How to Get the PS File name

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

Hi..

Can you please suggest that how can we get the PS file name which has generated in the previous steps, in the same job? I want to write that in a work file so that can use it in my application program.

expat · Posted: Mon Jan 11, 2010 9:52 pm

Please explain exactly what you want to do. Along with the requirement to do it in this fashion.

Dataset names are specified in the JCL and remain constant through the job execution lifecycle. Unless you are dynamically generating a dataset name from within a program, which IMHO is pretty bad practice, then the sucessor programs within the same job will need to dynamically allocate the file using the name generated originally.

ofer71 · Posted: Mon Jan 11, 2010 10:29 pm

You can refer back to dataset(s) in previous step(s) by DDNAME.

O.

dbzTHEdinosauer · Posted: Mon Jan 11, 2010 10:37 pm

10 to 1 Anthony is generating a PS file with a date as part of one of the DSN nodes.

MBabu · Posted: Tue Jan 12, 2010 3:29 am

If your program needs to find the actual dsname allocated to a ddname, search the web for programs that search the TIOT and JFCB control blocks, and of course, have someone at your site do a review of them with you to insure they suit your needs, are correct, and that you are legally allowed to use them for your particular pursuits. As ofer71 mentioned, the step looking for the dsname will need a referback dd to point to the ddname in the step that does the original allocation.

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

Requirement goek like this, i have to make use of the NAME of the PS file generated say in third step. Now, in the fifth i have my application program which is getting called and it will take the PS file NAME also as an Input and based on that it will create few out put files (By suffixing some name at the last qualifier).

So, now, its important to paas the name of the PS file (Third Step) to my application program. Now, here i have to use Appliaction code to achieve this as there are few other task what my code is performing aprt from this.

Now, after generating an PS flat file in step-3, can we store it's NAME in an Work file so that i can read it in my program ?

ofer71 · Posted: Tue Jan 12, 2010 11:24 am

You can use LISTALC from within REXX , parse the output and get dataset name.

O.

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

Ihave never worked on REXX. yes, i have tried iN "TSO LISTALC" in ISPF and ot gives you all the datsets currently allocated.
But, in my case when i will have a new datset allocated in third step, can i write an JCL step to execute "TSO LISTALC" command. even if i do that, i will be getting the list of all the datasets allocated currently, but i require only the one which has generated in the third step...

So, to do this, i will have to store the output of the "TSO LISTALC" in a work file do an SORT out of it, to find the dataset what i am looking for ? am i correct ?

dbzTHEdinosauer · Posted: Tue Jan 12, 2010 12:32 pm

how is the DSN being 'generated' in the third step?

expat · Posted: Tue Jan 12, 2010 12:40 pm

As asked previously, how is the dataset name being generated by the program.

Unless you answer the questions we ask with the information we ask for, rather than what you deem we should be allowed to know, there isn't a lot we can do to help you.

Why not just do

dbzTHEdinosauer · Posted: Tue Jan 12, 2010 12:43 pm

or if the DSN is being magically generated in the third step,
use the same brew of mushrooms to generate the DSN in the 5th step.

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

Step 1 is the Extract step, Step 2 is the filter step, then three is the SORT step where we gets an PS flat file with all the attributes (like DCB, DSN, DISP) coded in the JCl itself.

And i require the name of this PS flat file to pas on to my application code.

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

I do not want to make use of the DSN name generated in the Third step to create an new DSN for the next step. i only need to store (or pass) that NAME so that i can apply few logical operations on that in my code.

expat · Posted: Tue Jan 12, 2010 12:53 pm

Anyone here have a map of this thread .................... I'm totally lost

If, and only if, my interpretations are correct, you can specify any DSN you want for the SORTOUT DD name and then use it later in your job/application without having to do anything clever (aka stupid) anywhere at all.

PeterHolland · Posted: Tue Jan 12, 2010 12:56 pm

To summarize all solutions provided :

1. create dataset
2. use IRXJCL with the created dataset allocated
execute LISTALC
get the datasetname
write it to a workfile
3. use the contents of the work file as input

Now was that too difficult?

MBabu · Posted: Tue Jan 12, 2010 1:13 pm

A much better solution is what ofer71 said near the beginning. Use a ddname with referback and don't worry about the data set name at all. If your data set is a &&TEMP dataset, you might not be able to allocate it dynamically anyway. Just use a ddname and read from that.

