Rickerism

SOAP Representation

WS-EDI Level 4 provides a SOAP representation of EDI. With Level 4, there is only one message in each SOAP envelope. If the original EDI file contained several messages, then the Level 4 representation will create several SOAP messages, one for each EDI message in the original file. When EDI files are split apart, the interchange and functional group headers are repeated in the SOAP headers.

X12

In Level 4, there is only one ST loop in each SOAP envelope. The SOAP body contains an element which contains the segments within the ST loop except the ST and SE segments themselves. The SOAP header contains an element which contains only the ISA, GS and ST segments. If the original EDI message contained several ST loops, then the same ISA and GS segments are repeated in the SOAP header of each SOAP envelope.

<s:Envelope>
  <s:Header>
    <w:wsedi>
      <w:segment code="ISA" name="Interchange Header"/>
      <w:segment code="GS" name="Functional Group Header"/>
      <w:segment code="ST" name="Message Header"/>
    </w:wsedi>
  </s:Header>
  <s:Body>
    <w:message>
      ... segments of the message
      (after the ST segment and before SE) ...
    </w:message>
  <s:Body>
</s:Envelope>

Dropped segments

Why do we drop the SE, GE and IEA segments? Afterall, WS-EDI strives to keep a one-to-one relationship between EDI and XML. Within any level of representation, WS-EDI must be able to recreate the EDI document. It is far easier to ignore information left in the document than it is to recreate information stripped out the document. However, the SE, GE and IEA segments are only used as check sums in EDI. To place these segments in the SOAP header would require the processor to buffer the entire EDI document before producing the WS-EDI document. Experience has proven that buffering very large EDI documents is unfeasable. It is far easier for an Uneveloper to recreate the SE, GE and IEA segments than it is for an Enveloper to place those segments in the SOAP header.

EDIFACT

UN EDIFACT messages are represented in SOAP the same way as X12. The only difference is the segments that appear in the SOAP header.

<s:Envelope>
  <s:Header>
    <w:wsedi>
      <w:segment code="UNA" name="Service String Advice"/>
      <w:segment code="UNB" name="Interchange Header"/>
      <w:segment code="UNG" name="Functional Group Header"/>
      <w:segment code="UNH" name="Message Header"/>
    </w:wsedi>
  </s:Header>
  <s:Body>
    <w:message>
      ... segments of the message
      (after the UNH segment and before UNT) ...
    </w:message>
  <s:Body>
</s:Envelope>

The Level 4 representation drops the UNT, UNE and UNZ segments from EDIFACT for the same reason segments are dropped in X12.

Combinations of levels

What happened to Level 3?

Level 3 is the Enriched and Nested Representation, that is, it is a combination of Level 1 and Level 2. The levels comply with binary flagging. The complete set of levels are as follows:

Flags	Level	Description
0000	0	Minimal
0001	1	Enriched
0010	2	Nested
0011	3	Nested and enriched
0100	4	Enveloped (SOAP)
0101	5	Enveloped and enriched
0110	6	Enveloped and nested
0111	7	Enveloped, nested and enriched

Note that the Level 4 representation is not enriched or nested. That is, the WS-EDI tags within a Level 4 SOAP document do not contain the human readable metadata or the loop structures inside the message. A Level 7 representation would be a SOAP message with the descriptions and loop structures.

Example documents

• An UN EDIFACT document and its Level 4 representation.
• An ANSI X12 document and its Level 4 representation.