Tuple sorting method, query method and device in relational database management system

Abstract

The present invention provides a tuple sorting method, query method and device in a relational database management system, wherein the sorting method includes: storing the data used for tuple sorting extracted from the original structure to the basis of the application in memory In the array of data types, and store the tuple copied from the original structure in the tuple array of the original relational database structure type applied in the memory; by setting the kernel function to output the stored tuple sorting data, To use the external hardware accelerator to sort the data used for tuple sorting; receive the sorting calculation result returned by setting the kernel function, and according to the sorting calculation result, use the tuples in the applied tuple array to assign values ​​to the original tuple array , to sort the tuples in the original structure. Through the above solutions, the present invention can improve the execution efficiency of the relational database management system.

Description

technical field [0001] The invention relates to the technical field of databases, in particular to a tuple sorting method, query method and device in a relational database management system. Background technique [0002] With the rapid development of the Internet and e-commerce, more and more relational databases are used on the Internet, and they assume the core responsibility of enterprise business and information systems, so they are increasingly facing the test of ultra-large-scale data volume and high-intensity instantaneous concurrency. In response to these problems, for database enterprise-level applications, database technology with higher execution efficiency is required to ensure the interests of enterprises. [0003] When a relational database executes SQL statements such as order by and group by, it will complete the SQL statement to be queried through query compilation, query optimization, and query execution. During this process, the sorting process most affect...

AI Technical Summary

Problems solved by technology

The method of iterative mode returns a tuple each time, which requires multiple coordination between different operators, and the execution efficiency is not high; while the materialization mode processes all data at one time, when the amount of data is larger than the memory of the tuple, it will cause insufficient memory (out of memory) problem

Images