Proceedings

Abstract:

As XML formatted content and data becomes pervasive on intranets and the Internet the requirement to minimize individual process times becomes great. XPath has been evolving into a rich expression language to query and extract data in a precise way. While it has been designed to be used by a host language such as XSLT and XQuery, an XPath processor can be used quite usefully standalone or as part of an application framework.

This paper discusses the design of such an XPath-powered framework which supports registration of processes to individual XPaths. These XPaths are compiled with a plug-in analysis engine and satisfied through stream-based document parsing for optimal scalability and performance. A publish-subscribe model allows data consumers to register XPath expressions and function callback handlers to the framework. Through real-world examples, we will discuss how this streaming XPath XML data processing framework can efficiently extract XML data for content management applications populating electronic warehouses or serve as high-performance data router for Web services and application integration applications. We will also present how this framework can be used as a high-performance data acquiring processor for XSLT and XQuery engines.