On Thu, 7 Aug 2008 23:05:24 -0500, Joel C. Ewing wrote: >Ken Leidner wrote: >> I have a tape dataset cataloged over 10 years ago with a very strange - >> invalid name. I am not sure how it was created, but I would like to >> remove it from the catalog. The tape has scratched years ago and I >> found this odd dataset name and would like to remove it from the catalog >> without having to zap the catalog to change its name first. >> >> Any suggestions? >> >> I have tried; >> >> In TSO a DELETE 'P.FSPOL.SBHF38AA(0)' noscratch >> IKJ56709I INVALID DATA SET NAME, 'P.FSPOL.SBHF38AA(0)' >> >> ISMF doesn't even find the dataset, even when looking for a prefix of >> the dataset name (like 'P.FSPOL.SB*.**) >> >> Running IEFBR14 in batch gives >> //STEP0 EXEC PGM=IEFBR14 >> //DD1 DD DSN='P.FSPOL.SBHF38AA(0)',DISP=(OLD,UNCATLG) >> >> IEF212I CJ4130BT STEP0 DD1 - DATA SET NOT FOUND > It was discussed here within the past couple years that IBM has done two or three about-faces on catalog syntax rules.
In the oldest implementation, such a name could never have been catalogued. At some relatively recent release, IBM relaxed the constraints so practally any 44-character string could be catalogued, although I suspect most application-level software would have intervened to enforce the classic syntax rules -- a new-style name might not be catalogable except by an assembler call to catalog services. Some customers found this newly-granted liberty psychologically disturbing (Stockholm Syndrome?) and complained to the extent that IBM provided an option to reinstate the classic syntax rules. At first, the default was lenient; in a later release the default was changed to strict. Your offending data set may have been catalogued (by an assembler program) during the interval of freedom, or the catalog may have been zapped. So, you have at least two options: re-zap the catalog, or temporarily set the option to lenient syntax (PARMLIB? restart required?) and uncatalog the data set. If you then choose to reinstitute strict checking, you can only hope that during the interim no other users have catalogued nonconforming data set names or you will once again be in a similar situation. IBM made a poor design choice in the manner of implementing the strict checking option: the existence of an object should always overrule the invalidity of its name -- if it's there, don't quibble over grammar. (I'm reminded of a Woody Allen movie in which a teller disputes a bank robber over a misspelling in the holdup note.) At the very least, access to objects with unconventional names should be allowed if only for the purpose of deleting or renaming them. -- gil ---------------------------------------------------------------------- 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
