Compare 2 PDS member & store to same name mem of new PDS

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

hi, Please help on this. As far as searched, when 2 PDS members are compared, the output results are stored into one PS file. But I want it to store into new PDS with same member whichever compared. Example is given below.

MY.COMPARE.PDS1
MEM1
MEM2
MEM3
MEM4

MY.COMPARE.PDS2
MEM1
MEM2
MEM3
MEM5

The mismatch results to be stored into new PDS and with same name of members whichever compared and there is a mismatch.

MY.COMPARE.PDS.NEW (assuming MEM3 don't have any mismatch)
MEM1
MEM2

Note: Compare input should be PDS1 & PDS2 names only. Not with specific member names.

Bill Woodger · Posted: Tue Jul 26, 2011 11:41 pm

I think you'll have to explain some more.

What software are you using for your compare?

Are you comparing member names only, or contents of members as well?

I do not understand your output data. Why 1 & 2 but not 3?

superk · Posted: Tue Jul 26, 2011 11:58 pm

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

superk · Posted: Wed Jul 27, 2011 12:02 am

That seems like the right process. What's the problem?

enrico-sorichetti · Posted: Wed Jul 27, 2011 12:08 am

that most probably the TS wants to split the isrsupc result dataset to make the overall analysis <simpler>

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

I need the solution thru JCL or thru some sort utility. But this topic is moved to "SMS & VSAM" from "JCL". Is this right forum to discuss about this?

superk · Posted: Wed Jul 27, 2011 12:31 am

Yes, it is. This topic is dealing with datasets, in particular with the nuances of partitioned datasets, which almost always require specific programming functions to deal with them properly. If you just state that you need a batch-based solution here, that's fine.

Akatsukami · Posted: Wed Jul 27, 2011 12:32 am

superk · Posted: Wed Jul 27, 2011 12:35 am

What is the target RECFM and LRECL for the new PDS? I'm wondering if you can possibly use IEBUPDTE for what you want to do?

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

enrico-sorichetti · Posted: Wed Jul 27, 2011 12:50 am

the layout of the isrsupc output is not so easy to parse,
and doing it trhu <sort> will require at least two passes over it

if I had to do it ( but only with a gun pointed at my head ) I would do it with rexx

if You google for superc rexx You will find some tools to help You in <smart> browsing the superc output

superk · Posted: Wed Jul 27, 2011 12:53 am

Convert your SUPERC utility output:

dbzTHEdinosauer · Posted: Wed Jul 27, 2011 1:01 am

superc will only generate one list per step. that step can contain a complete pds to pds compare, or a member compare.

the only way you are going to get separate files (pds members in your case)
is to run separate SUPERC Steps for each member match.

as Akatsukami indicated,
generate some jcl, then massage it with rexx.

i am lazy and and after the panel displays the generated jcl
i invoke an edit macro to massage/create new jcl either there,
(normally in another dataset and then submit the new dataset,
and throw the panel generated jcl away.

actually, I don't even bother going to 3... for pds to pds matches.
i have a rexx script that I invoke that:

accesses some dummy (skelton) jcl
accesses a member list of both pds's (actually I build them on the fly with lmm routines)

and generte a step for each matching members,
and populate a separate list 'member' with a list of unmatched members.

gotta be careful of the 255 step limit - more than 255 matching members, means more than one job.

Akatsukami · Posted: Wed Jul 27, 2011 1:15 am

dick scherrer · Posted: Wed Jul 27, 2011 1:28 am

Hello,

Nanda kumaar · New User Joined: 09 Jul 2011 Posts: 8 Location: India

Nic Clouston · Posted: Wed Aug 17, 2011 1:31 am