Method and system for implementing database access

Abstract

The invention discloses a method and a system for implementing database access. The method comprises the following steps of: forming a Bucket instance by using the same class of structured query language (SQL); configuring various SQL and Bucket in a data access configuration file; if non cross-statement database access is performed, acquiring Pooled Statement by using a borrow Statement () interface method; if cross-statement database access is performed, acquiring the Bucket by using a borrow Bucket () interface method, and acquiring the Pooled Statement therein one by one by using a get Statement () method provided by the Bucket; and performing data operation of the database access by using an interface method provided by the Pooled Statement. The method and the system can implement high-availability, high-performance, low-cost, and easily managed, adjusted and optimized database access for a distributed system of massive data.

Description

technical field [0001] The present invention relates to information application technology, in particular to a method and system for realizing database access. Background technique [0002] Data access technology is an important issue to be considered in the distributed system of massive data. [0003] The current data storage generally adopts relational database. In Java implementation, the conversion between relational database and Java object is called O-R Mapping. At present, there are many O-R Mapping solutions, among which open source such as Hibernate (object-relational mapping framework ). [0004] For distributed systems with massive data that require high availability and high performance, these existing implementations generally have the following deficiencies: [0005] 1. It is difficult to optimize database connections with different reading and writing characteristics and different priority levels according to the characteristics of the application. For examp...

AI Technical Summary

Problems solved by technology

[0005] 1. It is difficult to optimize database connections with different reading and writing characteristics and different priority levels according to the characteristics of the application. For example, one type of database access has exclusive database connection resources, and several written statements share one database connection; some database connections Keep it all the time, release other database connections as soon as possible, etc.

[0006] 2. SQL (Structured Query Language, Structured Query Language) statements are not centrally configured, it is difficult to perform SQL statement-level tuning, fault analysis, and centralized management of database access statements, and it is difficult for system development managers and database managers to use A centralized way to manage and tune the SQL that accesses the database

[0007] 3. Using Java's reflection mechanism (reflect) to realize O-R Mapping has a large performance consumption and cannot meet the requirements of systems with particularly high performance requirements

In this case, the mass storage function of the database system cannot be completely relied on, and the database access layer needs to provide corresponding support. The general existing implementations are not flexible enough in this regard, such as not providing a variety of distributed writing strategies

[0009] 5. Cross-statement transaction management depends on the implementation of the application server, which consumes a lot of performance

Images