BiLCLP: Storage Efficient Prefix Labeling Scheme

Prefix scheme is popular to label XML tree. In order to overcome the defect of rapidly the code size increasing with the depth increment, we propose a novel prefix labeling scheme based on layered structure. In our scheme, the XML document tree firstly is partitioned into some subtrees according to the segment granularity, whose roots are uniquely labeled by prefix codes as an outer code identifying all the nodes in the subtrees. And then the nodes inside the subtrees are also labeled by prefix codes as node´s inner codes. Finally, it is necessary to establish a linear table to link the root-of-subtree outer code and the code as a leaf node in the upper level subtree, which can help to restore the layered code to the complete one in the whole tree. Extensive theoretic analysis and experimental results show that the proposed prefix labeling scheme has relatively smaller mean code size, which is not increased with the increment of XML document, and shorter query response time for the smaller code size.