Research On XML Data Cube Based On Tree Structure

Following the development of computer science, the internet has been playing more and more important role in people’s daily life. The web was only formed by linked static files originally, and now it has become the interactive social media. It will be the information center base on everyone in this world in the foreseeable future. People are gradually attaching importance to how to use information on the web efficiently. Extensible Markup Language (XML) is one good tool to get information on the web. With feature of simplicity, flexibility and cross-platform, XML has been more and more used for data storing and exchanging. Because of this trend, designing data warehouse based on XML also becomes a hot topic in realm of data warehouse researching now. But the flexibility of structure of XML brings a challenge to this researching. In this thesis, a data cube model named XTree Cube whose constructing data cube based on a tree structure is proposed. XTree Cube transforms dimensions of data cube into a tree-style structure and stores them in an XML file. This method reduces the redundancy causing by saving dimension relationship in XML data cube and increases the efficiency of constructing XML data cube and data querying. XTree Cube also takes advantage of flexibility of construct of the tree-style to solve questions in data cube constructing and data querying causing by complicated inheritance relationship between dimensions in the process of constructing XML data cube. The main work in this thesis is as following:(1) Designing XTree Cube model, constraint and algorithm. Discussing OLAP based on XTree Cube and resolution for designing tree model based on complicated dimension data source.(2) Building an XTree cube which data based on a real data source. After the cube being built, the thesis executes the data querying process, and shows the process to how to update the cube contents when the data contents in data source increasing.(3) The thesis compares to star model, XCube model and XTree Cube model by experiment analysis. It includes time and space efficiency for building data cube, time efficiency for data querying based on the built data cube, and time efficiency for data update in three models. The thesis discusses the current hot topic about data warehouse and proposes a new data cube model. A series of experiments show that XTree Cube has higher time and space efficiency than other models when building data cube. The XTree Cube is also faster than other data cube models when OLAP and has more feasible structure to store and represent data with complicated inheritance relationship. So the XTree Cube owns value for data ware house researching.