PDT-based document fragmentation of XML streaming data

Unlike in conventional databases, queries on XML stream data are bounded by not only the memory capacity but also the real time processing. Based on the Hole-Filler model, a path frequency tree (PFT) is defined according to the statistic information on queries about XML to set out a sibling-based document fragmentation policy including corresponding algorithm. Then, an alternative membership-based document fragmentation policy and corresponding algorithm are proposed. Both algorithms can effectively enhance the utilization and cohesion of XML fragments. Testing results showed that the PFT-based document fragmentation algorithms perform well on query cost and other properties.