余國倩1，陶光毅2，趙天宇3，劉 冰2

(1. 山東省水文局，山東 濟南 250002;2. 山東國基光曄信息科技有限公司，山東 濟南 250021;3. 南水北調東線山東干線有限責任公司，山東 濟南 250019)

　　摘 要：針對水文結構化和非結構化數據中的文字信息分析，以及從分析結果直接調取原始數據進行溯源以驗證數據的真實性，提出基于非關系型數據庫的水文數據分析方法。將海量水文數據存儲在 TRIP 非關系型數據庫，搜索引擎集成在數據庫，利用索引、中文自動分詞、檢索和統計分析功能，對水文數據中的數值、日期、時間、文字內容進行分析。對水文觀測項目文字內容的分析，獲得水文測站的水文觀測項目分布;對長序列、全時段分鐘降水數據的分析，獲得降水變化特征;對單站單日降水量的分析，獲得單站單日降水量的分布，讀取數據庫中降水自記紙圖像文件并進行溯源。研究結果表明：在非關系型數據庫存儲海量水文結構化和非結構化數據的基礎上，能夠對海量水文數據進行分析和數據溯源，具有現實的應用價值。

　　關鍵詞：水文數據;非關系型數據庫;分鐘降水數據;數據分析;數據溯源

　　Methodology of hydrologic data analysis based on non-relational database

　　YU Guoqian1，TAO Guangyi2，ZHAO Tianyu3，LIU Bing2

　　(1. Shandong Provincial Hydrology Bureau, Ji’nan 250002, China;

　　2. Shandong Guojiguangye Information Technology Co., Ji’nan 250021, China;

　　3. Shandong Main Line Co., Ltd., East Route Project of South-to-North Water Diversion), Ji’nan 250019, China)

　　Abstract：This article explains how to use non-relational database to analyze hydrologic data. By using the non-relational database, people can analyze text information of hydrologic data either structured or unstructured, and verify authenticity of these data by drawing them from the analysis results and tracing them to the source. Massive hydrologic data can be stored in the TRIP non-relational database with a search engine. Options of index, Chinese automatic segmentation, retrieval and statistical analysis help users analyze the numerical value, date, time and text contents of data. It analyze texts of observation items to get distribution of those items, analyze one-minute precipitation of all intervals in a long term to get variation characteristics of precipitation, analyzes one-day precipitation for a station to get precipitation pattern on that day and fetches files of precipitation recording in the database to verify data’s traceability. What’s revealed in this article is that the non-relational database, where hydrologic data, either structured or unstructured, can be massively stored, analyzed and verified their traceability, holds the potential for wide application.

　　Key words：hydrologic data; non-relational database; minute precipitation data; data analysis; data traceability