Sybase数据仓库解决方案.pptVIP

THIS BUILD SLIDE EXPLAINS THE PROCESS INVOLVED IN BUILDING A DATA MART - IT HAS SOME SIMILARITIES TO THE PRODUCT ARCHITECTURE SLIDE. You typically start with (CLICK) a source and target database. When you buy PowerMart, you get (CLICK) the Informatica Server and Client Tools. The Repository Manager (CLICK) is used to build the Metadata Repository, where all of the dynamically generated metadata will be stored. Within the Designer, (CLICK) there are four tools to facilitate the Design process. The Source Analyzer (CLICK) is used to analyze source definitions, which are stored in the metadata repository. The next step is to use the Warehouse Designer (CLICK) to read and store the target definitions. The Transformation Developer and Mapping Designer are used to create re-usable transformations and to build the transformation pipeline. Once you’ve done that, you can use Server Manager (CLICK) to create a load session and set the properties and schedule for extracting, transforming and loading your data. At the scheduled time, the server engine is invoked (CLICK) and data is extracted from the source system, transformed within the engine, and loaded into the target warehouse. For the state column of a table, a LowFast index would be appropriate since there are only 50 distinct values. Sybase IQ will build 50 bitmaps, one for each state. For processing this query, Sybase IQ will select the bitmap for “AL” and count the “1”s. “How many MALES are NOT INSURED in CALIFORNIA? Gender M M F M M - 800 Bytes/Row 10M ROWS State NYCACTMA CA - RDBMS Insured YYN Y N M Y CA M N CA F Y NY M N CA 1 2 4 3 Gender Insured State + + 1 1 0 1 1 1 0 1 0 1 0 1 10M Bits 10M Bits x 3 col / 8 16K Page = 235 I/Os 800 Bytes x 10M 16K Page = 500,000 I/Os 基本上只能使用表扫描 查询过程读取了太多的无效数据 IQ Example: I/O 的明显减少 IQ的索引特点 索引即是数据 没有索引和数据的分别 任何一列可以建立多个索引 系统保证至少会存在一个索引（FP） 索引的选择和设计主要基于： 数据的级数（离散值的个数） 在查询中的使用方式 和SQL语句的顺序无关 索引的种类 Fast Projection(FP) 数据压缩存储 根据数据的特点会自动使用三种方式中的一种 Low Fast (LF) Bit map 索引 H