Computer resource usage by such queries can be extremely
high, and hence we use this benchmark to compare the performance of PMap
indexing and REBSI. The set query benchmark has the following key characteristics.
??? The queries for the benchmark are specified in SQL, and the data used is
representative of real applications.
??? These queries are chosen to span the tasks performed by different strategic
data applications (e.g., document search, direct marketing, and decision
support).
??? The benchmark specifies measurements for a wide range of selectivity
values within each query type.
The database has a single table called BENCH that contains 1 million rows of
200 bytes each (224 with overhead). Besides using the given size to measure
performance for large databases, we also use a size of 50,000 rows to mea-
Indexing in Data Warehouses
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission
of Idea Group Inc. is prohibited.
sure the performance for a smaller database size. Each of the 13 attributes has
integer values ranging from 1 to its cardinality, which is represented in the
attribute name. Thus K2 has 2 values: 1 and 2, K4 has 4 values: 1, 2, 3 and 4,
K100k has 100,000 values: 1, 2, ??¦, 100,000. One indexed attribute, KSEQ, is
a clustered primary key, with values: 1, 2, ??¦, 1,000,000.
Pages:
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371