|
In this section: |
A group of segments that are related to each other as a sequence of parent-child relationships, beginning with the root segment and continuing down to a leaf, is called a path. Because the path is a sequence of parent-child relationships, each segment is logically dependent upon all of the segments higher in the path.
In the following view of the EMPLOYEE data source, EMPINFO, SALINFO, and DEDUCT form a path. An instance of DEDUCT (paycheck deductions) can exist only if a related instance of SALINFO (the paycheck) exists, and the instance of SALINFO can exist only if a related instance of EMPINFO (the employee) exists.
Consider the full EMPLOYEE structure, which includes the EMPLOYEE data source and the JOBFILE and EDUCFILE data sources that have been joined to it.
This is a multipath data structure. There are several paths, each beginning with the root segment and ending with a leaf. Every leaf segment is the end of a separate path. The following diagram illustrates the concept of a multipath data structure.
The EMPLOYEE data structure has six paths. The paths begin with the EMPINFO segment (the root), and end with:
Each path is logically independent of the others. For example, an instance of DEDUCT is dependent upon its ancestor segment instances SALINFO and EMPINFO, but the ADDRESS segment lies in a different path, so DEDUCT is independent of ADDRESS.
This is because employee deductions are identified by the paycheck from which they came, so deduction information can be entered into the data source only if the paycheck from which the deduction was made is entered first. However, deductions are not identified by the employee address. An employee paycheck deduction can be entered without the address being known, and conversely, the employee address can be entered before any paychecks and deductions have been entered into the data source.