XBRL Framework Wiki
Brought to you by:
mapsilva
Menu
▾
▴
LMDQL
Authors:
Professor PhD Paulo Caetano da Silva, PhD (Lattes CV)
Professor PhD Valéria Cesário Times, PostDoc (Lattes CV)
The use of tools for data analytical processing (OLAP) to perform the strategic analysis of an organization allows users responsible for making the decisions to identify trends and patterns in order to better conduct the business of the company where they work. However, the academic and commercial development of analytical processing systems for XML data does not have all the features of the existing OLAP tools for traditional data and interrelated XML documents. Therefore, the need for developing
OLAP systems to assist in the strategic analysis of data from one organization, represented in XML format and interconnected by a set of references, is the main motivation for developing this work.
Currently, research is being developed in the academic context in order to perform analytical processing of data represented as XML. However, because these technologies were originally designed for different purposes, this is not a trivial task. To assist in the development of OLAP systems, this work proposes a computational system consisting of an analytic query language, LMDQL, to allow queries to multidimensional data based on interconnected XML documents be performed. As the W3C standard language to query XML documents (e.g. XPath and XQuery) does not support links, a navigation language for XML documents with links, namely XLPath, was specified and incorporated into the LMDQL processor. In addition, to model data cubes of interconnected XML documents and to deal with the syntactic, semantic and structural heterogeneities found in data represented as XML, a multidimensional metamodel, called XLDM, is given. From these specifications, we developed a prototype system that enables the analysis of interrelated XML documents. This prototype includes LMDQL, XLDM, XLPath and a driver, sql2xquery, which performs the mapping of SQL queries to XQuery, allowing the incorporation of this prototype on a OLAP server for relational data.
To evaluate the proposals presented in this thesis, three case studies from distinct areas (i.e. medical, financial and sales) were performed to show the expressiveness and use of the proposed system. We also discuss performance tests results aimed at evaluating the execution time of the analytical processing over interconnected XML documents. In these tests, we evaluated the efficiency of the driver for three DBMSs for XML data and investigated the impact on the execution time of analytical processing over XML documents that use links to express additional information to XML data. The costs of query processing in LMDQL was compared with the costs of processing the same queries written in MDX (an industry standard), showing that the benefits achieved with the use of LMDQL do not impact the execution time of queries.
