数据仓库缓慢变化维读书报告.pptVIP

“ ” “ ” 读书报告 姓名：王建荣 班级：大数据1501班 学号主要参考论文 Handling slowly changing dimensions  in data warehouses Sidra Faisal, Mansoor Sarwar? Punjab University College of Information Technology University of the Punjab, Lahore, Pakistan 报告题目：处理数据仓库中的缓慢变化维度 Framework Introduction Motivational example Types of SCDs Surrogate key-based temporal data warehouse (SKTDW) Settings for evaluation of temporal warehouses Performance results Conclusions Introduction the challenge is to crack the business value of information and utilize it for decision-making. One of these factors affect the quality of such analysis is slowly changing dimensions (SCDs) SCDs should maintain consistency and correctness of data The other problem that has been raised by SCDs is of query performance for state-oriented data representation. When representing state-oriented data using a time-series representation, information may be lost in cases where the sampling frequency is low, or redundancy may occur in cases where the sampling frequency is high. SCDs can be handled through regular star schema using surrogate keys and temporal star schema, but they lead to complex queries and temporal joins, respectively. Motivational example Fig. 1. Business entities and relationships among them are: Every product has a unique identifier (PId – natural key)and contains the attributes name (PName), price, and packaging. Every product belongs to a particular category. Every store has a unique identifier (SId)and a location. A row in sales fact table may show the information like, 100 items of product A from category C1 are sold in store S1 during August 2010. Here, a single fact table row represents the sales record of a particular product in a particular store on a given day. Thus, the granularity of measure values is product-by store-by day. Types of SCDs Type 1: New data overwrites old data in a dimension’s attribute and no history is maintain