>>>If everything is regular (order of fields do not change etc) it'd be easy:) However there are other messages saying that this is an HL7 format (which I don't know the details) the situation might be much more complex than it looks.
>>
>>Well, info about the line structure was sparse. But as long as each line maps to a special cursor type (perhaps 2 or 5 different "CUR_EVH" depending on fields 16, 23 ad 35 for instance I'ld still go for some cursor naming handling.
>>
>
>If every line (segment) truly maps to a completely different table or cursor, that is truly completely different animal than EDI...where information that normally would be stored in one table may span multiple segments.
>
>Would appreciate if someone would shed some light on this, just for curiousity's sake.

The fields (segments) can map to whatever field in the data tables. They are in fixed position in the HL7 data (text) stream. Meaning, I may not always have 7 rows, but when I see "PDI" at the beginning of a line, I know that there will be x number of fields to follow. Many will be empty, but the delimiters will still be there. It may be more complex than that - haven't read Greg's reply yet - but for the billing piece I'm working on at the moment, I think that's pretty much it.