That is true for any dataset. If you "create" (allocate space to and write an entry in the VTOC) a dataset but never open it for output, a subsequent step, or even a subsequent job days later, can open that dataset and read it, getting whatever data happens to be on those tracks, with no error indication, unless the reading program happens to get an S013(?) because the blocks are the wrong size, not valid RECFM=V format, etc. Results are unpredictable: one day you will get an ABEND, the next day you will read in unexpected data, and on the third try you will be lucky enough to hit an EOF marker (or empty tracks followed by an EOF marker) right away.
An interesting mainframe "hack" would be to write a program that allocated vast quantities of DASD DISP=NEW and then read whatever happened to be there (RECFM=U, of course), looking for "interesting" data. Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Veilleux, Jon L Sent: Thursday, July 06, 2006 11:47 AM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: avgrec/avgblk history ? At one time, and I don't know if it still works this way, if you allocated a temporary dataset with disp=mod,pass and didn't open it in the first step, when you opened it in the second step with disp=mod, it would read the residual data from the work disk at the point where the file was allocated. We ran into this in an ASML proc where the LINK step ---------------------------------------------------------------------- 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
