Normalized data is "feature" of relational databases. No-SQL databases like Mongo don't have normalized data. That's what makes them fast. If you're concerned about normalized data, no-sql is not what you should be looking at.

>LOL. I had seen this before - I think someone on UT (Craig?) posted the link a year or so ago.
>
>Of course the bias is obvious and some of the 'facts' are wrong. The SQL proponent, for example, states that MongoDb relies on a 'fire and forget' approach to writing to the DB and, whilst it is true that this is one option, you can also ensure that the change has been committed (and not just committed but also , if required, replicated). And the ability of Mongo to auto-scale is a valid advantage.
>
>OTOH, I do have reservations about Mongo:
>(a) harder to 'normalize' data.
>(b) transaction support
>(c) online backup capabilites
>(d).......
>
>We're also looking at the MS Azure Table Storage which is similar......