The operating system user management has two separate primary key:
The operating system automatically generates its own random ID (Linux, windows of the SID), login id. Both ID must be the only difference is, delete the user test and then add a user test, SID different, login id the same. The main purpose of a dual primary key is removed to prevent the same login id to increase the confusion. For example sales manager share files to the general manager of the native peter, one year after the general manager left the company, came in an ordinary employee peter, 2 peter use the same login id, if only to make the operating system login id user management primary key, then the loopholes: General Staff peter can access only the general manager be able to see the original document. Operating its own random ID generated automatically following the user is normally invisible. Pairs of the primary key is now widely used in a variety of database systems, not limited to the user management system.
To a fixed database table to meet changing customer needs
This is mainly based on the following considerations:
Large EPR The normal use of the system, maintaining the need for software vendors and their many partners to provide technical services, including a large number of secondary development. If a user during normal use of the software need to add new tables or databases, will give the software vendors and their many partners to bring problems.
software upgrade needs. There is no one software that allows customers to use tens of hundreds of years without upgrading. Software upgrades often involve changes in the structure of the database table. Software vendors will do an additional procedure for an earlier version of the software, database data upgrade to the new version, but for user-generated tables during processing is more difficult. 4.3 software development needs. Using a fixed database table from the database development, secondary development, the much easier. For users to use the process of generating the table, each time the data must first look-up table to find names, obtain other data, compare trouble.
For example, the early UF financial software for the database using Access, to create a new database each year. Quickly, users and UF companies have found that multi-year data analysis very difficult. So this is a bad design. In the ERP, very few have different individual separate from annual data. In general, the data for all years in the same table. For multinational corporations and even the entire group is an ERP system with the same time, all the company's data together. This benefit is relatively easy to do data analysis.
|
