In today’s episode of Data Vault Mysteries we demystify Zero Keys and Ghost Records!
In the construction of your fact table, if the measure falls in any one of those categories of data listed above, then the foreign key loaded to the fact table will be one of those pre-loaded keys. What that does is when a star-join optimized query is run against that star schema, it will return the dimensions and fact content selected and one of those default values as determined by your reporting rules. Unlike data vault, the default keys here are strictly used for reporting! Still, the same EQUI-join query is used, and database platforms with OLAP functionality recognize this type of query and optimally returns the data from this dimension – fact table configuration. See: bit.ly2O0pUXh
This concept is not useful to data vault because it is not intended to be an information layer business intelligence (BI) tools are accustomed to using. These BI tools will use the information marts built over a data vault and not directly query it!
The difference in mindset between dimensional modeling and Data Vault is never so clear; dimensional modeling focuses on the business reporting outcomes, whereas data vault molds to the platform environment, technology, and business processes to build a data warehouse. The focus of either is so different, including its implementation. It has never been so plain to see in its definitions of default keys; what I mean is: the Data Vault is designed to build a data warehouse, and Kimball modeling is designed around an information mart. Let that sink in – data warehouse versus an information mart. An information mart conforms data to suit the business need now. A Data Vault integrates data sources through agility, automation, and audit-ability as a base to deliver conformed information marts. The two solve two different purposes in a data warehouse. Data Vault does not replace the dimensional mart, but rather it complements it. As the methodology is used as the base, it is the audit history. It is non-destructive to change, and thus it implies that the information mart layer has become disposable.
Disclaimer: The opinions expressed in this article are entirely my own and will not necessarily reflect my employer’s.
Author: Patrick Cuba
Patrick has nearly 20 years working on data-inspired problems utilizing his experience and he has embraced Data Vault 2.0. He works by understanding the business before innovating the technology needed to ensure that his data-driven delivery is agile and automated. He is a Snowflake Solutions Architect, Data Vault 2.0 certified and regularly contributes to Data Vault Alliance.