The lack of a decent shell in Windows is a killer.   appropos newline would have shown you "unix2dos".  

tr couild have done it.   For a custom solution, I'd almost certainly turn to Perl before going back to sed. 

The text file customs of both OSs are well documented so should not have been tough to research.

Stranger still is the idea that having to write filters in data processing would freak anyone out.   That seems to me to be a huge part of what we do - take data however it is provided and turn it into something we can handle.

I don't understand the fellow who had a problem with the listing order of a data dictionary provided by an outside firm - maybe the mainframe weenies understand that "ordinal position" is a lame sort of concept to begin with.   None of us wants to do the drudge work, but its what we get paid for.  One thing I will say for the mainframe weenies (never was one but some of my best friends...) - they at least know basic database concepts, unlike 84.3% of the SQL Server jockies I've met.