A custom-based index building method, device and system

Abstract

The present invention provides a method, device and system for establishing an index based on customization. By adding customization for domain information structures and customization of mapping information from different database structures to the same domain structure during the indexing process, the customized The mapping information encapsulates the read data, writes the encapsulated data to the index server according to the domain structure information, and then builds the index, thus realizing the addition of the retrieval system without affecting the security and efficiency of the retrieval system. The versatility and scope of application realize the high degree of customization in the indexing process.

Description

technical field [0001] The invention relates to the technical field of information retrieval, in particular to a method, device and system for establishing an index based on customization. Background technique [0002] At present, with the development of the network, more and more information retrieval systems have been maturely applied in the Internet, e-commerce and other fields. Information retrieval systems usually adopt the retrieval structure of the popular inverted index of search engines, and are embedded in the Go to the site of the modern service industry, as a favorable tool for searching within the site. However, as the business process of the site becomes more and more complex and the data accumulates more and more, the high efficiency requirements for information retrieval become more and more urgent; among them, the application fields with higher requirements mainly include: library database, Information database, patent database, medical database, office aut...

AI Technical Summary

Problems solved by technology

[0003] However, the data structures of information retrieval among different service industries are very different, and the data retrieved from the massive data of different databases in different industries are also very different. If the retrieval service is moved from one industry to another , even from one database to another, it requires a lot of manpower to configure and re-develop and deploy, which is easy to cause redundant cost overhead, and is not conducive to the increasingly complex demand for information retrieval in the service industry

[0004] In addition, current information retrieval systems generally lack customizability; the various domain structures of the system and the data structures that need to be indexed are often solidified in the entire system, and the entire system can only target a specific type in a certain field. The indexing of the data structure makes the system inflexible; for example, in the news domain, when users search for news, they often search for the news title, release time or content. This kind of information is solidified in the news retrieval system; and when users need to retrieve talent information, the key words searched may be the person’s name, education level, etc., which requires retrieval of a completely different domain structure from news, and It is obviously not feasible in the traditional retrieval system that solidifies various domain structures in the system; at the same time, for the set various domain structure information, the corresponding database definitions of the same type are also very different, for example, two Completely different news databases A and B, the table structure and field names of the news information in database A are completely different from the table structure and field names of news information in database B. In database A, the news_title field represents the title of the news, while In database B, the title field is used to represent the title of the news; although the names of these two fields are different, there is no difference in their content, both of which represent the title of the news

[0005] Therefore, for the traditional retrieval system, since the indexing data structure it is oriented to is often solidified in the system, in the process of indexing the two databases with different structures on the same domain structure, It is often necessary to re-develop systems for different data structures, and the re-developed systems often have many similar logical operations, such as reading databases, building news indexes, etc., which easily causes a lot of waste of resources

Images