Database Reference
In-Depth Information
Figure 8. Experimental Results
cessing mechanism proposed in this chapter. There are two keys of advantage in this mechanism. First
key is the derivation of boolean conditions form fuzzy ones at optimizer level. It avoids superfluous
computation of unnecessary satisfaction degrees for fuzzy conditions. Second key is the computation
of satisfaction degrees into access methods and physical operator. It avoids the rescanning of result
set for the computation of satisfaction degrees. It has been possible because we have proposed fuzzy
query processing at core of the RDBMS engine. With this approach we avoid overhead due to the post-
processing applied at the top for processing SQLf in previous works.
Proposed extension for fuzzy query processing with reported benefits would lead to wider acceptance
and use of fuzzy querying systems for real applications.
The development of PostgreSQLf is still open. There are features of SQLf that has not been imple-
mented yet at core of this RDBMS, it may be matter of future works. Up to present time we have as-
sumed that cost model for the optimizer is not significantly affected and we have shown in practice that,
nevertheless, it would be very interesting to do a formal study of cost model for this kind of queries. Also
it is possible to think about more powerful fuzzy algebra specific physical operators to be implemented
and considered bye the planner optimizer.
Additionally we are designing various benchmark tests to compare our approach with other imple-
mentations of flexible queries.
ACKNOWLEDGMENT
We acknowledge financial help of Venezuela's FONACIT Project G-2005000278 and France's IRISA/
ENSSAT Project Pilgrim. “I will lift up mine eyes unto the hills, from whence cometh my help. My help
cometh from the LORD, which made heaven and earth” (Psalms121:1-2).
