Due to the enormous increase of semi-structured data over the last decade generated by services like Twitter, Google, Facebook or Instagram efficient ways of processing are necessary to fulfill the needs. Traditionally structured data are stored in relational databases, for storing semi-structured and unstructured data NoSQL databases or Hadoop solutions are the most performant way to do it. Bridging those two worlds is often necessary. Therefore semi-structured data like XML in conjunction with a relational database is still a hot topic in computer science. This thesis analyses XML storage and retrieval possibilities within the open source database MySQL. MySQL is chosen as RDBMS because it is free, popular and offers free and well documented interfaces for extension. A data set with one million XML documents (~149MB) is set up for testing purpose. First, the built in functionality of storing XML of MySQL is investigated. The limitations of XML processing through MySQL are listed and discussed. In order to support full XPath 1.0 functionality under MySQL and to improve XML processing performance a generic table approach is introduced. For two XPath expressions a sample mapping to SQL were presented and explained. Therefore all possible XPath axis are shown and discussed first, because for each of those another mapping would be needed. Data retrieve time is compared with the standard MySQL approach and BaseX, a native XML database. As a result it is shown that the generic table approach is the most performant solution for the specifically discussed XPath expression.