POI结构化和非结构化数据存储检索策略

POI（兴趣点）数据具有结构化和非结构化属性，使用关系型数据库存储时，非结构化占用内存空间大，检索时直接从数据库读取，且POI空间信息需要通过HBase的rowkey进行多条件检索，效率低下。通过研究POI数据特点，构建数据模型，针对POI结构化属性数据和非结构化数据提出不同的存储和检索策略并创建索引。最后使用Sqoop 完成 POI 属性数据和非结构化数据的交互并嵌入Solr 技术进行检索。经实验验证，应用该方法针对POI非结构化数据检索效率有所提高。

The Storage and Retrieval Strategy Structured and Unstructured Data of POI

POI data has both structured and unstructured attributes.When relational database storage is used, unstructured storage occupies large memory space,The data is directly retrieved from the database. In addition, the HBase rowkey is used for multi-condition retrieval of POI space information, which is inefficient.By studying the characteristics of POI data and building a data model,Different storage and retrieval strategies are proposed for structured attribute data and unstructured data of POI and create index.Finally, Sqoop is used to complete the interaction between POI attribute data and unstructured data, and Solr technology is embedded for retrieval.Experimental results show that the proposed method can improve the efficiency of POI unstructured data retrieval.