Analyzing the data when it is large.

While i was trying to improve the performance of one of my project functionality, its not always good to drill the data from top to bottom where you may end up with so many transactions in the server.

When you got to know , in future that this data will be growing, then while implementing any of the functionality on that data , identify and consider the driving parameters to get the data and try analying it. So when dealing with large datas, its good you go retrive the data in denormalized way so that you get minimum data to be processed rather than large.