Get the name of the Input file where search term is found

sriky86 · New User Joined: 30 Sep 2009 Posts: 8 Location: india

Hi All,

I did attempt to find an existing topic but in vein..

When we search for a particular keyword in a complete GDG file, I want to put the particular version dataset (complete name) which contains the term in a output file.

Eg:
AAA.BBB.CCC - GDG Base
AAA.BBB.CCC.V01
AAA.BBB.CCC.V02
AAA.BBB.CCC.V03

// INPUT1 DD DSN=AAA.BBB.CCC,DISP=SHR
//OUTPU1 DD DSN=XXX.YYY.OUTPUT,DISP=MOD

WHen I give the search term as 'JOHN', if this value is present in AAA.BBB.CCC.V02, I want the output PS datafile to contain.

XXX.YYY.OUTPUT:
JOHN - AAA.BBB.CCC.V02

Thanks for your assistance.

Phrzby Phil · Posted: Thu Dec 23, 2010 6:04 pm

There are so many confusing things here it is hard to know where to begin.

The main issue, leaving the rest for other commentators:

1. If AAA.BBB.CCC is a GDG base, then there will never be a dataset named AAA.BBB.CCC.V01. The first
generation will be named AAA.BBB.CCC.G0001V00. If you
replace that generation with another version of that same generation, then you will have a dataset named AAA.BBB.CCC.G0001V01, which is the next version of the first generation.

Once you can use the correct terminology, then you'll know what you are talking about, and only then will we know what you are talking about, and only then can you pose a cogent question.

sriky86 · New User Joined: 30 Sep 2009 Posts: 8 Location: india

Phil,

Whatever I've put was just a sample. I didn't actually mean that a GDG is in that convention.

Wat I'm talking about is :

Scan through the whole GDG for a SEARCH TERM and finally put the VERSION dataset filename (where the search term is found) into a flat file.

Hope this further explains.

Thanks for your post.

Akatsukami · Posted: Thu Dec 23, 2010 9:04 pm

daveporcelan · Posted: Thu Dec 23, 2010 9:14 pm

sirky86,

sriky86 · New User Joined: 30 Sep 2009 Posts: 8 Location: india

Apologies for the mis-directing posts. I did go through the GDG basic stuffs and now I realize where I went wrong in my request.

Below are the GDG base and the generations
ABC.STOCK.YEAR
ABC.STOCK.YEAR.G0001V00
ABC.STOCK.YEAR.G0002V00
ABC.STOCK.YEAR.G0003V00
ABC.STOCK.YEAR.G0004V00

The JCL should scan through all the generations of the GDG for a specific keyword and when found, it must write the whole name (Eg: ABC.STOCK.YEAR.G0002V00) that holds the keyword and the line in which it is present, in the output dataset.

daveporcelan · Posted: Tue Dec 28, 2010 7:31 pm

In my opinion, this can be done, but is not as simple as one would like.

The problem is if you concatenate all generations, using the base alone, with ISRSUPC or SORT, the data can be identified, but the specific generation can not be accurately identified.

If this was a requirement I had to fulfill (because I was being paid to do so), I would:

1) Use a REXX program to identify the specific generations
2) For each generation, call ISRSUPC internally for each one, writting the output to a dataset with disp=mod
3) Re-format the output dataset as needed to meet the requirement

Like I said, it can be done, but not real easy, especially if you are a novice.

If someone else has a better solution, I am sure the Sriky86 would appreciate it.

dick scherrer · Posted: Tue Dec 28, 2010 10:30 pm

Hello,

One very easy way to do this is to create a small inline PROCedure to run SUPERC/searchfor that has the generation defined as a symbolic parameter - i.e.:
//NEWDD DD DSN=MY.GDGBASE.&GEN,DISP=SHR

Then after the PROC/PEND, execute the PROC once for each generation:
//S1 EXEC THEPROC,GEN=G0001V00
//S1 EXEC THEPROC,GEN=G0002V00
etc.

This requires no knowledge other than basic JCL. The search jcl can be generated from the online screen and copied into the PROC.

Phrzby Phil · Posted: Wed Dec 29, 2010 10:19 am

A SAS program can determine the sequential number of the dataset in a concatenation a record is read from. This would require specifying a concatenation of all generations.

I do not know, however, if it can do so when the concatenation is presented merely by specifying the GDG base.

When specifying the GDG base, the generations are read in reverse order, so I think it unlikely that this would look the same to SAS.

I may just test that some day.

sriky86 · New User Joined: 30 Sep 2009 Posts: 8 Location: india

Hi Dick,

dbzTHEdinosauer · Posted: Mon Jan 17, 2011 6:05 pm

It's your screen, how do you expect us to tell you?
especially since we do not even know the environment within which you are receiving the search request.

use a batch superc
goto 3.15 ispf options. - and then save the jcl generated.
then you will have examples of multiple search statements,

then use the gdg base as your newdd.

then you can write a rexx (or use dfsort) to parse the output and generate your little report.

yeah, you have to go to the end of the superc output to find out which dataset name belongs to the concatenation order.

you could even write a REXX edit macro to accmplish that while you view/edit the dataset.