An Efficient Data Extraction and Storage Utility For XML Documents

In this paper, a mechanism to provide selective extraction of data objects from XML documents, the storage of these documents in an object-relational database, and retrieval and reconstruction of XML documents from extracted data objects is discussed. The motivation is provided by a need for a Workflow Process Repository in a Workflow Management System (WFMS) [6], namely METEOR WFMS, to store meta-data about workflow designs, organizations, informational resources and computational resources. Thus meta-data is composed of different XML documents representing different components of a workflow process. The repository, involving the Data Extraction and Storage Utility (i.e., Extractor), has the following main capabilities: Filtering of XML objects that need to extracted, Generating relational schemas for on-the-fly storage of XML documents, Loading data from XML documents into relational tables, Re-creating original XML documents as needed, Querying, browsing, and versioning. Our XML storage scheme is so practical and flexible. Practicality comes from the broader acceptance and use of Object-Relational Database Management Systems (ORDBMSs); flexibility is provided by selective extraction mechanism (i.e., filtering) employed by the Extractor, which is not available in similar approaches [3] using an ORDBMS. The comparison of our approach with XML databases (e.g., Lore [5]) in terms of efficient storage and querying XML documents [1,2] requires more research and performance testing. Although they support a native data model for XML documents, many XML databases do not provide high performance. Recently, some database vendors have started to support XML. Axielle by Ardent Software, eXcelon by ODI, and XML Repository by Poet are some examples. In general, these products provide an import and export programming interface, which is compliant with Document Object Model (DOM) from the W3C.