Publication at Faculty of Mathematics and Physics |
2008
Abstract
We have significantly enhanced the Xord framework for streaming processing of large XML data by introducing buffers for input fragments. The analyzer examines given top-down XSLT and XSD, and generates fragments which identify parts of XSD that need to be buffered during XSLT transformation.
The transformer uses auxiliary memory buffers to store temporary data according to the analyzed fragments. We describe the implementation of the engine within the Xord framework and provide evaluation tests which show that the new engine is much more memory-efficient comparing to the common XSLT processors.