Removing last 8 bytes for gdg file names

kitchu84 · New User Joined: 02 Dec 2006 Posts: 33 Location: chennai

hi ,

my input file is like this: LRECL=100, RECFM=FB (File names are of varying lengths)

XXXX.ABC.FILE.ABBYS.LART.TEST.G1423V00 - GDG file
XXXX.MNO.FILE.SAMPLE.ART.TEST01.G0003V00 - GDG file
XXXX.KLM.FILE.TES101.G0203V00 -GDG file
XXXX.ABC.FILE.TEST - flat file

Output file should not contain GDG versioning in the file names

XXXX.ABC.FILE.ABBYS.LART.TEST - GDG file
XXXX.MNO.FILE.SAMPLE.ART.TEST - GDG file
XXXX.KLM.FILE.TES101 -GDG file
XXXX.ABC.FILE.TEST - flat file

I have SYNCSORT available . Can I achieve this without using COBOL

enrico-sorichetti · Posted: Thu May 19, 2011 11:58 am

You /the powers of Your organization should revisit the requirement
a GDG datasetname/generation without the last qualifier is not a Dataset name any longer
it is a GDG base name

and why would You mix GDG base name with normal datasets ???

enrico-sorichetti · Posted: Thu May 19, 2011 12:02 pm

follow on
the question might be reasonable if You only had GDGs
and wanted to check for orphans
( missing gdg bases/uncataloged datasets )
something very commendable for proper storage management

but the presence of a normal dataset is puzzling/baffling

vasanthz · Posted: Thu May 19, 2011 1:42 pm

Hello,

You could try the below job for the requirement,

enrico-sorichetti · Posted: Thu May 19, 2011 1:47 pm

Hi Vasanth

while the answer is perfect from a technical point of view
applying a perfect solution to a flaky requirement is prone to troubles, sooner or later !

it would still be nice to understand why the TS feels right to mix GDG bases with true datasets

expat · Posted: Thu May 19, 2011 1:58 pm

Actually, given the terms used by the OP, and the expected output from the example, I raise the following question.

If the GDG versioning (not my choice of words) is to be removed, then surely the results would be

INPUT:
XXXX.ABC.FILE.ABBYS.LART.TEST.G1423V00 - GDG file

OUTPUT:
XXXX.ABC.FILE.ABBYS.LART.TEST.G1423V - GDG file

or perhaps
XXXX.ABC.FILE.ABBYS.LART.TEST.G1423 - GDG file

Am I being pedantic here by showing yet another misuse of terminology

vasanthz · Posted: Thu May 19, 2011 2:09 pm

enrico-sorichetti · Posted: Thu May 19, 2011 2:10 pm

just curious and found this sort control cards alternative ( worked with dfsort )

Nic Clouston · Posted: Thu May 19, 2011 2:53 pm

Also: a GDG dataset is not a specific type of dataset - it could be a flat (PS) dataset, a dump or whatever. Just follows a naming specification. I haven't tried it but I suppose it could be that you could catalog a dataset that is not a part of a GDG but has a dsn ending in the same format as a GDG. Must try it some time.

vasanthz · Posted: Thu May 19, 2011 2:58 pm

Hello,

kitchu84 · New User Joined: 02 Dec 2006 Posts: 33 Location: chennai

Thanks Vasanthz, Enrico - it worked really well.

Now I can use it to suit my requirements.

[/code]

enrico-sorichetti · Posted: Fri May 20, 2011 12:30 am

it would be nice on Your side , since we spent time giving You a solution,
to tell us the illogic reason for mixing up GCG bases and normal datasets

but naturally You did not care to acknowledge this previously expressed wish, uhu ???

Robert Sample · Posted: Fri May 20, 2011 12:34 am

vasanthz, I tried with IEFBR14 and got back

Nic Clouston · Posted: Fri May 20, 2011 12:58 am

Ah - I was just about to play with that. Saved me some (minimal) effort, Robert!!!

kitchu84 · New User Joined: 02 Dec 2006 Posts: 33 Location: chennai

hey Enrico ,

Yea Ofcourse I was about to post the details... I am also multi tasking with other office work as well, just though to thank u first and then provide a detailed explainations once I reach home.

I have requirement to get the exact file names being used in a JCL and their counts .

1. At the end of all the job steps, I am planning to read the JESYSMS while the job is running and then extract all the exact filenames being used in the JOb. (Filenames could be a GDG version or Flat file both).

2.Now in the earlier job steps which use the i/p files and create the o/p datasets. I am planning to create a report file containing names of the files and their counts. However these file names will not be exact filenames. (these will not have exact GDG version numbers)

3. Now i need to compare the exact file names in Step 1 (excluding GDG version) and Step2 (which contains file names and counts) and then finally create a FINAL file with exact filename and counts.

Not sure if this sounds confusing. Please let me know.

enrico-sorichetti · Posted: Fri May 20, 2011 1:19 am

counts of what ...
do not tell me that You want to count the record of each file

kitchu84 · New User Joined: 02 Dec 2006 Posts: 33 Location: chennai

To add to it :

Since I need to match on 2 files one with exact file names and other file
base file name and count. I asked this query.

Also, I tried changing the sort card to achieve this.

Input file: LERCL=100,RECFM=FB
File Name: 80 bytes(1:80) - last 9 bytes are for version number EX: '.G1423V00'

File count : 11 bytes(90:11)

last bytes view of the input file:

+----9----+----0
****************
00024560010
00003300010
00330330010
00000005510

Sort Card I used:

//SYSIN DD *
SORT FIELDS=COPY
INREC FIELDS=(1,80,JFY=(SHIFT=RIGHT),90,11)
OUTREC IFTHEN=(WHEN=(72,1,CH,EQ,C'.',AND,
73,1,CH,EQ,C'G',AND,
78,1,CH,EQ,C'V',AND,
74,4,FS,EQ,NUM,AND,
79,2,FS,EQ,NUM),
BUILD=(1,71,JFY=(SHIFT=LEFT),72:72,9,90:90,11)),
IFTHEN=(WHEN=NONE,
BUILD=(1,80,JFY=(SHIFT=LEFT),90:90,11))
/*

Output that I got :
LRECL=100,RECFM=80

I am getting first 71 bytes of file names correctly
Next 9 bytes of version number also properly starting position 72. But the
file counts are not coming at position 90 correctly.

last bytes view of the output file:

7----+----8----+----9----+----0
*******************************
.G1423V00 10
.G0003V00 10
.G0203V00 10
10

Please suggest me.

kitchu84 · New User Joined: 02 Dec 2006 Posts: 33 Location: chennai

Pardon me for the file snapshots, it is not alinged properly while pasting ...

kitchu84 · New User Joined: 02 Dec 2006 Posts: 33 Location: chennai

At enrico:

"counts of what ...
do not tell me that You want to count the record of each file "

Yes we need the count of files being used in the job. However, it will be upto our discretion for which files we want to display the counts. So now we are planning to make changes to existing cobols/ezt that create that file or call ICETOOL to get the file counts. For volume intensive files, we might not want to call ICETOOL, just that we will make change to modules to have the counts displayed while creating the files itself.