Creation of Datasets dynamically on the count basis

Raghu M N · New User Joined: 12 Mar 2018 Posts: 2 Location: India

Hi Guys,

Could someone help in resolving this using JCL.

1. I am having a file with n records (zip files) in it.

2. I will get the count from it to one variable (assume COUNT =25)

Now according to this count I need to create datasets dynamically and need to copy files one by one dynamically into the datasets created to process the files.

Thanks

sergeyken · Posted: Fri Mar 16, 2018 2:10 am

You cannot do it in JCL.
Period.

Advice: start learning information technology from scratch.

enrico-sorichetti · Posted: Fri Mar 16, 2018 2:53 am

unfortunately the description of the requirement is clear as mud

Nic Clouston · Posted: Fri Mar 16, 2018 10:53 am

Robert Sample · Posted: Fri Mar 16, 2018 6:38 pm

This is a VERY poorly worded request. Things that have not been answered include:
- where do the output dataset names come from?
- where do the input dataset names come from?
- what is the maximum number of datasets to be created?
- what does "n records (zip files)" mean?
-- Are the input datasets zipped?
-- Are the output datasets to be zipped?
-- Does the site have a mainframe zip program?
--- If not, how is the zip / unzip process being done?
- "create datasets dynamically"
-- using BPXWDYN?
-- using Asembler's dynamic allocation?
-- using some other method not yet described?
- "need to copy files one by one dynamically"
-- using IEBGENER?
--- If so, how are the DD names going to be allocated / freed?
-- using a program?
-- using some other method not yet described?
- will there be times when the count is zero?
-- should the process handle this?

steve-myers · Posted: Fri Mar 16, 2018 8:19 pm

Many, many years ago I looked at ZIP files very carefully. There is no way, repeat, no way to stack multiple ZIP files into a single file except by using ZIP to ZIP multiple ZIP files into a single file. This is rather silly, as the new file will be larger than the sum of the sizes of the component files. In other words, what Raghu M N appears to be thinking cannot be done. Period. End of story.

Rohit Umarjikar · Posted: Fri Mar 16, 2018 8:31 pm

GIMUNZIP?

enrico-sorichetti · Posted: Fri Mar 16, 2018 9:06 pm

from the link posted and by the program name prefix GIM
GIMUNZIP is probably only available for SMPE related/structured data

steve-myers · Posted: Fri Mar 16, 2018 9:08 pm

I'm rather out of date with SMP. I was not aware of GIMUNZIP and its related functions. I followed the GIMUNZIP link Mr. Umarjikar provided and other links in the GIMUNZIP topic.

The issue is this. On toy machines where the number of bytes in a file is an attribute of the file, the first thing a program that is going to process the file does is go to the end of the file to extract something about the file. There is nothing in a ZIP file except, perhaps, the PK data in the first two bytes of the file (in memory of the late Phil Katz) to reliably say this is the start of a ZIP file, and it is all to easy to think PK can be in the data, not just the beginning.

It appears to me GIMZIP performs a PKZIP type function on regular data sets. I have absolutely no idea what the internal format of a GIMZIP data set is, but I very much doubt it is exactly the PKZIP format.