Can someone provide me with some resources on when to de-normalize data? If possible could someone give me three reasons on why to denormalize data. I am aware that normalized data can have many joins and this affects performance. Basically, I would like something that is a bit more formal.
Denormalizing data might make sense in a datawarehouse application because the data doesn't change, and it can speed up queries, but I would never do it in a OLTP application. The beauty of normalized data is that no data is duplicated. Just my 2 cents. As for references go to google and put in the following keywords
normalized data denormalized
You do not need a parachute to skydive. You only need a parachute to skydive twice.