The Research Of 3NF And BCNF Based On XML

XML(eXtensible Markup Language)has become one of the Primary standards for data exchange and representation on the Web. It is widely used in various fields. Likes the relational database to be same, the XML documents also possibly include some redundant information that causes the update anomalies. These questions all are caused by certain functional dependencies among paths in the document. It is well known, relational database design and concept of normal form are crucial for the relational database technology. Similarly, normal form research based on the XML aim at processing directly to model in the web world from the database design angle, thus obtains the database model normalized, like this, not only the semantics and the structure information in the XML documents are preserved integrally, the database design requirement has satisfied. The purpose of this paper is propose the concepts of partial functional dependence and transitive functional dependence as well as local candidate key for XML with DTDs as their schemas, and 3rd normal form(X3NF) and the BC normal form(XBCNF) based on the XML. Finally using the concept of lossless join decomposition for DTD and normalization rules for XML schema, find a way of converting an arbitrary DTD into document satisfies X3NF and the XBCNF. This will lay the theoretical foundation for holding functional dependence, the XML key words, the XML integrity constraints, the inference rule, the XML multiple-valued functional dependence as well as the further research of normalization for XML in the future. This paper main content including following two aspects:The first question studied in this paper is the functional dependence in XML documents. The functional dependence is the key question for XML documents design, also has the significant theoretical significance to the XML other correlation research, like inquiry optimization and normalization and so on. However, functional dependence in the XML documents has many remarkable differences from functional dependence in the relational database: In the XML documents, whether a functional dependency satisfied, first needs to consider this functional dependence scope where satisfied; In addition, possibly also must compare the node which the element corresponds in the XML documents is whether same. Because above difference and complexity of XML documents structure, XML functional dependence and its corresponding inference rule must be more complex than corresponding question in the relational database. What is more important is that the function dependence that defined here can represent the element (node) in XML documents, the attributes of an element and the functional dependences between the values of elements. And expended the concept of the functional dependence in the relational database. This also considers the differences between characteristic of XML documents itself and relational database. Therefore, one of the merits of the function dependence which defined in this paper is that its applicable scope is much broader than the functional dependence relations that only considered the attribute of the element and the value of element. After has defined the functional dependence the paper proceeds to research the inference question of functional dependence in XML documents. And it has produced a set of self-contained (complete) inferences rule sets of function dependence in XML documents. Moreover, the key also is one kind of data restraint. Therefore, this paper has also studied the relation between functional dependence in XML documents and the keys further. According to the key theory in relational database and the characteristics of XML documents itself, the paper has proposed the concept of local candidate key. In other words, it means that there are candidate keys in local scope, which is local candidate key. The paper has produced the algorithm for solving the local candidate key in XML documents and has produced the analysis and the proof for the algorithm to establish the foundation for the research on posterior normal form.The second question studied in this paper is the standardized question for the XML documents that is relation to the functional dependence. It has defined the two kinds of functional dependence which can cause the data redundancy in XML documents and the unusual operation: Partial (local) functions dependence and transfer function dependence that can be regarded as separately the concept exudes for the partial (local) functions dependence and transfer function dependence in the relational database but cannot equate completely. They both have the big relations with the structure of the XML documents itself. Then based on this, the paper has defined the X2NF and X3NF in XML documents. They generalize X2NF and X3NF in relational database. And eliminated the data redundancy and the unusual operation which are caused by the leaf nodes cause to partial functional dependence and transitive functional dependence for local candidate key. Then defined the second kind of XBCNF for XML and emphasized its tenable premise is complete tree. It is exudes for BCNF in the relational database. The relation of the X3NF and XBCNF is the same one of 3NF and BCNF in relational database. That is if XML documents conform to XBCNF, then it also conforms to X3NF. Moreover, according to the concept of the DTD lossless decomposition and using the concept of local candidate key proposed the two algorithms that lossless decomposed the DTD so as to conform to the X2NF and the X3NF. The paper proposed the normalization rules of XML and using the rule has given the normalization algorithm that can lossless decompose the any XML DTD model so as to conform to XBCNF. And introduced the concepts of redundancy cover and normalization cover into normalization algorithm of X3NF and XBCNF. In the normalization algorithm of X2NF and X3NF. By using the solving algorithm of local candidate key to find the local candidate in XML documents. That because the condition of premise for definition of X2NF and X3NF is XML complete tree. Therefore, only carries on the judgment to the non- attribute leaf node and eliminated the partial functions dependence and the transfer function dependence of the local candidate key caused by them and has achieved the normalization target.For the further research, it mainly concentrates in three aspects: 1.The decomposition algorithm for functional dependences preserving. The three algorithms which produced in this paper for the decomposition of the X2NF, X3NF and XBCNF didn't consider holding the functional dependence in the process of the decomposition. Therefore, we should look for the decomposition algorithm and decomposition rules that hold the functional dependence in the future research to meet the needs of the occasion that holding the decomposition of the functional dependence. 2. The decomposition order of the algorithm. The result of the decomposition algorithm for the BCNF is not the only resolves possibly. The decomposition result concerns the order of the selected unusual data dependence. Therefore, by what kind of order selecting the abnormity dependence to make decomposition so that can conform to the actual relations of the elements in DTD is the second worth researching question, that means in which order selecting the abnormity dependence to make decomposition is the most optimization one. When XML tree is incomplete the improvements of X3NF and the XBCNF is the third aspect. This paper mainly discussed X3NF and XBCNF definition and normalization algorithm when XML tree is complete. because the basic theory for XML incomplete tree is not very mature, these questions wait for in doing more thorough research.