NoSQL is cool. Cool as in “Throw any documents you like in here and we’ll store it for you!” It makes complex relational schemas a snap, since each entity you store in the database can have whatever fields you like and even coexist with other entities that may not be of the same type. This means that instead of spending lots of time solving a problem with a relational schema, database designers can organize information by purpose or intent. The problem, however, is that to run a query on the database to get, for example, the average salary of the employees in the system would require using a mapreduce function - essentially using a sledge hammer when a small mallet will do.
Well, version 2.2 just arrived and has brought the querying benefits of relational SQL to the NoSQL world. I’m beginning to suspect that, as Oracle continues to absorb and dismantle MySQL, PostgreSQL will become more popular as well as MongoDB and other NoSQL solutions. One of these days, NoSQL will be a very mainstream database option, and today’s release is definitely a step in that direction.