DataWarehousing

Possibly with an Agile flare?!

The Tenth Dimension

http://www.tenthdimension.c... The tenth dimension is an interesting concept that is being put together by author Rob Bryanton. This video below is an interesting representation of the concept of dimensions. We are in the third dimension. If you follow and believe the concepts presented below, a Data Warehouse would be in the concept of the 4th dimension (time as points). We would be at the 5th dimension (a branch of any possible moment) when we look at the Data Warehouse. You can see ......

Data Warehouses Should Be Granularity Agnostic

When you are designing a data warehouse solution that publishes out data marts, always be sure that your warehouse is granularity agnostic. Granularity means the level of detail you track (defined more in depth below). Being agnostic means that you don't care about or have knowledge of something. To be "granularity agnostic" means that your warehouse doesn't care or have knowledge of how granular your data is. In essence, you design it up front so that it can get as much or as little detail as possible. ......

Data Marts Slowly Changing Dimensions

In my last post, I put in the diagram a Type 1 and a Type 2 insert. I think I need to elaborate on that just a bit. Type 1 updates don't have any concept of history. Type 2 updates are what I use to encapsulate all history types (Type 2, Type 3, Type 4, and Type 6). These dimensions are where Ralph Kimball's dimensional modeling come in. So before anyone points out that I didn't talk about the other types in my diagram, Type 2 updates for me is synonymous with all history in a data mart ......

Data Warehousing, Data Mart Dimensions and Fact Table Growth

NOTE: This is up for discussion. I am not an expert in Data Warehousing (yet). I hear from my Data Modeler that we are mostly Kimball in our approach, but I think that only applies to the Data Marts. When I think of a Data Warehouse, I am thinking of a central place to store historical information. Then we have a concept of Data Marts, which are separate databases of information in a slightly denormalized fashion (this is where you hear about Facts, Dimensions, and Aggregates). For Data Warehousing ......