Why do you need to build a data warehouse in analysis and decision-making instead of directly on the traditional database?

Transaction (traditional) database is not impossible, but you should not do statistics and analysis directly on the transaction database, because it will affect the speed of front-end business processors.

Can be exported separately for analysis and statistics;

The carrier of data warehouse is also the carrier of traditional database, similarly, such as ORACLE\SQLSERVER and so on. , you can use.

If you can model the data warehouse, that's good. You can find that you will find that multidimensional query statements are more convenient and faster than SQL statements.

The reason why the data warehouse can query faster is because the computer has calculated the possible ways of people's query in advance when people are sleeping, and calculated the results of most possible ways, and saved them directly to an address, and then people can query this address directly when they wake up.