COBOL program to delete datasets

MoganaKumaran · New User Joined: 26 Jun 2007 Posts: 20 Location: Chennai

I'm in the process of automating a daily proces that happens , I have around 4000 datasets to be deleted, but they are with different qualifiers.
I just want to know is there any way to write a COBOL program to delete the datasets.

Regards,
Mohan

dbzTHEdinosauer · Posted: Tue Apr 15, 2008 8:15 pm

you could use dynamic allocation for a cobol program.

easier would be:

1. generate a list of datasets.
2. use this list as input to (anything - rexx, sort, cobol) to build DD statements for a step with IEFBR14, with disp=old,delete,delete.

keep in mind there are limits to the number of dd statements for a step and a job.

a cobol program with dynamic allocation might be quicker to write, but
a generated job that would have a dd statement for each ds would leave an easy audit trail to follow.

my personal opinion (no technical grounds) would be that the IEFBR14 job would run quicker.

superk · Posted: Tue Apr 15, 2008 8:19 pm

I'm confused!

How would a COBOL program delete datasets, and more importantly, why would you write a program to do this? Usually you'd use Access Method Services (IDCAMS) to delete a given list of datasets.

MoganaKumaran · New User Joined: 26 Jun 2007 Posts: 20 Location: Chennai

Hi all,

1) i have 4000 DSN s copied in a dataset.
2) I jus want to know is there any utility or program or code to delete the datasets(4000) that are located in a dataset.

could u assist me which process would b simple , JCL iefbr14 r IDCAMS or is there any COBOL prog.

Thanks,
Mohan

Craq Giegerich · Posted: Tue Apr 15, 2008 8:36 pm

What is the format of the dataset that contains the list of datasets. If you can format it properly then you can use IDCAMS to do the deletes.

mmwife · Super Moderator Joined: 30 May 2003 Posts: 1592

Is this intended as a production job? Test? Where does the file get updated? Prod? Test?

Some solutions are OK in test; not OK in prod. Please provide ALL details for the best answer.

MoganaKumaran · New User Joined: 26 Jun 2007 Posts: 20 Location: Chennai

It s all gonna be a test datasets , that we have used during our testing process.It s of no use absolutely in future. So we want to delete it , since the count is very high around 4000 datasets , i need a jcl code or cobol code to delete the all 4000 datasets at a single stretch , will it be possible.

Kindly help me in this regard.

thanks in advance,

regards,
Mohan

ofer71 · Posted: Wed Apr 16, 2008 10:52 am

I would go for REXX.

O.

enrico-sorichetti · Posted: Wed Apr 16, 2008 11:01 am

if all the datasets with the same high level qualifier can be deletete

idcams delete hlq1.hlq2.hlq3.*
wiil do the job
the star is a generic placeholder for just one qualifier

another solution could be to run a dfdss job wit a dummy output

(the filtering capabilities of dfdss are quite sophisticated)
check the manual

MoganaKumaran · New User Joined: 26 Jun 2007 Posts: 20 Location: Chennai

In my requirement I have hlq1.hlq2.****.*** here hlq1 and hlq2 alone is common in all my datasets, the 3rd and 4th qualifier differs in all datasets.

Can you please gimme the sample JCL to get through this process.

Thanks ,
Mohan

enrico-sorichetti · Posted: Wed Apr 16, 2008 11:25 am

here is a link to the filterin specs for dfdss
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/dgt2u261/3.2.1?ACTION=MATCHES&REQUEST=filtering&TYPE=FUZZY&SHELF=DGT2BK70&DT=20080128192112&CASE=&searchTopic=TOPIC&searchText=TEXT&searchIndex=INDEX&rank=RANK&ScrollTOP=FIRSTHIT#FIRSTHIT

here is the link to the dfdss reference
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/DGT2S261/CCONTENTS?SHELF=DGT2BK70&DN=SC26-7402-09&DT=20080128160125

sample jcl

MoganaKumaran · New User Joined: 26 Jun 2007 Posts: 20 Location: Chennai

Thanks enrico!!

Lemme try this , alsoI want to know is there any other simple way of doing it .

Regards,
Mohan

enrico-sorichetti · Posted: Wed Apr 16, 2008 11:40 am

No ( or at least I have no reason/time to investigate other solutions )
isn' t the dfdss job simple enough

the filtering capabilities of idcams delete are not sophisticated as dfdss' s
(with dfdss You can filter aon substrings of a qualifier )

furthermore with dfdss has a very powerful approach to check for results

MoganaKumaran · New User Joined: 26 Jun 2007 Posts: 20 Location: Chennai

Thanks enrico.

I can go with this ...

Thannk u very much...

mmwife · Super Moderator Joined: 30 May 2003 Posts: 1592

Probably too late, but here's another:
3.4 hlq1.hlq2.*
in cmd line: SAVE list4k
This creates yrTSOid.list4y.datasets

Using EDIT delete everything except the DSN; insert " DELETE " at pos 1 of each line. e.g.:

C ALL 46 140 P'¬' '';res chg;C ALL 1 P'=' 'bDELETEb' (note: b = space)

save the results

Use following JCL: