A spatiotemporal data indexing method in a non-relational database

Abstract

The invention discloses a spatio-temporal data indexing method in a non-relational database. A spatial index based on a quadtree and a Hilbert curve is established for spatio-temporal data in a distributed environment; temporal attributes of the spatio-temporal data are combined to form a unified spatio-temporal index; the spatio-temporal index is divided into two parts of partition keys and sorting keys, and consistent hash is used to realize distributed storage and routing of spatial data in a dynamic network topology; and when a query request is received, the partition keys and the sorting keys of a query range are determined according to temporal and spatial ranges, retrieval is carried out in storage nodes, and a query result is returned. According to the method, the storage and query efficiency of the spatio-temporal data is improved, each storage node only needs to maintain allocated data thereof, addition or deletion of the storage nodes will not cause an index error, real-time performance requirements on data accessing under a high-concurrency condition are met, and the database extendibility is good.

Description

technical field [0001] The invention relates to the technical field of databases, in particular to a spatio-temporal data indexing method in a non-relational database. Background technique [0002] With the development of technology, big data widely exists in all walks of life, such as enterprise data, medical data, Internet data, Internet of Things data, etc., including structured, semi-structured and unstructured data, and there are many types of data And the number is large, and its scale or complexity has far exceeded the range of data sets that can be managed and processed by commonly used traditional database and software technologies. [0003] Big data brings new development opportunities for all walks of life, but also poses challenges for the effective management and utilization of big data. Although there are certain differences between different types of massive data, in general, a system that supports massive data management should have the following characteris...

AI Technical Summary

Problems solved by technology

[0006] Traditional spatial indexes include quadtrees, R-trees, etc., but for the storage of spatiotemporal big data, these index structures expose the following deficiencies: (1) Not suitable for databases that store data in Key-Value (bondingbox The information cannot be converted into 1-dimensional information and used as part of the Key value for retrieval); (2) For distributed databases, the cost of maintaining and updating indexes is high; (3) For distributed databases (especially master-slave architecture), when When data is updated, it is difficult to guarantee the consistency of the indexes of each node (especially in the case of high concurrent reading and writing); (4) For database servers, the index needs to be resident in memory, which consumes too much memory resources and is not suitable for massive data Build an index; (5) When the amount of data reaches a certain level, it takes a long time to build an index, and the index file is large

[0007] It can be seen that the use of non-relational databases to store spatio-temporal big data has the problem of indexing. The use of traditional indexing methods leads to low efficiency of data storage and access, and it is difficult to meet the real-time requirements of high concurrency and large data volumes. The performance advantages of non-relational databases are obtained. less than play

Images