Consultant and former CIO, I put a point of honor that people rather think twice before scheduling a, knowing that the use of computers wasted a lot more time and money reformulate on paper or using a model that the user desires. The more time passes, the more I asked a "test game", an example of desired data transformation, since few people understand data models, graphical representation of the concepts of the user, valid whatever the volume data, invoice or writing representing the same concept regardless of company size. Interested in this concept of Big Data, wanting to know what it alters this approach, I fall of the cabinet must be reading "discover" what is in the data, then "produce" that is, say exploit !!!! In short, the opposite of common sense. Why? to save costs, make faster and better decisions and innovate. We can not do worse waffle .... The worst is that the author admits that most of the work is to structure the data. Obviously he knows nothing of the subject, confounding basic hierarchical and relational data, masking his ignorance by empty words like "With a traditional DBMS, data is expected in the form of rows and columns of values, while large data comes in different formats. " Chapter 5, written by a person of SAS, brings some professionalism in approaching the SQL. Unfortunately, it refers to developpez.com to make us understand the ETL software, data translation. I suggest the author to go to this site to check the books I wrote it 20 years ago on the structuring of data. This will allow him to think before promoting an Nth thing to sell to "decision makers" who will bring them nothing if they did not think before. What function to cover? What project? Which user project manager? What expense? What return on investment?