One of the problems associated with XML databases is the poor performance of XPath queries. Although this has attracted much attention by the research community, solutions are either partial (not providing full XPath functionality) or unable to manage database updates. In this work, we exploit features of Oracle 10g in order to rapidly build indexes that improve the processing times of XML databases. Significantly, we can also support XML database updates as the rebuild time for entire indexes is reasonably fast and thus, provides for flexible update strategies. This paper discusses the process for building the index repository and describes a series of experiments that demonstrate our improved query response times.