g., primary key violations).
??? On-Demand.ETL:.An ETL process of this kind is executed sporadically,
and it is manually initiated by some user demand. The process is responsible
for retrieving external data and loading them in the DW after the appropriate
transformations. For instance, consider the case that some users request data
to be brought in from the Web. The administrator/programmer is assigned the
task of constructing an ETL process that extracts the dates from the specified
sites, transforms them, and ultimately stores them in some (possibly novel)
part of the warehouse. Any time the user needs this data, this on-demand ETL
process brings in the relevant information. As research challenges in this area,
we mention the following issues:
?‚? The need for appropriate operators, since this process is mostly focused
towards Web data.
?‚? The computation of minimum effort/time/resources for the construction
of the process.
?‚? The provision of a framework easily adaptable to the changes of the external
data.
?‚? The finding of efficient algorithms, due to the fact that this process is
initiated by the user.
Conclusion
In this chapter, we have delved into a crucial part of the data warehouse architecture:
the backstage area. We have presented the state of the art concerning the existing
ETL technology. In practice, a designer/administrator uses a commercial ETL tool
or an in-house developed software artifact to create ad-hoc ETL workflows.
Pages:
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266