The Research Of Compression Storage Organization And Query Processing Methods Of XML Data Modeled As A Tree

With the rapid development of Web Service these years, XML (extensible markup language) has played an active role in the field of data storage as well as data exchange. More and more semi-structure data has been represented by XML and XML has become an important criteria of data representation and data exchange on Internet. Because of the rapid growth of XML data and the much redundancy in XML data, efficient storage as well as query techniques on large xml dataset has become a hot problem of XML research these days.The principal problem is to build effective disk storage structure when processing large XML dataset. So at first in this paper, a disk-based xml data storage structure LDF&B based on F&B index is proposed. On one side, LDF&B wipes off the redundancy existing in traditional storage structure which stores the same value information repeatedly; on the other hand, the index on the tape and the inverted index introduced in this storage structure enable nodes with the same property to be clustered when stored on disk, and enable nodes satisfied with the query to be searched effectively. That is to say, LDF&B can support XML query processing effectively.Then we propose a series of query processing methods based on LDF&B. Firstly, when processing queries with value, the algorithm proposed could filter a large amount of intermediate results which could not satisfy the query pattern, at the same time, it reduces the operation of further matching of these intermediate results. In this way, it enables query processing with value efficiently, which could not be processed by Disk-based F&B[1]. Compared with another method which first matches skeleton information and then value information, its query processing efficiency has been promoted remarkably. Secondly, when processing branching queries, compared with the Join method, the amount of nodes that need to be traversed in the XML data tree here could be reduced to a great extent, what is more, it would not need to store intermediate results to obtain the final result, which makes it engrosses little memory. Branching query processing efficiency has also been promoted remarkably. Thirdly, when processing simple path queries, this paper proposes methods that could locate the appropriate nodes...