v11 and v12 do offer more up to date PDF import engines to deal with different layouts, but I doubt that there will be a magic bullet that will line everything up 100% accurately. But you can handle the data fine in 10.5 - using numeric or trapping, floating traps, and, if things get really hairy, text blocks chopped up using TextLine() and Column().
If you email me a problem PDF file I'll happily take a look.
I have packed up for the night and am going through to see the Client early in the morning, so I can't easily get at the file tonight.
The variability of the field position is quite significant, though, and it is a "code" field adjacent to a text description field so gets a bit messy to segregate.
I'm going to try another tack in the morning and see if I can get the report (which is over 200 pages long and takes 6 hours to run) produced as a Print to File, so we by-pass PDF altogether.
If that doesn't work I'll be back to you to see if you can make me a better Model!
Many thanks for taking the time to come back to me.