Applications Of XML Constraints In XML Data Storage, Publishing And Transformation

In the past several years, XML has been one of the standards for data exchangeon the Web. It is widely applied in many fields, and much research work is done on it.Many great achievements are gained in research on XML data model, XML datastorage, XML query processing, query optimization, and indexes, and so forth. Weresearch the following topics based on the previous work on XML: transforming keysfrom XML to relation, mapping XML documents to relations in the presence of XMLfunctional dependencies, and XML schema-oriented XML data transformations,which include XML schema-directed relational data to XML data publishing, andXML schema-conforming XML data to XML data transformation. As a result, wepropose effective algorithms and prove their correctness. We also implement themethods in our prototype systems, and conduct preliminary experiments. Thoseexperimental results verify the effectiveness of our techniques. Our research resultsare valuable not only in theory but also in practical applications. The maincontributions of the paper are as follows: 1) In this paper, we consider an important class of XML semantic constraints,XML keys, and try to find the relationship between XML keys and relation keys. Wedesign a heuristic algorithm to transform XML keys into relation keys. We also givethe proof that our algorithm can guarantee the completeness of the set of relation keysafter transformation. Based on this, we develop a system named as KeyGen withVisual C++ for key transformation. When users want to import online XML data intolocal relational databases, the facility KeyGen in this paper can be adopted to predictwhether importing will be completed successfully or not in advance in the constraintlevel. 2) Many methods for mapping XML to relations have been proposed withoutconsidering the semantic constraints of XML data before. But those constraints arevery important to design schemas for storage, optimize queries, and check updateanomalies, etc. We consider another important class of XML semantic constraints,XML functional dependencies XML FDs, and then present a method based on thehybrid inlining to map XML to relations for storage in the presence of XML FDsspecified over DTDs. It can preserve the XML FDs, as well as the content and thestructure of the XML data, in relational databases. Borrowing the concept of XML FD,a better relational schema can be obtained by our method. Therefore, much storage 3Abstractredundancy can be reduced. Furthermore, we prove that those relations mapped fromXML are all in Third Normal Form. 3) When data is exchanged on the Web, the XML schemas of the two entities(their standards) must be considered. If the source data is relational data, it needs to bepublished into XML data conforming to the schema of the target entity. On the otherhand, if the source data is XML data, it also needs to be transformed into the XMLdata which conforms to the schema of the target entity. However, the majority ofresearch on XML publishing and transformation ignores this point. In this paper, weassume that there always exists a predefined XML schema and users require thatXML data must conform to the schema after publishing or transformation.Specifically speaking, the schema is represented by DTD. We discuss the problem ofDTD-directed relational data to XML data publishing, and the problem ofDTD-conforming XML data to XML data transformation. We present two sistertechniques, Attribute Translation Grammar ATG and XML Transformation GrammarXTG, to solve the above two problems, respectively. Then we use Java to develop twosystems, PRATA and TREX, as two facilities for pratical applications. 4) Because the techniques for relational data to XML data publishing are allrelated to backend relational databases, the optimization approaches presented inXML publishing systems rely on relational databases. Considering the parti...