Great article on "Schema-less" MySQL

18 Feb 2010

Just dug up this nugget: How FriendFeed uses MySQL to store schema-less data

This is similar to the postgres feature of h-stores

I haven't had the need to index the unstructured data, but I have used a blob to store key-value pairs in a JSON form. Why JSON? With cJSON modules it's as fast if not faster than cPickle,and it's text so a bit easier on the eyes. Also there is a possibility of just slamming it to the client as-is