I have referred to a lot of materials before, and its main core idea is that DV(Data Vault) is a set of modeling theory combining ER and DM. In fact, it is better to be the ER group than to find a middle zone, both of which can be considered. Recently, I saw the explanation about DV in Wikipedia, which mentioned DM: For this purpose, the hubs and related satellites on those hubs can be considered as dimensions and the links and related satellites on those links can be viewed as fact tables in a dimensional model. In short, dimension table is the satellite table of Hub and Hub, while fact table is the correlation table and related satellite table. I always feel that it is a little reasonable, but I can't deny the model concept of DV. From the book "Big data Road" published by Alibaba, we can see that DM has evolved. Dimension tables have been used as auxiliary tables, and dimension table information has been redundant in fact tables. In other words, it seems that the same evolution mode can be seen in ER, with fewer topics. In each topic, table data of the 3NF type is no longer used, but through redundancy, Some commonly used table fields are formed. Some of the fields that need to be associated through a link history zipper table can be used to obtain data. Now, due to redundancy, they also become auxiliary tables, however, the auxiliary table is only for this topic. In other topics, it may become an associated table. That is to say, for DV, the existing modeling form of data warehouse conflicts with the concept, because DV is still based on the two basic models of ER and DM, for the evolution model of ER and DM, the Hub defined in DV is completely different, but it cannot be completely abandoned. As long as we change the definition, the original DV framework will continue to be adopted. How to define the Hub? In my opinion, this part of table data is redundant, because we often use data, we will be redundant, whether it is the DM redundant part or the ER redundant part. It should be emphasized here that the redundant parts remain their own characteristics, that is, the entity and relationship, dimension and fact table can still be seen. Satellite table seems to be the auxiliary table of the two models mentioned above. Let's talk about Link again. In my opinion, it only plays a role in decoupling, that is, aiming at the phenomenon of many-to-many. It seems that a new model combining DM and ER has a framework. Next time we will talk about this model. Welcome to ask more questions and discuss together.
Start Building Today with a Free Trial to 50+ Products
Learn and experience the power of Alibaba Cloud.
Sign Up Now