Database Reference
In-Depth Information
2. terms corresponding to the non-standard aggregation operators, including:
a. fuzzy (linguistic) quantifiers (e.g., “most”),
b. importance coefficients (e.g., “important to a degree 0.8” or “very important” etc.)
The query languages of the classic DBMSs do not provide any means for representing such linguistic
terms, and some extensions of a classic query language, the SQL are needed. These extensions concern
both the syntax of the language and a proper interpretation of particular new linguistic constructs ac-
companied with some scheme for their representation, elicitation and manipulation. Here we discuss
these issues on the example of FQUERY for Access.
The Syntax
We will focus our attention on the well-known SELECT...FROM...WHERE command of the SQL, and
will deal only with its WHERE clause. Starting with its simplified version, e.g., excluding subqueries,
we propose the following additions to the usual syntax of this clause providing for the direct use of
linguistic terms:
<WHERE-clause>::= WHERE <condition>
<condition>::= <linguistic quantifier> <sequence of subconditions> ;
<sequence of subconditions>::= <subcondition> | <subcondition> OR
<sequence of subconditions>
<subcondition>::= <importance coefficient> <linguistic quantifier>
<sequence of atomic conditions>
<sequence of atomic conditions>::= <atomic condition> |
<atomic condition> AND
<sequence of atomic conditions>
<atomic condition>::= <attribute> = <modifier> <fuzzy value> |
<attribute> <fuzzy relation> <attribute> |
<attribute> <fuzzy relation> <number> |
<single-valued-attribute> IN <fuzzy-set constant> |
<multi-valued-attribute> <compatibility operator> <fuzzy-set constant>
|
<attribute>::= <numeric field>
<linguistic quantifier>::= <OWA-tag> <quantifier name>
<OWA-tag>::= OWA |
<modifier>::= VERY | MORE OR LESS | RATHER | NOT |
<compatibility operator>::= possible matching | necessary matching |
Jackard compatibility
Now, let us discuss particular categories of linguistic terms listed above. In what follows, we mainly
use examples referring to a hypothetical database of a real estate agency. Particular houses are charac-
terized by: price, land area, location (region), number of bedrooms and bathrooms and other life quality
indicators as, e.g., an overall assessment of the environment, transportation infrastructure or shopping
facilities.
