An Approach for Generating an XML Data Warehouse Schema using Model Transformation Language

Traditionally, the multidimensional schema of the data warehouse is derived from data sources that are mainly the company’s internal data, well-known and structured, by identifying facts, dimensions and numeric measurements through a manual analysis of the operational schemas. With the proliferation of new platforms of communication in today’s information societies, there has been growing numbers of web-based applications such as online social networks that generate huge amounts of XML data on the web. Therefore, it is increasingly important to develop an appropriate warehousing approach for such ever-growing XML data sources. However, XML documents have a complex hierarchical structure. Moreover, designing and building DWs is tedious, timeconsuming, error-prone and expensive process. In this paper, we describe an approach for automatically generating and building the star schema for data warehouse from XML schema. This approach is extensively based on standards (UML, XML, QVT, and XSLT). First, we model the structure of XML (XML Schema) using the Unified Modeling Language, which is the standard language for object oriented analysis and design. Then, we provide an algorithm that automatically selects the multidimensional concepts. After that, a representation in an XML schema language as the XML data warehouse schema description is derived automatically from UML star schema. Furthermore, we choose Query/Views/ Transformation (QVT), which is also OMG standard transformation language for defining and formalizing transformations between models. Finally, a prototype tool is implemented for testing and evaluating our approach and its transformations. Subject Categories and Descriptors H.2.7 [Database Administration]: Data warehouse and repository; D.3.3 [Language Constructs and Features] Data types and structures General Terms : Data warehouse Design, XML, XML Schema