Logical Dependence: The Parent-Child Relationship

Logical dependence between segments is expressed in terms of the parent-child relationship. A child segment is dependent upon its parent segment. An instance of the child segment can exist only if a related instance of the parent segment exists. The parent segment has logical precedence in the relationship, and is retrieved first when the data source is accessed.

Note that if the parent-child relationship is logical and not physical, as in the case of a join, it is possible to have a child instance without a related parent instance. In this case, the child instance is not accessible through the join, although it is still accessible independently.

If a join relates the parent-child segments, the parent is known as the host segment, and the child is known as the cross-referenced segment. The fields on which the join is based (that is, the matching fields in the host and cross-referenced segments) are known respectively as the host and cross-referenced fields.

A Simple Parent-Child Relationship

In the EMPLOYEE data source, the EMPINFO and SALINFO segments are related. EMPINFO identifies an employee by ID number, while SALINFO contains the pay history of the employee. EMPINFO is the parent segment, and SALINFO is a child segment dependent upon it. It is possible to have in this relationship an employee identified by ID and name for whom no salary information has been entered (that is, the parent instance without the child instance). However, it is meaningless to have salary information for an employee if one does not know who the employee is (that is, a child instance without the parent instance).

The following diagram illustrates the concept of a parent-child relationship.

A Parent-Child Relationship With Multiple Segments

The same general parent-child relationships hold for data structures containing more than two segments. Consider the following diagram of a portion of the EMPLOYEE data source, containing the EMPINFO, SALINFO, and DEDUCT segments. DEDUCT contains payroll deduction information for each paycheck.

The following diagram illustrates the concept of a parent-child relationship with multiple segments.

EMPINFO is related to SALINFO, and in this relationship EMPINFO is the parent segment and SALINFO is the child segment. SALINFO is also related to DEDUCT. In this second relationship, SALINFO is the parent segment and DEDUCT is the child segment. Just as SALINFO is dependent upon EMPINFO, DEDUCT is dependent upon SALINFO.

Understanding a Root Segment

The segment that has logical precedence over the entire data structure, the parent of the entire structure, is called the root segment. This is because a data structure can branch like a tree, and the root segment, like the root of a tree, is the source of the structure.

In the following diagram, EMPINFO is the root. It has no parent, and all other segments in the structure are its children, directly (SALINFO) or indirectly (DEDUCT).

Understanding a Descendant Segment

Direct and indirect children of a segment are collectively known as its descendant segments. SALINFO and DEDUCT are descendants of EMPINFO. DEDUCT is also a descendant of SALINFO. A descendant segment with no children is called a leaf segment, because the branching of the data structure tree ends with the leaf. DEDUCT is a leaf.

The following diagram illustrates the concept of a descendant segment.

Understanding an Ancestral Segment

Direct and indirect parents of a segment are its ancestral segments. In the following diagram, SALINFO and EMPINFO are ancestors of DEDUCT.