The Best Foot Forward needs now to define actions to undertake in the future. The previous examples are not sufficient, we need to analyze the evolution of already filed data, and if possible to anticipate their evolution in the future. 数据挖掘研究院
OLAP databases are very powerful for that, they provide functions allowing to manipulate temporal inquiries. For example, we could know what week corresponds to 12 June 2000, how much days are between 06 May 1998 and 15 June 2002, or if 29 February 2000 exists or no.
Some engines provide standard datatypes for time dimensions. These dimensions are then characterized by their periodicity (day, week, semester, . ..) and functions exist to convert a periodicity to another. The OLAP engine include all the calendar specifications, and developers no longer worry about complex algorithms, building and maintaining temporal hierarchies.
With small and simple programs, we can now define formulas replying to questions like this: 数据挖掘研究院
- What is the evolution of my net profit in comparison with the same month of the last year.
- What is the evolution of my net figure in comparison with the average of the last three months.
- What will be the market trend for the next 12 months.
With relationnal databases, such questions will require very expensive programs, perhaps never giving a response to the user...
