I support a customer where the SMF DAILY and WEEKLY files are GDG-based and we use a relative Julian date (or week of year) to identify the particular file of interest. On days where no processing occurs, it's conceivable to have a stub generation created. A new year requires creating a dummy or "starter/adjustment" generation until that first daily or weekly job completes, then delete the stub, if required.
Possibly, a similar technique could be used if the generation # represented a "**.GmmddV00" naming pattern, though it would be necessary to "insert" a dummy generation "**.G0200V00" and so on, then maybe go back and delete the stub generation. Obviously "daily" generations for different years (using the "mmdd" technique) would not be able to span years. So, as I mentioned, this technique works well with Julian dates for a daily file and also a week of year (weekly being named "**.GyywwV00"), though some admin is required at the start of a new year. Suggestions for using "derived" JCL SET symbolics using some standardized naming convention have been mentioned previously on this forum and I would consider that the same could apply in this application but clearly with some research, test scenarios and well-documented restart/recovery procedures. Sincerely, Scott Barry SBBWorks, Inc. ________________________ From: Tim Hare <[EMAIL PROTECTED]> Subject: GDG question I'm pretty sure I can do this, but I'm looking for a second opinion (and "you're ugly, too" as my coworker said doesn't count <grin>). I have the assistance of a job scheduler which can substitute variables in JCL to help with this idea, so bear that in mind. I want to create twelve GDGs, prefix.whatever.JAN, prefix.whatever.FEB and so forth. Then, instead of creating prefix.whatever.JAN(+1) I want to create prefix.whatever.JAN.G(day-of-month-with-leading-zeroes)V00 with DISP=(NEW,CATLG). In essence one generation per day of month, with missing generations if we don't run that day (state holiday or whatever). Then I want to be able to read prefix.whatever.JAN as the complete set, as you can do with GDGs. Are there any holes in this theory that I'm overlooking? Tim Hare Senior Systems Programmer Florida Department of Transportation (850) 414-4209 ---------------------------------------------------------------------- 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
