A relevance comparison between interval and prefix labelling schemes

Abstract

Improving XML database management system has attracted researchers to consider whether the indexing system is equivalent to a relational database management system. The indexing system is based on labelling the nodes of the XML tree. Different types of labelling scheme have been proposed to label the document quickly and without consuming too much storage space. However, most the studies focused on evaluating the performance of new labelling schemes. The appropriateness of various existing schemes to the particular structure an XML document has not been addressed sufficiently. To investigate this aspect two common XML labelling schemes were employed: Prefix (Dewey Encoding) and Interval (Containment) to label three different examples of XML documents with very different structures. The time and storage space requirements were investigated to compare the relevance of each scheme to the structures of the documents. A number of experiments were conducted and it was found that Dewey Encoding and Containment techniques are relatively fast when labelling shallow tree structures. Dewey required little storage space to save labels of wide tree structures, however, Containment used less storage space when storing the labels of short trees.