Am new to DataVault, our DV architect is vaulting product information. we get a flat product file from our source system. one of the attributes is the manufacturer name. ie Coke, Pepsi, etc. thats all we get, no other information about the manufacturer. Now our architect has created a main product hub/satellite with many data points in the satellite. but decided to create a separate hub/satellite for manufacturer. the natural key is the manufacturer name. furthermore, he only created a hub but no satellite stating that there are no attributes in the satellite so its optional. again the only piece of information we get about the manufacturer is the name. would you normally create a hub/satellite for something where you only get the name? i totally understand that sometime in the future, we may get more manufacturer data like address, inception date, etc. in this case the satellite concept makes sense. but we may never get manufacturer information, its not something we are expecting as of now. so should the manufacturer name just be an attribute of the product, or should it be in its own satellite. and note that its a many to 1, with many products having the same manufacturer.
sounds like a ref table and not necessary to build out as a hub+sat
I agree that for now this sounds like a Ref table. If in the future more information about the manufacturer becomes available, you can build out the Hub/Sat at that time. And you will need to refactor the FK joins to the Manufacturer Ref table as a Link to the new Hub.
From your perspective, what is the manufacturer? Is it a first class business entity? Or is it just an attribute on a [product] (business entity) that is provided by a [supplier]. If it is something that your business cares about, model it properly. If it is just an attribute then its reference data.
If it is something more fundamental then model it as a hub but push back to the source to get a better source file (if it is more fundamental, then it is also more likely that the sources will have better managed data around it)
I have a very similar problem, except the hub in question is product. For some source systems, I get a file that gives detailed product information, which of course goes to HUB_PRODUCT and it’s satellites. However, other source systems just have PRODUCT_TYPE_CD as an attribute of Account and nothing else, just like your Manufacture name situation. Ref table makes sense, but the hub already exists in my example, so I’m torn between allowing it to go to HUB_PRODUCT with no attributes, and just leaving it as an attribute of the account, or creating a ref table.
In this case the manufacturer is an attribute. Dont be “smart” and solve a problem that you dont have. Meaning that if you get additional data for the manufacturer then you deal with it, not before.