.2
Advanced Ad Hoc Star Query Processing 4
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission
of Idea Group Inc. is prohibited.
Each h-surrogate can trigger the creation of a foreign key in the fact table. The
concatenation of all these foreign keys produces the fact table??™s primary key. This
is depicted in Figure 2(a). Note that for the fact table we have two alternative composite
keys: (a) (Customer_id, Product_id, Store_id, Day) that links to the corresponding
lowest hierarchical attribute of each dimension and (b) (Cust_hsk, Prod_
hsk, Loc_hsk , Date_ hsk) that links to the corresponding h-surrogate attribute. Note
that the former is not necessary to achieve hierarchical clustering of the data, and
thus could be omitted, in order to reduce storage overhead.
The h-surrogates play a central role in processing ad hoc star queries, first because
they enable the clustering of the fact table according to the dimension hierarchies,
and second because they can be exploited to optimize the query evaluation plans.
Experiments in Karayannidis et al. (2002) have shown speed-ups up to a factor of
20, over the state of the art bitmap-based star join processing (see section on other
methods for star query processing). Even more interestingly, this factor doubled
when query optimization techniques that are discussed later in this chapter were
exploited.
Pages:
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284