Using a mapping knowledge base in a system for (semi-)automatic loading of XML data into relational databases

Very often the need is felt to store information extracted from the web, from databases or from documents and converted to XML (eXtensible Markup Language) in a database, in most cases, in a relational one, in order to be able to query this database for further operational or strategic decision-making. Unfortunately the XML files do not always have the same structure, and their elements (tags) do not always have the same name. In many companies, a correspondence (mapping) between elements of the XML files and the columns of the tables in the target database is still established manually. In this paper we present a system which loads the data contained in XML files into several tables of a relational database. The system supplies the user with mappings which are based on similarities with mappings already performed by the user in the past. User intervention is therefore required less frequently over time, resulting in full automation of the process of loading data into the target database.