A data warehouse recipe book - recipe 1, the basic mixture
Ingredients:
1 data source 1 database to store the data warehouse 1 reporting layer (optional)
Method:
Add data source to database, mix well. Decorate with optional reporting layer.
Sounds simple, and not a Kimball or Inmon in sight! But as with all great recipes it’s the quality of the ingredients that counts.
The database: I only “cook” with Oracle. Not because it is the “best” database (whatever that may mean) but because I know Oracle, and what it can do for me. Other enterprise quality databases can be used but the techniques required to exploit them to their full potential varies between the vendors. If you want to know the Teradata way to do something, I’m not going to be your best source of information.
The data source: Traditionally, this is one (or quite possibly more) remote system(s); it is most unlikely that your data warehouse will perform well if it is in the same database instance as the source system. On previous posts I mentioned the two fundamental classes of DW data, “fact” and “reference” or dimensional data. How these two data classes arrive in the DW is a big topic and worthy of a post (or two) of their own
Soon - data sources (or is it sauces)