Fuzzy Querying Capability at Core of a RDBMS - Advanced Database Query Systems - page 173

Box 3.

#define RelationFuzzyPredId 2859
CATALOG(pg_fuzzypred,2859) BKI_BOOTSTRAP BKI_WITHOUT_OIDS{
NameData predname; // predicate name

int2 predbegd; // domain range begin

int2 predendd; // domain range end

int2 predminfp; // support's left bound

int2 predcore1; // core's left bound (included)

int2 predcore2; // core's right bound (included)

int2 predmaxfp; // support's right bound

int2 predtypefp; // shape 1(trapz.) 2(increases) 3(decreases)

NameData preddisd; // discrete domain

text predcompfplist[1]; // fuzzy predicates compare list name

NameData predexprfp; // expression

} FormData_pg_fuzzypred;

Box 4.

#define RelationFuzzyModId 2879
CATALOG(pg_fuzzymod,2879) BKI_BOOTSTRAP BKI_WITHOUT_OIDS {
NameData modname; // modifier name

int2 modtype; // modifier type

int2 modpower; // modifier power

NameData modnorms; // t-norms and t-conorms name

NameData modfirstarg; // t-norms and t-conorms left arg

NameData modsecarg; // t-norms and t-conorms rigth arg

} FormData_pg_fuzzymod;

Fuzzy Terms Nodes

If the parser module finds a fuzzy term then a new fuzzy node is created, i.e.: A_FuzzyPred (a fuzzy
predicate), A_FuzzyComp (a fuzzy comparator) or A_FuzzyQuan (a fuzzy quantifier); we record in these
memory structures the data retrieved from the fuzzy catalog tables.
Box 8 shows the node structure for holding in memory fuzzy predicate information.
In a similar way, the node structure for fuzzy query information in parsing structure is (see Box 9).
For the case of a fuzzy comparator, the representation as a node is in Box 10.

Fuzzy Parse Tree

The parser module has the mission of recognizing the statements in SQLf. In case of a well formed query,
the parser generates a in memory abstract structure named the parse tree. When the query contains some