可配置组合式数据校验方法

摘要

In the construction of some information system for oil&gas exploration and production, it’s likely to result in irregularity of reported data types, disunity of data accuracy, inconsistency of data statistics and dissatisfaction of interrelated data items to business rules. This is mainly due to the coverage of exploration, development and engineering business, the large quantity of reported data items and data, as well as the variety of reported ways. In order to solve these problems, we summarize six verification methods including data item, column, row, cross multi-table, historical data, Singular Value, and also their combined verification methods. We also design some verification policy on cross-business and cross-time comparison and Singular value analysis, which is in accord with petroleum exploration and production industry. On this basis, we further develop encapsulated technology of configuration and maintenance. It promotes the implementation efficiency of verification method.%在油气勘探生产信息系统建设过程中，由于涵盖勘探、开发和工程等业务，上报数据项多，数据量大和上报方式多样化的情况，容易出现上报数据类型不规范、数据精度不统一、数据统计时间不一致、相互关联的数据项不满足业务规则的问题。为解决这些问题，总结出数据项校验、列校验、行校验、跨表校验、历史数据校验和奇异值校验六种数据校验方法及组合式数据校验方法，并在跨业务和跨时间对比与奇异值分析上设计出符合石油勘探生产行业的校验规则。在此基础上，研发了可配置易维护的封装技术，提高了数据校验方法的实施效率。