Another
technique is log ???sniffing,??? that is, the scanning of the log file in order to ???reconstruct???
the changes performed since the last scan. In rare cases, change
detection can be facilitated by the use of triggers. However, this solution is
technically impossible for many of the sources that are legacy systems (such
a technique adds an enormous load to the source systems) or plain flat files.
20 Simitsis, Vassiliadis, Skiadopoulos, & Sellis
Copyright ?© 2007, Idea Group Inc. Copying or distributing in print or electronic forms without written permission of
Idea Group Inc. is prohibited.
In numerous other cases, where relational systems are used at the source side,
the usage of triggers is also prohibitive both due to the performance degradation
that their usage incurs and the need to intervene in the structure of the
database. Moreover, another crucial issue concerns the transportation of data
after the extraction, where tasks like FTP, encryption-decryption, compression-
decompression, and so forth, can possibly take place.
??? Transformation.and.cleaning: It is possible to determine typical tasks that
take place during the transformation and cleaning phase of an ETL process.
Rahm and Hai Do (2000) further detail this phase in the following tasks: (a)
data analysis; (b) definition of transformation workflow and mapping rules; (c)
verification; (d) transformation; and (e) backflow of cleaned data.
Pages:
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247