Index generation method and device based on KV storage, equipment and medium

Abstract

The invention relates to the technical field of distributed deployment, and discloses an index generation method, device and equipment based on KV storage, and a medium, and the method comprises the steps: obtaining to-be-generated index data which comprises a plurality of pieces of target service data, service classifications and service labels; determining service category data according to theindex data to be generated; according to all the target service data, the service classification, the service label, the service category data and the algorithm data, establishing an inverted index according to a KV generation rule to obtain index data to be evaluated, the index data to be evaluated comprising Key data and Value data; and performing effective estimation and effectiveness processing on the to-be-evaluated index data to obtain target index data. Therefore, enough inverted indexes can be constructed when the number of the candidate sets is small, multi-dimensional query is supported, recall requirements of most of recommended services are met, dimension disasters and data explosion of data in the target index data are avoided, and the effectiveness of the target index data isimproved.

Description

technical field [0001] The present application relates to the technical field of distributed deployment, in particular to a KV storage-based index generation method, device, device and medium. Background technique [0002] In the recommendation scenario, the most critical thing is how to store the candidate set, which can not only meet the high-performance requirements, but also meet the multi-dimensional requirements. Among many databases, high-performance distributed storage engines - ES or Redis - are often chosen. [0003] As a search engine, ES (Elasticsearch) automatically builds indexes, supports multi-dimensional combined queries, and quickly searches out the desired data, which is loved by developers. However, ES maintenance is required when there are not many candidate sets and needs to meet the needs of various scenarios. The cost far outweighs its contribution. At this time, developers are more inclined to use Redis's KV (Key-Value) storage engine, which can me...

AI Technical Summary

Problems solved by technology

[0004] The main purpose of this application is to provide an index generation method, device, device, and medium based on KV storage, aiming to solve the problem of how to meet the requirements of multiple scenarios when using the KV storage engine of Redis in the prior art. Build an inverted index to support multi-dimensional queries, and there is no technical problem with an effective construction method

Images