(btw, LISTALC is a TSO command, you need IKJEFTxx, not IRXJCL. From IRXJCL you could use BPXWDYN but that is way beyond this discussion)

PeterHolland · Posted: Tue Jan 12, 2010 1:21 pm

dbzTHEdinosauer · Posted: Tue Jan 12, 2010 1:25 pm

PeterHolland,

experience has shown that just purely providing 'what the TS wants',
is often

very poor programming standards
illogical
illegal - as in audit
a question based on lack of experience
not really conducive to furthering professionalism

.
The fact that step 3 is a sort step and will have the DSN HARDCODED in the DD statement,
the question arises: Why does the DSN need to be contained in a file?
when we ever receive the answer to that question, a good technical answer will probably be provided.

since your step 1

PeterHolland · Posted: Tue Jan 12, 2010 1:42 pm

Hello Dick,

are you accusing me of the following ? :

expat · Posted: Tue Jan 12, 2010 2:07 pm

Peter,

I can not see where the OP has said that the dataset triggers anything, merely that he wants to use it in a later step of the same job.

The step creating the dataset is a sort step and as DBZ and I have already stated that the dataset name should be hardcoded.

Even if the OP uses the code below without a dataset name

PeterHolland · Posted: Tue Jan 12, 2010 2:18 pm

Hi expat

expat · Posted: Tue Jan 12, 2010 2:29 pm

Peter,

Unless the OP can be bothered to actually tell us anything of any value we will never know what the requirement is or be able to offer any help whatsover.

PeterHolland · Posted: Tue Jan 12, 2010 4:52 pm

expat,

that leaves only a few candidates for help, i guess.

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

As suggested i have used IRXJCL to submit an REXX which will take the input file and will copy the PS file name into OUTSRCH step. Here i will be getting the value of &SFILE from the Previos step-3 which is the actual PS file name say AAA.BBB.

IN the beow PROC i have to capture the file name which is coming in &SFILE ..

Here is the JCL PROC
-----------------------------------------------------
//STEP1 PROC
//IRXJCL1 EXEC PGM=IRXJCL
//SYSEXEC DD DSN=AAA.PROCLIB(REXXPS),DISP=SHR
//SYSPRINT DD SYSOUT=*
//SYSTSPRT DD SYSOUT=*
//SYSTSIN DD DSN=&SFILE,DISP=SHR
//&SFILE
//INSRCH DD DSN=&SFILE,
// DISP=SHR
//OUTSRCH DD DSN=AAA.BBB.OUT,
// DISP=(NEW,CATLG,DELETE),
// DCB=(RECFM=FB,LRECL=80,BLKSIZE=0),
// SPACE=(10,(10,5),RLSE)

Here is the REXX (Well, this is going to be my First REXX Code, so there can be mistakes)..

---------------------------------------------
/* REXX */
"EXECIO * DISKR INSRCH(STEM IN.FINIS";
PARSE UPPER PULL DSN
SAY "DSN=" DSN
PS = SUBSTR(DSN,19,8)
SAY PS
SAY "HELLO"
QUEUE PS
"EXECIO * DISKW OUTSRCH (FINIS";

I am getting the error as "CC 3637" . Can you please help me resolving this ?

nemaji · New User Joined: 29 May 2007 Posts: 22 Location: Pune

Oh, i have missed a lot of post replies..

Actually, I have to code an SPLITTER in Natural language which will take an Big PS Flat file as an Input (This is the STEP-3 File we are talking here) and will Split it into the segments of 65535 (As this is further going to get downloaded as an CSV).

Now, I have my Natural Code ready which will READ the Input file and depending upon the number of records, it will generate the Segments and based in the Segment count (Say 4) it will generate number of IDCAMS step (From Natural). Here, i have the Problem.

When i am writing the IDCAMS Steps using Natural, i need to pass the Infile to it. (However i am generating the outFile programmatically in Natural).

So, Thats the reason i have started this POST so that i can pass the Input Big PS file to Natural.

Also, as this is going to be an generic SPLITTER so the JCL what i have coded is a PROC (Where i will call my Natural Module). So, whatever Job, which creates the Big PS Files and requires an SPLIT will directly Include This PROC in their JCL and will only Provide their InFile name as an input to this PROC. (As i have given the example in my previous post as AAA.BBB).

Now i think i have made my exact requirement very clear :-)