Rule Evaluation Algorithm for Semantic Query Optimisation

Statistical query-based rule derivation system by backward elimination algorithm

“…Although somehow complex rules can be expressed, using them in the query optimisation becomes too difficult; also the rule maintenance becomes problematical. Therefore the rule representation is limited to using simple rules [8,13,7,3,24,20,10,6,1,25,12,14,18,9,23,2,22]. Assuming a table is R (x, y, z) where x, y and z, are attributes of R. A rule can be formed as x⊕a → y⊗b where a and b are constant values; ⊕ and ⊗ are one of comparison operators ¡, ¡=, ¿, ¿=, =, !=.…”

“…These derived rules may be violated during SQO when database updates are made. Therefore, the logically equivalent queries are clearly semantically equivalent but the semantically equivalent queries may not be logically equivalent [8,13,7,3,24,20,10,6,1,25,12,14,18,9,23,2,22]. 3.…”

“…When updates are made on the database, these rules may become invalid. Therefore derived rules need to be maintained to be consistent with the current database state [25,12,14,18].…”

“…Another issue of the rule maintenance manager in Figure 1 is to keep the rule set accurate for new coming rules in case of having the same antecedent attribute and the consequent attribute with different comparison operator and values in order to manipulate their boundaries correctly. For this purpose, our previous papers can be taken into account [21,19,17].…”