陳真玄 1，王 琳 1，張 怡 2，孟慶學 3

　　(1.水利部信息中心，北京 100053;

　　2.北京金水信息技術發展有限公司，北京 100053;

　　3.北京沛海科技有限公司，北京 100083)

　　摘 要：如何更加高效地存儲和處理海量非結構化數據，已成為水利行業信息化建設亟待解決的難題，水文預報業務作為水利行業的核心業務，選擇合適的數據庫是首要任務。對目前國產主流數據庫進行梳理，將非結構化數據庫分為數據庫結合 NAS、數據庫大字段、多模數據庫 3 種類型，針對每種類型，分析原理、特點、數據處理時效性、支持的數據量、數據查詢效率等多個方面，對部分類型進行壓力測試，獲取量化性能指標，使選型者可以結合自身應用場景進行比較分析選擇合適類型，并基于水文預報業務數據庫選型經驗，提出水利非結構化數據庫選型建議。通過分析認為不同的業務需求適合不同的數據庫類型，對于水文預報業務這類海量非結構化數據存儲管理、且數據處理能力要求高的場景建議選用數據庫大字段或多模數據庫。

　　關鍵詞：水文預報業務;非結構化數據;數據庫選型;NAS;數據庫大字段;多模數據庫

　　Discussion on selection of unstructured database for hydrological forecasting

　　CHEN Zhenxuan 1，WANG Lin 1，ZHANG Yi 2，MENG Qingxue 3

　　(1.Information Center，Ministry of Water Resources，Beijing 100053，China;

　　2.Beijing Jinshui Information Technology Development Co.，Ltd.，Beijing 100053，China;

　　3.Beijing PowehiDB Technology Co.，Ltd.，Beijing 100083，China)

　　Abstract：How to store and process massive unstructured data more efficiently has become an urgent problem to be solved in the information development of the water conservancy industry. As the core business of the water conservancy industry, hydrological forecasting is the fundamental task to select an appropriate database. By researching on the current domestic mainstream databases, the unstructured databases are categorized into three types: database combined with NAS, database with large fields and multi-mode database. For each type, this paper analyzes its principle, features, effectiveness of data processing, supported data size, efficiency of data query and other aspects. Pressure tests are conducted on some types to obtain quantitative performance indicators, so that the users can select the appropriate type through comparative analysis in combination with their own application scenarios, and propose suggestions on unstructured database selection of water conservancy based on the database selection experience of hydrological forecasting business. The analysis shows that, different database types are suitable for different business needs. For hydrological forecasting business with high requirements for storage and management of massive unstructured data, as well as for data processing capacity, it is recommended to select database with large fields or multi-mode database.

　　Key words：hydrological forecasting;unstructured data;database selection;NAS;database with large fields;multi-mode database