Metadata Management and Data Governance - Data Governance: Creating Value from Information Assets - page 60

feature called Hybrid Columnar Compression. With this feature,
one can compress the data by a factor of 10. The data is still avail-
able; however, access times are often over two times longer. The
main advantage of this approach is that the data is available and
stored in the database so that, when the database is upgraded, the
compressed data is upgraded as well.
DATA VERSIONING
Data governance provides business rules for the versioning of data. There
are different data versioning techniques depending upon the data format in
which the data is stored—star dimensional model versus third normal form.
Data Versioning Techniques
Data versioning is different depending upon the data format. The two
main data formats are star dimensional model and third normal form.
Star Dimensional Model Data Versioning
Star Dimensional model is very different from the 3 rd normal form. It is
a much flatter or simpler data model, and consists of fact and dimension
tables. The fact tables represent the WHAT. A fact table consists of foreign
keys from the dimension tables and measurements or metrics of a business
process. It is the center of a star schema surrounded by dimension tables.
For example, what are the total sales by product, by state, and by year? Total
Sales is the fact (metric). The dimension tables are the HOW. A dimension
is a category used to view/access the metrics. In our example, product,
state, and year are the dimensions, or how we want to view the data.
In the Star Model in Figure 3.9 , we have one large table called the Sales
fact table and a set of smaller tables called Dimension tables: Product,
Date, Customer, and Location.
Star Dimensional Model data versioning is supported through slowly
changing dimensions, which are dimension data values that change over
time. For example, the customer dimension stores information, such as
a customer name. Often when a female marries, she will change her last
name to that of her husband's last name. Data governance will give us the