Vector data storage method, efficient retrieval method and device and storage medium

Abstract

The invention relates to a vector data storage method, an efficient retrieval method, equipment and a storage medium, and the vector data storage method comprises the following steps: circularly decomposing a plane space corresponding to vector data into multiple levels of grids with the same size by using a space filling curve algorithm, stringing all the grids through a curve, the whole plane space is uniformly filled with the curve; sequentially coding each minimum grid according to the segmentation hierarchy to obtain a unique code of each grid, and further obtaining a grid coordinate code corresponding to the vector data; and storing the grid coordinate codes corresponding to the vector data. By the adoption of the method, efficient retrieval of the vector data can be achieved when massive vector data are stored and managed.

Description

technical field [0001] The present application relates to the technical field of data storage and retrieval, and in particular to a vector data storage method, efficient retrieval method, device and storage medium. Background technique [0002] Presto usually handles the user's query statement request through the following process: the client sends the query statement to the coordinator of the Presto cluster through the http protocol, and the coordinator receives the query statement sent by the client, parses the statement, and generates a query execution plan , and generate SQL query executors in turn according to the query execution plan (for query SQL parsing and query plan construction) -> SQL stage executor (for dividing the built query plan into multi-layer stage tasks) -> remote http (Hypertext Transfer Protocol) task; the coordinator distributes each task to the work node where the data to be processed is located for execution; the task of executing the source ...

AI Technical Summary

Problems solved by technology

[0003] Although presto can realize data management of major data storage engines (such as kudu, hbase, etc.) through connectors, such as: data storage, data retrieval and data analysis, but this technology does not support index query and management of vector data, and cannot be realized Search vector data

[0004] In the prior art, there is a method of establishing a multi-level depth table based on the spatial index to realize the storage and indexing of vector data, but this method is only suitable for the efficient visualization of vector data when mobile devices such as mobile phones publish and browse maps, and cannot Realize efficient retrieval of massive big data in general scenarios

Images