Cluster database system and data query processing method thereof

Abstract

The invention discloses a data query processing method of a cluster database system. The data query processing method comprises that a query server enables a query task to be analyzed into at least one subquery task after receiving the query task; the query server sends the at least one subquery task to multiple database systems so as to execute in parallel the ate least one subquery task; first database systems in the multiple database systems carry out the subquery task while receiving relevant query result data which is obtained by executive relevant subquery tasks which are sent by other database systems; received relevant query result data are merged with query result data which is obtained by the first database system executive subquery task to form a query middle result; the query middle result serves as an input of a next subquery task, the next subquery task is carried out, or the query middle result is sent to the query server. The invention further discloses a cluster database system. By means of the cluster database system and the data query processing method thereof, overall execution time of the query can be reduced.

Description

technical field [0001] The present invention relates to the technical field of computer data processing, and more specifically, to a data query processing method of a cluster database system, a database system and a cluster database system. Background technique [0002] In the existing cluster database system, data is usually distributed in multiple database systems interconnected with a shared-nothing structure, and multiple database systems provide a single database image externally by way of framework middleware. The middleware system receives the user's SQL request, executes it in multiple underlying database systems, and then returns the final result to the user. Therefore, for users, using a cluster parallel database system is the same as using a single database. Usually, in a database, query statements are parsed (including lexical analysis, syntax analysis, and semantic decomposition and optimization) into query execution plans. The query execution plan may contain ...

AI Technical Summary

Problems solved by technology

However, in the link of intermediate results obtained by the database system through subtask queries, for each intermediate result, it is necessary to wait for the completion of all subtask queries related to it before merging the results of each subtask. A lot of time will also be wasted when subtask results are merged, resulting in a slowdown in query speed and a slowdown in response to users

Images