Design And Implementation Of Geographic Name And Poi Data Retrieval System Based On Elasticsearch

In the context of the era of big data,people pay more attention to data mining,data analysis and data retrieval.The geographical name information data has a special strategic position under the boost of mobile internet.From the national level,geographical name information is an important part of national geographic information construction;from the people's livelihood level,with the continuous warming of the mobile Internet and global positioning system,location-based service technology is widely used in all walks of life.Nowadays,people are pursuing more intelligent and precise results with geographical limits.Therefore,how to effectively combine traditional search and place name information retrieval is particularly important.The traditional way of storing geographical names information is to use relational databases.This method has many drawbacks.For example,it does not support massive data storage.The table schema lacks flexibility and scalability,and the search method provided is simple.It only supports fuzzy queries for keywords,and the matching efficiency is low.It does not support for spatial location retrieval.It cannot satisfy people's need for place name retrieval.The emerging No SQL database HBase has the advantages of distributed structure,high performance,flexible storage modes,and support for mass storage.HBase can solve the problem of storing geographical name data.However,HBase provides a single retrieval service with only a primary key index and does not provide a secondary index.It does not provide direct support for spatial data storage and management.In view of the above problems,the thesis designs and implements a Web-based geo-name and POI data retrieval system based on the core technologies such as HBase and Elasticsearch.The system implements functions such as keyword search,point of interest type search,self-recommendation,range search,area search,and combined query with the advantages of rapid retrieval and diversification of queries.The main work of this thesis is as follows:(1)The structure characteristics of place name data and POI data are studied.In terms of its unstructured features and spatial location attributes,a HBase table model suitable for the retrieval and processing of place name data is designed using the No SQL database HBase as a storage layer;(2)The secondary indexing scheme based on HBase is studied.Through analyzing their advantages and disadvantages,the full-text search engine Elasticsearch is used to establish secondary indexes for HBase.The data is separated from the index.Through using the powerful search service provided by Elasticsearch,the system provides users with efficient search efficiency and diversified query services.At the same time,the system uses HBase's co-processing mechanism to achieve data and index synchronization.(3)Research full-text search engine named Elasticsearch based on Lucene.Analyze its overall framework,operating mechanism,principle and application.Deploy ES clusters,compile and install auxiliary plug-ins,monitor,schedule and tune the clusters.Finally,the service system was deployed on stand-alone and clusters.Based on 5 million national names and 1.26 million Xi'an POI data,the system was tested for server and client functions.Through the test tool,system performance is tested.Experimental results show that the system provides diversified query services and millisecond-level retrieval response times.At the same time,the system has good scalability and high throughput.