Interestingly, a recommendation form IBM (under "DFSMSdfp: Accommodate changes 
in LISTCAT LEVEL output"):

"Take steps to convert any programs that rely upon LISTCAT output to use the 
Catalog Search Interface (CSI) instead. CSI is a supported general-use 
programming interface for the catalog, and will remove any dependency you have 
on future possible changes that may occur to LISTCAT output."

Do IBM provide COBOL layouts for the mapping of the output from CSI? Assembler? 
Anything else? My on-going concern with COBOL accessing "system" stuff for 
Production use is that IBM don't provide COBOL layouts (control blocks, for 
instance) and you have to rely on notification of changes to the Assembler 
macros as indicators that you have to change the COBOL layouts, and usually it 
is not the same people dealing with the OS upgrade as those who have hand-coded 
(or copy/pasted) full layouts or layouts which rely on specific positions for 
specicif items. Then, BANG! at best, wrong output at worst.

It is unclear to me if a use of CSI can guarantee to obtain all the GDSs 
associated with the GDG in one shot. If done in two (or more) shots there's a 
chance of a "change of state" since the first access of the CSI (new, delete, 
rename).

Anyway, I'd better state that using LISTCAT has issues, IBM do not hold 
themselves to keeping the output the same.

I'd not attempt the rename in the COBOL program. What would you do if it failed 
(data set in use, data set exists)?  Output a file for input to IDCAMS for the 
rename, and allow for identification of non-zero results.

You'll need a report of data sets processed.

Your generation-range processing is going to need to know about what to do with 
a range which includes G9999V00, and you mustn't miss out in the magical 
situation that an action Version is ever used.

What about a re-run situation?

Is there a need to worry about the number of steps? With LISTCAT busted as an 
idea, I'd consider one COBOL (or other) program to use the CSI and produce a 
definitive file of data set names from the GDG. 

Then a COBOL program to do the work.

IDCAMS to do the renames on successful completion of the work-program.

Keep everything as simple an controllable as possible. I'm assuming it needs to 
be in COBOL for the "skills shortage" issue with it being in Assmbler? So keep 
everything as simple as possible. Decide what needs to be processed, step one. 
Process, step two. Post-process (the renames) step three. Step two and three 
running of the output of step one, with the obvious dependencies on successful 
completion.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN

Reply via email to