repetition separator question

[hancockks]

I have an EDIFACT document that isn't using repetitions, but it's of course present in the UNA segment. When writing out the file, I was trying to write back out the UNA segment, using the reader.getDelimiters() but the map was empty for the repetition character.

I came across issue #77 which states that the repetition sepaprator is onlly available in syntax version 4. When I changed the document to syntax version 4, I now see the repetition character in the getDelimiters() map.

I was unable to find in any EDIFACT docs why this was restricited to version 4. It seemed to be present in version 3 as well?

[MikeEdgar]

@hancockks , much of the documentation I used to support StAEDI is from the Joint Syntax Working Group (JSWG) site [1]. It seems that recently it's be reorganized and some of the documentation is either no longer available (for free) or it's been relocated.

This page [2] has information about syntax versions 1-3, but unfortunately it looks like the V3 PDF was incorrectly updated and it's only an amendment. If you look at the V2 PDF on page 14, it identifies the structure of the UNA, which is the same for V3.

Repr.	Name	Remarks
an1 M	COMPONENT DATA ELEMENT SEPARATOR
an1 M	DATA ELEMENT SEPARATOR
an1 M	DECIMAL NOTATION	Comma or full stop
an1 M	RELEASE INDICATOR	If not used, insert space character
an1 M	Reserved for future use	Insert space character
an1 M	SEGMENT TERMINATOR

Compare that to the V4 syntax here [3] (the first PDF, page 23)

POS	REP	S	Name	Remarks
010	an1	M	COMPONENT DATA ELEMENT SEPARATOR
020	an1	M	DATA ELEMENT SEPARATOR
030	an1	M	DECIMAL MARK	The character transferred in this position shall be ignored by the recipient. Retained to maintain upward compatibility with earlier versions of the syntax.
040	an1	M	RELEASE CHARACTER
050	an1	M	REPETITION SEPARATOR
060	an1	M	SEGMENT TERMINATOR

Specifically, the "reserved for future use" in the earlier versions was replaced with the repetition separator in V4, and the decimal mark is decommissioned and shall be ignored.

Let me know if this clears things up. If there is some better source of documentation that is not behind a pay wall I would be more than happy to take it into consideration.

[1] https://service.gefeg.com/jwg1/Current/id1.htm
[2] https://service.gefeg.com/jwg1/Current/d31.htm
[3] https://service.gefeg.com/jwg1/Current/dx1.htm

[hancockks]

That's exactly the information I was looking for. Thanks so much for taking the time to dig that up and explain in detail!

[hancockks]

Maybe this now belongs in issues, but I tried adding the following lines to a ORDRSP EDIFACT doc which I believe are compliant:
CTA+PD+Contact Person'COM+foo@example.com:EM*bar@example.com:EM'

When I use createEDIStreamReader() and parse the body, reader.hasNext() never hits START_LOOP and the segment gets parsed as multiple elements. Is that expected?

[MikeEdgar]

I am not sure about the loop without seeing more, but you should see each repetition of the element as a separate ELEMENT_DATA event. You can check which occurrence it is by fetching reader.getLocation().getElementOccurrence().

[MikeEdgar]

The loop question would be a good issue if you can provide a small reproducer file/schema without any sensitive data.