info.rst

Ideas for path syntax for referencing elements or other parts of msg

Specification

Main specification documents

• http://www.unece.org/trade/untdid/texts/d423.htm
• http://www.unece.org/trade/untdid/download/r1241.txt

Spec archive: http://www.unece.org/fileadmin/DAM/trade/untdid/d14b/d14b.zip

Information about files contained in spec archive: .edify/downloads/d14b/_TextFiles/CONTENT.TXT

Reference message

(Indentation for demonstration only)

UNA:+.? ' UNB+UNOC:3+Senderkennung+Empfaengerkennung+060620:0931+1++1234567'

UNH+1+ORDERS:D:96A:UN'

BGM+220+B10001' DTM+4:20060620:102' NAD+BY+++Bestellername+Strasse+Stadt++23436+xx' LIN+1++Produkt Schrauben:SA' QTY+1:1000' UNS+S' CNT+2:1'

UNT+9+1'

UNZ+1+1234567'

Query language

Querying using spec information

msg:ORDERS[0]|seg:BGM[comp:C104:0]|[simp:1004]

--> "B10001"

msg:ORDERS[0]|seg:BGM[comp:C104:0]

--> CompositeDataElement (with component elements)

msg:ORDERS[0]

-> Message object

msg:ORDERS[*]

-> [Message object] (potentially many)

TODO: addressing segment groups

• Precompute paths? - only a small subset of data elements may ever get requested - message layout might change (e.g. by adding groups) --> better: on-the fly
• only fully qualified (abs.) paths; selection only by data element amost never makes sense

Web application examples

e.g. http://www.truugo.com/edifact/d14b/

Validation & Navigation

Message segment sequence validation

Basic operations while matching message sequence:

• match segment -> XX, increment current segment spec index
• repeat segment -> increment repeat count
• skip segment -> increment current segment index
• descend into group -> recursive call findSeg
• finish group (when: end of msg or segment not found in current group) -> return from findSeg
• repeat group -> loop within findSeg
• skip group -> increment current segment spec index

• (msg end ? -> spec complete (no more mandatory segment/group) -> ok

  incomplete: -> error missing_mandatory_segment