Improving XML Processing Using Adapted Data Structures

Neumüller, Mathias; Wilson, John N.

doi:10.1007/3-540-36560-5_16

Cited by 5 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The authors are aware of only two published works on querying compressed XML, XGRIND [20] and DDOM [18], which both simply compress character data in a query-friendly fashion. No attempt is made to compress the skeleton or to do any form of "global" compression.…”

Section: Introductionmentioning

confidence: 99%

Path Queries on Compressed XML

Buneman

Grohe

Koch

2003

Proceedings 2003 VLDB Conference

147

143

View full text Add to dashboard Cite

Central to any XML query language is a path language such as XPath which operates on the tree structure of the XML document. We demonstrate in this paper that the tree structure can be effectively compressed and manipulated using techniques derived from symbolic model checking. Specifically, we show first that succinct representations of document tree structures based on sharing subtrees are highly effective. Second, we show that compressed structures can be queried directly and efficiently through a process of manipulating selections of nodes and partial decompression. We study both the theoretical and experimental properties of this technique and provide algorithms for querying our compressed instances using node-selecting path query languages such as XPath. We believe the ability to store and manipulate large portions of the structure of very large XML documents in main memory is crucial to the development of efficient, scalable native XML databases and query engines.

show abstract

Section: Introductionmentioning

confidence: 99%

Path Queries on Compressed XML

Buneman

Grohe

Koch

2003

Proceedings 2003 VLDB Conference

147

143

View full text Add to dashboard Cite

show abstract

“…Alternative approaches involve the use of schema information to drive the compression (Bőttcher et al, 2007). In earlier work, we investigated the use of dictionary compression techniques for representing both tags and values in XML structures (Neumüller et al, 2003;Gourlay et al, 2007) and other investigators have also examined the decomposition of XML into array-based structures (Buneman et al, 2005).…”

Section: Related Workmentioning

confidence: 99%

Efficient data representation for XML in peer‐based systems

Tripney

Foley²,

Gourlay

et al. 2010

International Journal of Web Information Systems

View full text Add to dashboard Cite

Purpose -New directions in the provision of end-user computing experiences mean that we need to determine the best way to share data between small mobile computing devices. Partitioning large structures so that they can be shared efficiently provides a basis for data-intensive applications on such platforms. The partitioned structure can be compressed using dictionary-based approaches and then directly queried without firstly decompressing the whole structure. Design/methodology/approachWe describe an architecture for partitioning XML into structural and dictionary elements and the subsequent manipulation of the dictionary elements to make the best use of available space. FindingsThe results indicate that considerable savings are available by removing duplicate dictionaries. We also identify the most effective strategy for defining dictionary scope. Research limitations/implicationsOur evaluation is based on a range of benchmark XML structures and the approach to minimising dictionary size shows benefit in the majority of these. Where structures are small and regular, the benefits of efficient dictionary representation are lost. Our future research now focuses on heuristics for further partitioning of structural elements. Practical implicationsMobile applications that need access to large data collections will benefit from the findings of this research. Traditional client/server architectures are not suited to dealing with high volume demands from a multitude of small mobile devices. Peer data sharing provides a more scalable solution and the experiments that we describe demonstrate the most effective way of sharing data in this context. Social implicationsMany services are available via smartphone devices but users are wary of exploiting the full potential because of the need to conserve battery power. Our approach mitigates this challenge and consequently expands the potential for users to benefit from mobile information systems. This will have impact in areas such as advertising, entertainment and education but will depend on the acceptability of file sharing being extended from the desktop to the mobile environment. Originality/valueThe original work we have done characterises the most effective way of sharing large data sets between small mobile devices. This will save battery power on devices such as smartphones, thus providing benefits to users of such devices.Keywords Peer-to-peer, Data compression, Database management, Extensible Markup Language Paper type Research paper IntroductionGrowth in personal computer sales dropped almost to zero in 2008. Meanwhile two hundred million smartphones were sold, an increase of 13% on the previous year (Meyer, 2009). Although smartphone sales are set to exceed personal computer sales, 2008 also saw the purchase of 1.22 billion mobile phones. Given the overlap of functionality between these devices, there appears still to be a significant potential expansion available in the market for smartphone devices. Smartphone users are becoming accustomed to their needs...

show abstract

“…These allow efficient path queries over the data while still needing large parts of the data to be decompressed in the worst cases. We have already investigated the use of dictionary compression techniques for representing XML data structures [25] and other investigators have also examined the decomposition of XML into array-based structures [12].…”

Section: Related Workmentioning

confidence: 99%

Compressed Materialised Views of Semi-Structured Data

Gourlay

Tripney

Wilson

2007

24th British National Conference on Databases (BNCOD'07)

View full text Add to dashboard Cite

Query performance issues over semi-structured data have led to the emergence of materialised XML views as a means of restricting the data structure processed by a query. However preserving the conventional representation of such views remains a significant limiting factor especially in the context of mobile devices where processing power, memory usage and bandwidth are significant factors. To explore the concept of a compressed materialised view, we extend our earlier work on structural XML compression to produce a combination of structural summarisation and data compression techniques. These techniques provide a basis for efficiently dealing with both structural queries and valuebased predicates. We evaluate the effectiveness of such a scheme, presenting results and performance measures that show advantages of using such structures.

show abstract

Improving XML Processing Using Adapted Data Structures

Cited by 5 publications

References 12 publications

Path Queries on Compressed XML

Path Queries on Compressed XML

Efficient data representation for XML in peer‐based systems

Compressed Materialised Views of Semi-Structured Data

Contact Info

Product

Resources

About