A query method and query device

Abstract

The embodiment of the present invention discloses a query method and a query device. The method can determine N execution plans corresponding to each of the N query requests according to the received N query requests, and determine at least Two identical first sub-plans, generating a first sub-query result corresponding to any first sub-plan in the at least two identical first sub-plans, and generating a result corresponding to the query according to the N query requests In the process of query results corresponding to each of the N query requests, all the same first sub-plans reuse a first sub-query result corresponding to any of the first sub-plans. It can be seen that, Using the query method shown in this embodiment can reduce a large number of repeated calculations, save the resources of the database system, and improve the query efficiency and throughput of the database system in the application scenario of large-scale concurrent query.

Description

technical field [0001] The invention relates to database technology, in particular to a query method and a query device capable of realizing large-scale concurrent query. Background technique [0002] Large-scale concurrent query is an important database application scenario in the era of big data, and it is also an important criterion for testing database capabilities. Large-scale concurrent queries will consume a large amount of database system resources in a short period of time and put a lot of pressure on the database system. In order to ensure that users can quickly obtain query results, the query response time must be controlled within the tolerance range. The application scenario of large-scale concurrent queries will cause a sharp increase in database pressure in a short period of time, but the duration of the peak period is short. [0003] A large-scale concurrent query method is provided in the prior art, multiple queries are compiled together, that is, figure 1...

AI Technical Summary

Problems solved by technology

However, if the query changes, the entire integrated execution plan needs to be regenerated. The integrated execution plan is a static structure, and its execution cannot be split. Each execution of the integrated execution plan is equivalent to executing all the included queries once.

[0004] The large-scale concurrent query method shown in the prior art is suitable for solidified query scenarios, but in practical applications, the query will change frequently, and as long as one of the multiple queries is changed, the integrated execution plan will be changed as a whole. Change, which wastes a lot of execution plan generation time, thus prolonging the query response time and reducing query efficiency

Images