Related Studies On Incomplete Information Database

Incomplete information database is an important research direction in database field.Most of existing commercial databases are based on traditional relational data model,thus lack of formalized definition and a unified logic system when dealing with queries relating to NULL values.These defects are especially obvious when dealing with queries involved with duplicate tuples and inapplicable information,and the inadequate processing would lead to counter-intuitive query results,what's more,may bring some hidden dangers to the database security.Aiming at the problems above,this thesis has carried on thorough research based on traditional relational data model and some classical three-valued logic systems.The major work is listed as follows:Firstly,proposing a new logic connective at the basic of K₃ logic system,and expanding the definition of multi-set relational data model in the environment of NULL values.Then we expanded the standard multi-set relational algebra MRA toMRA^N,which can deal with NULL values in duplicate tuples.To improve the practicability of research results,we discussed the properties ofMRA^N,and applied the algebra to query instancesSecondly,introducing the methods dealing with inapplicable NULL values in relational database query and fuzzy query respectively,then we improved the selection and difference operation algorithms in relational database.As to fuzzy query,we proposed a new method by combine the PTV?Possible Truth Value?and MV?Marked NULL Values?to classify inapplicable NULL values to corresponding values in relevant attribute domains.Thirdly,introducing some classical three-valued logic systems and pointing the defects existing in these systems.Then we proposed a new four-valued logic system Z₄ based on the systems above.InZ₄,the third truth value is divided into two new truth value: applicable but unknown and inapplicable,thus the problem in three-valued logic systems is harmonic.After that,we expanded the connectives in Z₄ and proved the truth value is completeness.Besides,we discussed about the connectives' properties and the relation between Z₄ and classical three-valued logic systems.Finally,we proposed the algorithms of comparison and selection operation based on Z₄,and applied the operations to query instances,which illustrating Z₄ 's practicability.