Research And Application Of Building Semi-structured Data Approach Based NoSQL

The data formats began to vary with the rapid development of network, but thetraditional data stored in the files and relational database which play role in thischange can’t gradually meet the processing requirements of growing data. Currentlyaccording to its structure, data can be divided into three categories:(1) Data stored inthe files with loose and not strictly structure such as audio, image et al.(2) Datastored in relational database with strict structure.(3) The semi-structured data whichbetween in before, although it unlike the relational database system that has strict andconstraint pattern, it has rules indeed.Researchers have proposed a variety of theories and methods describedsemi-structured data, and the relative data store technology is growing. Thedevelopment of XML which is a data format of exchange is to promote thedevelopment of semi-structured data and Web applications. Meanwhile,RDF(Resource Description Framework) as a standard framework of semantic data innetwork, its triples model provides a nice simple theoretical model for describingsemi-structured data in network. In addition, as a storage and management mode witha free schema model, NoSQL compensates the shortcomings of the traditionalrelational database in scalability.The main content of this paper is to introduce how to describe and store thesemi-structured data in network. This paper introduces various methods for ofdescribing and storing semi-structured data, and pointed out that the major problemsfaced. In this paper, we give the main research results and innovative points as thefollowing:(1) propose RDF-JSON semantic method: we make a study on thecharacteristics of semi-structured data in network, then propose RDF-JSON methodto describe semi-structured data in network which combines with JSON data formatand RDF, and considers the data requirements in practical application. This methodcan be well compatible with traditional JSON data and can be a semanticaldescription of semi-structured data.(2) use MongoDB database to achieve dynamicstorage: with analyzing and summarizing the dynamic characteristic ofsemi-structured data in network, we take advantage of MongoDB database with freeschema mode which stores semi-structured data to solve the dynamic needs.(3)propose a multi-documents storage solution: we make pre-operation for the data storage and take multi-documents storage. Meanwhile, we establish thecorresponding index based on the characteristics of data. The solution can increasethe retrieval rate.(4) build an application for sharing data.