Wow! What an awesome group and resource. Five answers in twenty minutes. Responding more or less to everyone
- I realized after I hit send what I had implied but failed to state: the file is "binary" -- that is, many of the characters are not printable -- basic "mainframey" data. I suspect that editing it under Windows would be problematic. - I have both Rexx and assembler skills. Yeah, I was thinking that if there were not another approach I might write a little Rexx program. - The sed approach looks interesting. I have very basic UNIX skills so it would be a little bit of a challenge. - What I like best I think is Dave Salt's approach. The file is only 36+K. I suspect it is a couple of hundred records at most. I could pretty easily do a FIND on the hex "delimiter", do an eyeball check on whether it was a real record boundary or a false positive (I can recognize the start of a real record), and then either hit REFIND or a PF key that invoked a macro that split the record at that point. Easier to debug than the Rexx program, and no false splits to repair. I will still have to re-join the records that currently break erroneously at the FB boundaries. Whoever it was, thanks for the TF suggestion -- that should work. I confess I have never written an ISPF edit macro. (I wrote a bunch for XEDIT back when dinosaurs roamed the earth.) Dave, if you wanted to consider this "contacting you" you could send me whatever clues you wanted. I think you can make out my e-mail but if not it is charlesm at mcn dot org. I think I just need a PF-key-invocable macro that would split a record at the cursor position, putting the character under the cursor into the latter record. Hmmm, it's currently FB. I will have to first copy it into a VB dataset so that short records stay that way. Thanks all! Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Paul Gilmartin Sent: Tuesday, January 15, 2013 9:25 AM To: [email protected] Subject: Re: Break a dataset into new record boundaries? On Tue, 15 Jan 2013 11:21:37 -0600, Roberts, John J wrote: > >If binary, I would just write a one-off ASM program to recover the original >records. It's probably a 30 minute task, easier than trying to learn anything >new. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
