For each area, the population is stored as feature
attribute. Therefore, the dimension has three hierarchical attributes (Store_id, Area,
Region) and one feature attribute (Population) that is assigned to the Area level.
Note that the key attributes of the dimension tables are Customer_id, Item_id,
Store_id, and Day and the corresponding foreign keys (Customer_id, Product_id,
Store_id, Day) define the fact table??™s primary key.
In order to create a fact table that is clustered according to the dimension hierarchies
we first need to apply a hierarchical encoding (HE) on each dimension table.
To achieve this we use the hierarchical surrogate key (or h-surrogate) attribute, a
special attribute which is defined for each dimension table. The value of this attribute
is computed based on the value of the hierarchical attributes of the dimension.
The h-surrogate encodes not only the values of the hierarchical attributes but also
the hierarchical relationships defined among the levels of the dimension. Although
there are several equivalent ways to define such an encoding, it is sufficient to present
only one such technique for the reader to understand how h-surrogates are used.
All the query processing and optimization techniques presented in this chapter work
regardless of the particular encoding technique used.
Customer_ d (FK)
Product_ d (FK)
Store_id (FK)
Day (FK)
______________
sales
SALES_FACT Customer_ d
Profession
Name
Address
CUSTOMER
Day
Month
year
DATE
Item_ d
Class
Category
Brand
PRODUCT
Store_id
Area
Region
Populat on
LOCATION
(a) (b)
Year
Month
Day
DATE
Profession
Customer
CUSTOMER
Region
Area
Store
LOCATION
Category
Class
Product
PRODUCT
Figure 1.
Pages:
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281