A method and device for realizing data aggregation based on clickhouse and java

By using ClickHouse to generate materialized views and Java scheduled tasks to handle special metrics, the problem of low query response efficiency under large data volumes is solved, and efficient metric calculation and storage are achieved.

CN117785922BActive Publication Date: 2026-06-26WUHAN GREENET INFORMATION SERVICE
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
WUHAN GREENET INFORMATION SERVICE
Filing Date
2023-12-22
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

In existing technologies, query response efficiency for indicators with large data volumes and special calculation requirements is low, and they cannot be implemented solely through SQL at the database layer. Instead, they require Java program logic to perform indicator calculations.

Method used

ClickHouse is used to generate materialized view database operation instructions, static feature items and cumulative item execution code are added, and analysis code for generating special result items is combined with Java program. Special result items are processed through scheduled tasks to reduce indicator writing latency. Combined with Java scheduled tasks, it can make up for the indicator generation that materialized views cannot complete.

Benefits of technology

It effectively reduced the query response latency of indicators, improved query efficiency, and solved the query response efficiency problem of special calculation indicators under large data volumes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117785922B_ABST
    Figure CN117785922B_ABST
Patent Text Reader

Abstract

The application relates to the technical field of databases, and provides a method and device for realizing data convergence based on clickhouse and java. A materialized view database operation instruction is generated, a static feature item and / or an accumulated item execution code are added in the corresponding materialized view database operation instruction, a first index generation program instruction is obtained, analysis code of a corresponding special result item is generated by using a java program, the analysis code is stored as a second index generation class, an analysis period of the analysis code is obtained, and each execution process is performed, so that an original detailed list table of original data is processed. When the materialized view is used, the delay of writing the index into the table can be reduced to the minimum, and the java timing task is combined to compensate for the generation of other index items which cannot be completed by the materialized view in the index table.
Need to check novelty before this filing date? Find Prior Art