Method and device for data partitioning in key-value database

Abstract

The invention discloses a method and a device for data partitioning in a key-value database. The method comprises the following steps of: resolving partitioning number information and partitioning algorithm information from a pre-stored partitioning information basic list; and running a corresponding partitioning algorithm by a partitioning callback function according to the partitioning algorithm information, wherein the corresponding partitioning algorithm determines information of partitioned sub databases of key-value data to be partitioned according to the partitioning number information and a calculation value; and the calculation value is a value relevant to the key-value data to be partitioned. The partitioning number information and the partitioning algorithm information are recorded into the partitioning information basic list; when the key-value data are subjected to operation such as storage and query, the partitioned sub databases are determined for the key-value data according to contents recorded in the partitioning information basic list; and therefore, the partitioning mode can be flexibly changed by modifying the contents in the partitioning information basic list.

Description

technical field [0001] The invention relates to the technical field of databases, in particular to a method and device for partitioning data in a key-value database. Background technique [0002] Massive data storage with high reliability and scalability is a huge challenge for Internet companies. Traditional databases are often difficult to meet this demand, and most of the searches for specific systems are based on primary keys. Query, using a relational database in this case will make it inefficient, and expansion will become a big problem in the future. Under such circumstances, the use of key-value (key-value) databases for storage and query is widely used. [0003] In the key-value distributed storage system, the key value (key) and the content (value) are stored correspondingly. When querying, the corresponding content value can be quickly queried through the key value key, which has the advantages of fast query speed, large amount of stored data, The advantage of s...

AI Technical Summary

Problems solved by technology

[0005] In the existing technology, the partitioning of the key-value database only supports partitioning based on the number of partitions to be partitioned, and the partitioning method is single; this fixed mode partitioning method makes the partitioned database not necessarily well adapted to specific businesses Requirements, in the case of database expansion or changing query methods, have very large limitations

[0006] Moreover, the work of partitioning the key-value database is usually completed when the database is created. After the creation is completed, the partition method cannot be changed according to the future situation. That is to say, in the prior art, the database is created and partitioned After that, during the operation of the database, such as querying and adding data, the partition mode can no longer be changed

This makes the database inflexible and unable to make corresponding changes according to changes in actual conditions after a period of use

Images