOP-based extension to Cassandra. Drop the jar file in the Cassandra lib directory, update the start script to include the extension and the aspect will take care of indexing information as you mutate the data in Cassandra.
To support indexing a specific field within the column value, we added the ability for a user to specify a field to index (not just a column name). We then parse the JSON document as it is written in the mutation, extract the values and create entries in the wide-row index for that column family.
Specifics for the new JSON-based indexing configuration can be found on the wiki. Admittedly, we are still light on documentation. If you have any trouble using the new capability, let us know.