Oho! Now that is slick, in a gruesome way. I like it; just might use that. Bonus for doing exactly what I want in an almost unrelated way. It even seems like that could be the basis of a script that could figure out for itself what the probable pattern for each lousy file is, and just yank any outliers... Let's see how big a mess I can make with THAT!

And just the same, there is still that "D D D D D D" -- a non-indeterminate sequence so you just have to hope you don't run into any lines with seven Ds. I bet there's a way around that (and I know I'll never have more than nine -- in this file...) but that does point back to my original question: Can you make a single regex carefully validate a variable number of fields (and return all matches)? Will perl regex do that, or does it exceed the possibilities?

Anyway, thanks!