Development of an XML-document compaction method to improve data-processing performance

XML (Extensible Markup Language) is the global standardized form for representing digital data. It is flexible and extensible, but on the other hand uses vast amounts of main memory for data processing, thus representing a burden. In order to resolve this problem when processing data in large XML documents, in this paper the authors propose the “CSV compaction for XML” method to improve data processing performance by using a format conversion for an XML document with a record structure. In this method, redundant elements are packed in CSV (Comma Separated Values) format from the perspective of accessing the XML document prior to data processing. The results of evaluating this method confirmed that memory usage and deployment time were improved in proportion to the percentage of elements that could be packed and eliminated using the CSV format. This method is a reversible conversion for an XML document. Since the authors created software to execute the method as XSL (Extensible Stylesheet Language) Transformation, it can be executed using various programming languages. In this method, specifications that designate elements to be packed in CSV format in a record are needed, and creating them is troublesome. Thus, the authors created a support tool to create the specifications simply using a GUI (Graphical User Interface). As a result of these developments, complex documents with a hierarchical structure and documents with a large number of elements can be readily handled in this method. © 2007 Wiley Periodicals, Inc. Syst Comp Jpn, 38(5): 1– 13, 2007; Published online in Wiley InterScience (). DOI 10.1002/scj.20736