Method for changing data sheet structure

Abstract

The invention provides a method for changing a data sheet structure. By the method, short and controllable execution time of changing the data sheet structure is achieved. The method includes: corresponding to one or more primary structure data sheets in running state, each node in the computer network saves one or more new structural data sheets; with chosen nodes as novel publication nodes and the rest as novel subscription nodes, subjecting novel publication sheets and novel subscription sheets to the novel structural data sheets of the novel publication nodes and the novel subscription nodes respectively, adding data of conventional structural data sheets of the novel publication nodes to the corresponding novel structural data sheets, and publishing the data to the novel subscription nodes; in the state that the novel structural data sheets of the nodes and the conventional structural data sheets are synchronous, stopping data writing of the conventional structural data sheets of the nodes, and subjecting the novel structural data sheets of the nodes to the data sheets in the running status.

Description

technical field [0001] The invention relates to a method for changing the structure of a data table. Background technique [0002] With the popularity of computer network applications, the capacity of the database is also increasing, and the work of changing the structure of the data table is becoming more and more time-consuming. There are generally two common ways to change the data table structure: [0003] 1. Operate directly on the data table and change the definition of the table structure. There may be direct impacts, such as the following reasons 1 and 2. [0004] 2. Create a new structure table, import data in the maintenance window, and then modify the table name; the possible impact is that the data import time is limited by the size of the original table. [0005] The maintenance window is the period during which services are suspended. In order to minimize the impact on system operation, the shorter the maintenance window, the better. In the face of a data ta...

AI Technical Summary

Problems solved by technology

Because when the length of the fixed-length field in the data table is extended (including but not limited to changing the int type to bigint, char(10) to char(20), etc.), in order to ensure the ACID principle of the relational database, the data table will be added Schema update lock (Schema-M) to prevent new data from being written when the structure is changed; for large-capacity tables, the execution time of the structure change (hereinafter referred to as the execution time) will become longer, so that it exceeds The expected maintenance window time directly affects business visits;

[0007] 2. The risk of execution and rollback time increases

As the amount of data increases, the structure change time of the large-capacity table will increase faster than the increase in the amount of data. Once the maintenance window time is exceeded, whether to continue to wait or terminate the change and enter the rollback state will make it difficult for the maintenance personnel to make a decision. The rollback time may exceed several times or dozens of times the execution time

This greatly increases the risk of operational uncertainty; in addition, due to the complexity of the production environment, it is difficult to test in advance in a test environment that completely simulates the production environment, which reduces the applicability and reliability of test results, which also increases the impact on implementation. Uncontrollable risk of time

Images