It seems KDB excels at queries, joins, and aggregates on large datasets. I have limited experience with redis, but if that type of code would need to be written in c/python with redis then KDB may have a performance and productivity lift. I would like to see an example like that.

Another benchmark is here: http://kparc.com/q4/readme.txt