Electronic Data Interchange
The Data Segment (UN/EDIFACT)

The data segment is the fundamental unit of information in an EDI document.  A block of information that is layered one upon another separated only by a terminator to make up the document.  Basically, the document is just composed of layers of data segments.  The data segment consists of related simple data elements in a predefined sequence; or a combination of related simple data elements and composite elements in a predefined sequence.  Each elements and their relative position in the collection describe the type of information the data segment can hold to serve a special purpose in the context of the EDI document as a whole.  Each data segment has a unique purpose and properties from other data segments.

The data segment has the following general properties:

• Identifier .  The identifier of the data segment, or the segment tag.  Each data segment has a unique identifier.  For UN/EDIFACT, the character "U" begins the identifer of all service segments.
• Name.  The name of the data segment.

• Position Number.  The position number of the data segment.  Each data segment is assigned a position number indicating their relative position to other segments, and in the sequential order in which they are to appear.  A data segment with a higher position number cannot appear before a data segment with a lower position number, and vice versa. 
• Requirement.  The standard requirement of the data segment's presence at the specified position.  At a minimum, the segment identifier alone indicates its presence in the data stream.
  
  • Mandatory (M).  Must be present at the specified position.
  • Conditional (C).  May or may not be present at the specified position.
    
• User Requirement.  The custom requirement of a data segment's presence at the specified position as implemented by a trading partner.  This requirement supersedes the standard Requirement, that is, if the standard requirement designates that the data segment is optional (O,C) but the User Requirement indicates that it must be used (MUST USE) then the presence of the data segment is required.  At a minimum, the segment identifier alone indicates its presence in the data stream. The following is a list of allowed user requirements:
  • MUST USE
  • USE
  • NOT USED
  • DEPENDENT
  • RECOMMENDED
  • NOT RECOMMENDED
• Maximum Usage.  The maximum number of times the single data segment can occur in the same position.  In some implementations, there is no limit to the number of times it can occur; an infinite occurrence is described as having "more than one".  In a loop, the first data segment in the segment group can only have a maximum occurrence of one.
• Loop Repeat.  The maximum number of times a group of adjacent data segments can repeat as a whole.  A single occurrence of a repeating group of data segments is also referred to as a loop.  This property is only meaningful on the first data segment of the segment group, which is also known as the loop trigger.
• Loop Level.  A number that indicates how deeply the data segment is nested if it belongs to a loop.  Data segments not belonging to a loop has a loop level of zero.
• Area.  The area (table) that the data segment is in.  The ASC/X12 transaction set or UN/EDIFACT message is commonly broken up into three areas:
  • 1 - The header.
  • 2 - The detail
  • 3 - The summary.
• Ordinal.  A unique number assigned to each data segment.  The number does not necessarily follow any particular sequence.
• Reference.  A reference position to the data segment.  This is composed of the identifier and the position number of the data segment.
• Loop Tag.  The identifier of the loop that the data segment belongs to.  This is normally the identifier of the first data segment in the group of data segments that make up the loop.

Each data segment in the document is separated by a segment terminator in the document.

Composite Elements and Data Elements

A data segment can have a combination of composite elements and simple data elements.  Or they can have composite elements only, or simple data elements only.  Each element is separated by an element separator character in the data stream.

Both simple data elements and composite elements can have repeating instances.  A repeating instance of an element occurs repeatedly at the same position in the data segment, each separated by a repeating separator character.  However, it is important to note that two elements with the same identifier can be adjacent to each other and not be deemed repeating elements.  In the data stream, they can be differentiated by the fact that:

• repeating elements are separated by a repeat separator character;
• adjacent non-repeating elements are separated by an element separator character.

All elements in the collection can have a mandatory (M), or optional (O,C) requirement.  Additionally, an optional element may have a relational, or dependency requirement where its required presence is dependent on the presence of other elements in the same collection.  If this relational condition exists, then a syntax rule, or dependency note, that states the relationship of the elements, is included in the data segment definition

It is possible to have all elements in the collection to have an optional requirement.  If there is no data in the elements, the data segment is not considered empty as long as its identifier stands alone in the data stream. A simple data element is considered present only if its value contain at least one character; and a composite element is considered present only if one of its component element value contain at least one character.  However, even though the identifier alone is enough to indicate its presence, the stand alone identifier is present only to satisfy the mandatory requirement of the data segment -- it should be completely ommitted if it has an optional requirement.

It is also possible to have all elements in the data segment to have a mandatory requirement.  If some elements are mandatory, these elements will normally have a position at the beginning of the collection.  If an optional data segment is present, then all mandatory elements must be present as well.  If all optional and/or mandatory elements are empty in a mandatory data segment, the data segment should be omitted completely.  However, it is an error at the mandatory element if its parent data segment is omitted (not present) and is mandatory; in this case, there are two errors: the missing data element, and the missing data segment -- both mandatory.

An empty simple data element contain no character in its value; an empty composite element contain only empty component elements.  For both structures, they are completely omitted in the data stream, including their separators that follow them if they occur at the end of their parent structure.  If the empty element is followed by a non-empty element then its absence is indicated by its separator that follow it, thus preserving the positions of non-empty values in its parent structure's data stream.