OLAP is currently a hot BI method, but in fact, the success rate is not high, and users have spent a lot of money to analyze modeling design to purchase front-end tools, but the result is often not ideal.
In fact, it is also very simple.
The implementation of OLAP needs to analyze the industry business, causing data cubes beforehand before you can do query analysis. For the failure to be expected, there is no corresponding cube available, OLAP is also impossible. For example, we asked a certain institution for three years of employees who have been rated as advanced workers in five consecutive years. In general, there must be this information in the general personnel database, but if there is no such data cube, use OLAP can't answer this personnel. The agency is easy to ask and have no problems, or require the technical department to edit a program or write a complex SQL to make.
In fact, the model principle of OLAP can only deal with the data that has been cified, that is, solves the problem of a single table query change problem, and most of the multi-table queries are difficult to deal with, how can a single table query? . OLAP's most difficult multi-table change single work is transferred to modeling work, which requires business experience to complete, and the threshold is highly efficient, so this is hard.
However, for domestic business, even if this is done, it is still not. China's political and economic rules are in constant changes, we can't move directly to the current model of the United States (then there are people to rebellious), and they can't stop. (Then my country must be "expeling the ball") It can only be reforms step by step. In this case, it is difficult to have a strict industry expert, which cannot predict the possible changes, and the model built is basically unstable, and it has changed again. The implementation of OLAP is seriously depends on industry modeling. In this way, use OLAP solutions, this is impossible.
In fact, it is the case, the business system is the same, the domestic application system often starts maintenance of patch from the day of the year, and then pushing back in two or three years, it is not the ability to develop business, it is not possible It is predicted that all the colleagues who have experienced project experience must have the same feeling. Unlike the United States, the business is relatively stable, and many industries have used systems written in COBOL language several decades ago. BI for the United States will inevitably defeat.
Poor business intelligence, when can you smart?
