XML query algebra operators, and strategies for their implementation

We propose an improvement to the logical view of the selection operator in Niagara Algebra [L. Galanis et al. (2002), S. D. Viglas et al. (2002)]. We demonstrate the improvement due to our implementation by experimental results. We observe that the degree of improvement depends on the nature of the XML data. The new operator is particularly better at working with data that has a large number of elements that need to be unnested in order to run a select In addition, it shows a significant improvement in situations where a large number of elements must be evaluated to check if the selection criteria are met.