>>>My goal in to compare sold tickets from Amazon with BSP billing reports. Every ticket in billing report
>>>must have corresponding invoice issued to customer by travel agency.
>>
>>I looked at the document on the link you provided - that only seems to provide information on the added EMD segment not the complete interchange ?
>
>Yes. It provides main document name
>
>Ticketing - Amadeus Edifact Conventions and Directory PSEUDO EDIFACT AIR
>- VERSION 207 (Doc ID - ICDTIRAIR207
>
>I havent found this document after heavy googling.
>This format start AIR-BLK207
>
>In my case older version of this format, which starts with AIR-BLK206
>is used. I havent found its description and havent got response from Amadeus local
>company.
>
>>Was the sample that you initially posted complete ?
>
>Yes. Yuo see similar samples in the documents which I posted and which you posted.
>
>> A full EDI interchange would have a UNA segment defining the separators used in the interchange.
>
>This is probably pseudo edifact.
>Amadeous client conf manual which I found in internet describes that also second format ,
>Edifact is avaliable which is used by SAP onlu. Output format can specified in
>Amadeus client settings.
>This format is also not documented in internet.
>
>>In the sample this appeared to be using just CR or CR/LF as the segment
> separator and semi-colons for the component with no separator for data
>elements - the latter, where they contain multiple elements, look to be just fixed length fields.
>
>yes, Amadeus file contains lines separated by crlf.
>
>>If '2JW8HW' is the number then I'd guess the first occurrence is the main number. The repeat at the end is prefixed by 'OV' for Estonia Air followed by a second number (RZ3ZZ7) for the Turkish Airways leg?
>
>PNR locator must be unique. In pdf file samples this is always repeated at start and at end.
>Maybe turkish airlines uses other reservation system and this is its id.
>
>>I looked at : http://contentm.mkt3061.com/lp/10362/159354/ATC%20User%20Manual%20EN%20Feb0912%20v2_FINAL_1.pdf
>>The section near the end ('Back-Office') seems to bear out my assumption - this ticked is a round trip Paris - Frankfurt (all on Air France) and the ticket number (5PII3K) is only repeated once at the end (i.e. no other carrier involved). Also noted that the G-X segment is 'PARPAR' rather than 'CDGCDG' so seems to be based on the city rather than the airport.....
>>Can't fathom how the legs are listed. On your sample the second leg is listed first. On this one it's the other way around (and the prefixes are different) :-{
>
>5PII3K is reservation number in Amadeus. In travel industry terms it is called
>PNR record locator and it is described in Wikipedia article which I provided.
>
>One reservation can contain may tickets. Tickets are issued for every person
>and for every flight in reservtation.
>Ticket numbers are longer and contain only numbers, for example 3919928548 ,
>1810426516
>
>From your sample and sample in http://www.amadeus.com/corp/xml/documents/aco/gb/en/EMDAIROct2011.pdf
> it looks like ticket number is probably in line starting at T-K
>
>T-K057-2996031611
>
>057- is unknown element.
>After that there is ticket number: 2996031611

TBH, whilst trying to analyse and deduce is fun, unless you can obtain the full specs document mentioned above I don't see how you can offer a viable, robust solution to a client. I'm offline for a week but would be interested to hear if you do obtain the docs......