International Journal of Information Technology & Systems, Vol. 1; No. 1: ISSN: 2277-9825 (Jan-June 2012) www.gtia.co.in 35 Computing the Data Cube Cells Using Control Charts K.Dhanasree* , Dr.C.Shobha Bindu** *Assistant professor,DRK Institute of Tech ,Hyderabad. ** Associate professor, JNTUC, Anantapur. Email: dsree_99@yahoo.co.in, shoba_bindhu@yahoo.co.in Abstract: Most of the modern OLAP applications are frequently pre computing the multiple group bys.The pre computation of data cells is critically required to increase the response time. However in a distributed environment the user may not be intrested with all the pre computed cells. If the whole data cube is pre computed then the query response time will be less.This paper presents which data cells are to be pre computed owing to user requests. Here we present a pictorial approach called control charts to pre compute the required cells, there by increasing the query performance. Keywords: OLAP, data cube,lattice, control limits, contol charts. 1 Introduction In recent years, data in the real world has grown immensely. There by we are at a greater search for techniques to uncover useful information from these larger amounts of data. Data mining is the most powerful technique for extracting and analyzing useful information from large amounts of data. Since data in the data warehouse is of very high volume, There needs to be mechanisms like data summarizations in order to get only the relevant and meaningful information in less messy way. Data mining tools like OLAP [1] serves these summarized computations in a very efficient way. OLAP applications are at a great advent in solving modern business problems. OLAP software helps analysts and managers to gain insight into performance of an enterprise through a wide variety of views of data organized to reflect the multidimensional nature of enterprise data. OLAP characterizes the operations of summarizing, consolidating, viewing, applying formulae to, and synthesizing data along multiple dimensions. Since we are required to retrieve large number of records from different data bases, we are at an urge to summarize them on multiple dimensions. This multidimensional nature of data has led to OLAP applications. Recently introduced data cube operator is supporting such aggregates in OLAP data bases. A data cube allows data to be modeled and viewed in multiple dimensions. Many of OLAP tools can be used to compute the data cube. A data cube consists of two kinds of attributes : measures and dimensions. The dimensions consist of attributes like sales, city, time period etc. The measures are numeric attributes like profit, total sales etc. The pre computation of all or part of data cube can greatly reduce the response time and enhance the performance of online analytical processing. To support the goal of an OLAP application the most efficient ways are pre compute all cells in the cube, or pre compute no cells , or pre compute some of the cells. If the whole cube is pre computed then the query response time is faster. But the disadvantage is pre computation requires lot of memory. We can pre compute none of the cells in order to minimize memory requirements. The disadvantage is user query response time is slow. With these two disadvantages we pre compute the user required cells. The question is, which cells are to be pre computed to increase the decision support query retrieval. In our work we present a pictorial analyzation of which data cubes are to be pre computed using control charts. These control charts also present a way for pre computation and providing the availability of these pre computed cells in a distributed environment. 2 Data Cube For supporting decision queries effectively, a new operator, CUBE BY, was proposed [2]. It is a multidimensional extension of the relational operator GROUP BY. The CUBE BY operator computes groupbys corresponding to all possible combinations of grouping attributes in the CUBE BY clause. A cell (cuboids) of a cube is one group by. As the cube by attributes increases cube operator becomes more expensive. The huge size of a data cube makes data cube computation time-consuming [3]. Recently several methods have been introduced to reduce the size of a data cube and hence its computation time and storage overhead are reduced. condensed cube, Dwarf , Quotient cube and indexing using QC-trees[4] are some