| Adopting the structure of a kind of concentrating type in the data warehouse, it stores the data of the whole company in a data base ( repository) in unison. Usually, its data appear by way of data cube, and multidimensional modes of the storage can make different looking over and various kinds of associations to data. These data are like numerous trees in one large stretch of forest, the data administrator must hack one's way through all difficulties , and then could see the implicit meaning within them after doing the association of some dependence of these company data. OLAP different from data warehouse makes the business software catalogue , let users operate by data cube. It is concentrated that typical OLAP operation includes the data (consolidate), lay visit , cuts into slices , dices and the pivot . The result produced can appear with the database form of tabulation or in traditional way, can also even do them into charts . Though such a output may be only a kind of regular form , it usually allows user's direct operating data to make further analysis, for example verify the trend , array of dependence (correlation ) or time ( time series ) ,etc.. Using the application program on Web to analyse and process on the line ( OLAP, on-line analytical processing), can strengthen the might of Web technology greatly. The excavating of OLAM (OLAP Mining) is based on foundation of the data cube analysing and processing, and, in practical application, though multidimension of OLAM may need more dimension and strong visit tool to calculate, we can conclude that data cube of OLAM and data cube of OLAP have no essential difference. We can see , OLAM server receives users' analysis order through the user figure interface . under guidance of meta data , OLAM make certain operation to super cube ,and then represent excavated analysis results to users. This course is dynamic. The multidimensional data view (ultra data cube ) is the foundation of OLAM. The organization way of the multidimensional view plays an essential role to the execution efficiency and response speed of the system. We have carried on deep studying to this technology through three respects of data cube ----memory, computation, materialization tactics, and has made the satisfactory result.A series of complicated table excavate to database is excavate calculating to data cube after all. In this case the request arithmetical accuracy simply is given up originally and chased unfinishedly. A bit fuzzy on certain level can not only improve the speed excavated , but also there are no obvious losses on accuracy. It is important ways that the cube data are compressed to improve multidimensional data warehouse performance. It is the main application on the data warehouse to analyse and process on-line, and is one of the most frequently used operation while analysing and processing cube operates. The research of Cube algorithm compressed on the multidimensional data warehouse is the important task with challenge that the database circle faces. In recent years, people have launched a large amount of work in Cube algorithm, but seldom involve the multidimensional data warehouse and compress the multidimensional data warehouse . This text adopts to observation of every element value in the data cube basis on the further investigated and compressed data warehouse , we found that it was very similar for each other in cubedata. Have proposed compressing Cube algorithm of the multidimensional data warehouse, we carries out it to operate in compressing the data directly, needn't uncompress ,and have improved the processing speeds of Cube.assemble calculating is a kind of operation of occupying an leading position in DSS and data cube call that multidimensional database too which is a commonly used technology. Its main thought is make those costly daily operation together, for instance: Count, Sum , Average , Max , Min ,etc. then calculates out the result in advance , and store in a multidimensional database according to different attribute .As a case in...
|
PDF Full Text Request