Fuzzy Querying Capability at Core of a RDBMS - Advanced Database Query Systems - page 179

Figure 6. Derived Fuzzy Query Tree
EXPERIMENTAL RESULTS
We have designed and tested a set of queries on PostgreSQLf in order to verify functionality and perfor-
mance results. In general, the results obtained are good. The functionality tests and validation of results
are done with a 100% level of effectiveness. The performance tests are very good because the time is
not incremented much and in some instances the mean time is very similar than in the classical case.
We have designed and tested a set of queries on PostgreSQLf in order to verify functionality and
performance results over the TPC Benchmark™H (TPC-H) available at http://www.tpc.org/tpch/.
For such tests several fuzzy terms were created. Some of them are shown in Box 11.
We have designed 28 fuzzy queries with different querying structures and fuzzy condition. One of
such queries is in Box 12.
We compare the performance of processing fuzzy queries in PostgreSQLf regarding regular querying
processing in PostgreSQL. For each designed fuzzy query we run a boolean version obtained as the
corresponding derived query according the Derivation Principle. This comparisos allows showing the
added extra cost for fuzzy query processing with mechanism that has been proposed and implemented
in this work.
We generate two volumes of data (low 1 gigabyte and high 5 gigabytes) and we obtain execution
time of fuzzy queries and their corresponding classical queries, of course, in the first case we addition-