Parsing Bits

[Solecord]

It's not a trick question, more of an open-ended joke really. EDI is an old method of passing business documents between companies. Here's an example format of an EDI document:

ISA*00* *00* *ZZ*PARTNER *16*102096559PROD *060521*0300*U*00400*638504569*0*P*> 
GS*PO*PARTNER*102096559PROD*20090421*0300*4569*X*004010 
ST*850*891325 
BEG*00*NE*MH80312**20090421 
CUR*SE*USD 
REF*CT*123456 
DTM*010*20090421 
N1*BY*SOLD TO Org Name*ZZ*038181483 
N1*PO*PAYOR Org Name*ZZ*038181483 
N1*BT*BILL TO Org Name*ZZ*04332571 
N1*ST*SHIP TO Org Name*ZZ*0381814830030 
N2*Partner Name 
N3*100 Bramshott Rd. 
N4*San Jose*CA*95110 
PER*IC*John Doe*TE*8015551234*FX*8015554321*EM*[email protected] 
PO1*1*11*EA*1499.95*CA*BP*12345678*MG*12070205 
DTM*010*20090421 
CTT*1 
AMT*TT*12 
SE*18*891325 
GE*1*4569 
IEA*1*638504569

The bitch with these are not all lines and segments are required. Some lines are optional and may not always appear in the document. Othertimes, only specific segments within a line are required. And some companies use different portions of the full EDI spec.

I wrote a parser via PHP to read the segments I needed, but it wasn't very extendable towards other companies. In the end I've given up and decided to use an open source EDI parser called BOTS written in Python hah

[MadGypsy]

PER*IC*John Doe*TE*8015551234*FX*8015554321*EM*[email protected]

<person>
	<name>John Doe</name>
	<phone>8015551234</telephone>
	<fax>8015554321</fax>
	<email>[email protected]</email>
</person>

or

<person name="John Doe" phone="8015551234" fax="8015554321" email="[email protected]" />

?

[MadGypsy]

actually, can you show me what that document would look like if it was properly parsed in it's entirety into XML? If you can, I will fully understand all the EDI delimiters and I can parse it. I wont do it in php though. I will do it in Actionscript. However, these languages are so similar that it could easily be ported to php, even by someone that does not know AS3.

[MadGypsy]

Oh, I am also assuming that all the capitalized stuff are your standards*. Do you have a list of these?

*(ex) TE is what everyone uses all the time to indicate a telephone number

[Zop]

I suppose you could make a map, using the root characters of each line as a key, which return sublevel maps... How did BOTS end up coding it?