querying and maintaining ordered xml data using relational databases anil rawat wei pan chang chen...

11
Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Upload: dale-harrington

Post on 21-Jan-2016

213 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Querying and Maintaining Ordered XML Data using

Relational DatabasesAnil RawatWei Pan

Chang chenKelly Yu

Page 2: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Motivation

• XML is being used increasingly often for the exchange of data form heterogeneous data sources. But at present, most business is stored in relational databases, due to the decades of research have been put into the field.

• How to harness the versatility of the XML data model with minimal modifications to the mechanisms of relational databases?

This paper proposes an efficient structural join algorithm that can be implemented on top of existing relational database.

Page 3: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Aim of the work

We mainly introduce a new labeling scheme (XRU) and a new structural join algorithm (SS-Join) that have been presented in the Querying and Maintaining Ordered XML Data using Relational Databases(William M. Shui, Franky Lam, Damien K. Fisher& Raymond K.Wong, 2005).

The method is not based on the B-Tree or Stack, and do not use the indices ,just using a edge table that is stored in the relational database.

Page 4: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

• In XRU random insertion of nodes using local encoding scheme is generally faster than both Dewey and global encoding schemes

• worst case of XRU using local encoding scheme is the same as that of global encoding scheme

Result of the new approaches

Page 5: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Result of the new approaches

• XRU vs optimistic – compare the time and the number of

relabelings – XRU significantly better

Page 6: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Result of the new approaches

• SS-Join– Better when selectivity between low and

medium – Little bit worse than XR-Tree when the

selectivity is high– The SS-join have to spend add extra

overhead on join operation

Page 7: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

• successfully built two new approaches which are XRU Relabeling scheme and SS-join algorithm

Result of the new approaches

Page 8: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Experiments

• XRU relabeling scheme– Global edge table– local edge table (best order maintenance

performance)

Page 9: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Performance results

• Performance of relabeleing scheme– Global encoding scheme– Local encoding scheme– XRU vs. Optimistic

Page 10: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Performance of querying XML

• Performance of SS-D algorithm w.r.t. other structural join algorithms

• Performance of SS-A algorithm w.r.t. other structural join algorithms

Page 11: Querying and Maintaining Ordered XML Data using Relational Databases Anil Rawat Wei Pan Chang chen Kelly Yu

Summary of the results

• The XRU relabeling scheme achieves O (log n) cost for random and frequent insertions.

• The SS-A and SS-D algorithm proposed in this paper outperform all other structural join algorithms.