Having decided that you need a Data Warehouse, the first decision you need to take is what type of design/database you are going to use. There are quite a few options (Kimball, Inmon, Data Vault, NoSQL, Graph, etc.) but the vast majority of data warehouses follow the basic Kimball Methodology of dimensional modelling e.g. Facts and Dimensions.
If you are going to build a Kimball-style data warehouse (or follow any other methodology) then my first recommendation would be to employ someone with experience who can lead the work. It is very easy to make mistakes when designing a DW but very hard to correct them once people are using it, have built reports against it, etc.
If you're not going to employ someone who knows what they are doing then the next best option is to go on a course and/or read books on the subject. For Kimball, there are really 2 books that should be required reading:
- The Data Warehouse Lifecycle Toolkit : this talks you through all the components involved and the steps to follow in order to deliver a robust data warehouse
- The Data Warehouse Toolkit : this goes through the steps to design a dimensional model
Once you have read and understood these 2 books you will be better placed to understand the terminology and ask specific, focussed questions about any parts of the methodology (or your specific circumstances) that you don't understand.
This is absolutely not meant to be a criticism but from your questions it is very clear that you don't (yet) have the knowledge or experience to be designing and building a data warehouse - and you're not going to be able gain that experience by asking questions on this (or any other) forum.
与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
