大型企业报表引擎的优化设计与实现

Optimization Design and Implementation of Large Enterprise Reporting Engine

报表是企业级应用系统的重要组成部分,优秀的报表可以对基础数据进行过滤、合并、统计、分析,提供直观的表达形式,并为决策提供良好的依据。目前大型企业基于B/S的报表系统开发模式相对不成熟,开发人员大都是根据实际需求直接写代码实现数据的分析、处理、展现功能,这样的实现方式会做很多重复工作,比如很多项目中都需要通过报表数据和图表来展现业务趋势,至于展现的时间粒度、地域粒度、数据纬度和指标却各不相同,结果是不同的。项目中类似的功能需要重复开发,总结起来就是可重用性低、可扩展性差、可维护性差、开发成本高。为解决以上问题,文中提出对业务需求进行高层次的抽象,实现一个能够统一分析数据、处理数据、展现信息的优化报表引擎,该报表引擎专注于业务逻辑的实现,与具体的业务分离,能帮助开发人员快速稳定地构建报表应用系统。基于该报表引擎开发的报表系统已应用于某省级电信运营商,从实际的运行情况看,系统运行稳定,并表现出良好的可扩展性,经过多用户并发查询、大数据量查询的测试,性能表现良好。

The report is the important constituent of enterprise level application system, excellent reports on basic data filtering ,merging, statistics, analysis, provides an intuitive form of expression, and provides a good basis for decision-making. At present, large enterprises based on B/S reporting system development pattern is relatively immature, developers are mostly based on actual demand to write code directly to realize the data analysis,processing, display function, such implementations will do a lot of repetitive work, for example, many projects need to pass the report data and charts to show business trends, as demonstrated in the-time granularity, granularity, geographical latitude and index data are different, the results are different. Project of similar function requires repeated development, summed up is reus- ability is low, poor scalability and maintainability, the high cost of development. In order to solve the problem above, present the business needs for a high level of abstraction,realizing a optimized reporting engine of unified analysis data, processing data, displaying informa- tion, the report engine focused on the business logic to achieve, and the specific business separation, can help developers quickly build a stable report application system. Based on the report engine, the report system has been applied in a provincial telecom operators, from the actual operation situation, the system operation is stable, and exhibit good scalability, after the test of many concurrent users query, large amount of data query, show the good performance.