Jon, Thanks for that. I had considered an approach like this (Rexx is usually my first response to such tasks), but that kind of iterative processing in Rexx, plus all the allocation/deallocation, is likely to be just too slow for the kinds of searches we need to do. I don't mind if the job runs overnight, but given the size of what we're facing even that might not be long enough.
I only got two answers to this question, including yours. Oh, well, seems as if I'm gonna have to write another utility -- that's what I was trying to avoid, but it looks inevitable. Anyway, thanks for the creative reply. David ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html