Native PDS format is dependent on MVS DASD track geometry. What you
have, if it's on a PC, is a sequential file representation of a PDS
file, and the type of representation and its format is the issue, not
the native PDS structure. But, if you're talking about files downloaded
from the historically-named CBT tape site , they should all be in
"standard" TSO TRANSMIT (XMIT)or XMI format, which contains all the info
TSO RECEIVE needs to rebuild the PDS. All you really need to access
these on a PC is to install on your PC "XMIT Manager" (also free from
http://www.cbttape.org/), and you will be able to view, save, whatever,
EBCDIC PDS members directly without any additional work.
Paul Robinson wrote:
I am attempting to unpack on a PC, an EBCDIC Partitioned Data Set (PDS)
container file and extract the members (it's one of many I have
downloaded from the tape library, that one that has over 500 mainframe
tape files on it, whose name escapes me). This means I want to be able
to do extractions programmatically by reading the directory and any
associated metadata contained in the PDS file for that purpose.
I need to find out the format of an IBM Mainframe PDS file and have
been unable to find any references for it. I believe it is similar in
concept to the old LBR archive file format (as opposed to say, ARC or
ZIP archives because LBR did not compress files and ARC and ZIP do).
Short version of a long story: I do not have access to a mainframe and
Hercules won't work on my Windows XP, but I have MUSIC/SP (sans
utilites to read PDS datasets), which means I can't use tools like
IEBGENER or whatever is normally used on mainframes for that purpose,
so I wrote some quickie Turbo Pascal programs under MSDOS to, among
many tools related to this purpose, to read an EBCDIC coded file and
translate the EBCDIC data into ASCII so I can either extract the
members of the PDS untranslated as pure EBCDIC or extract and translate
(for doc files and source code).
I have looked at the PDS files directly and can see that they consist
of some kind of metadata such as prefix information, some kind of
directory, then individual files, but with some extra coding stuck at
points in the middle of the individual members which causes problems
because they are then contaminated and I don't necessarily know where
or how this extra code is inserted. It also causes problems for
extracting, say, object files where I might not necessarily know what
is actual ESD/RLD/TXT object record data and what is some of the file
structure or metadata of the PDS container to which it is a member.
I would like to find either the format of a PDS or a reference to one
of IBM's on-line manuals that I can download that would have it.
Once I extract the individual members I can either upload them in
binary as pure EBCDIC or I can translate them into source code as
appropriate.
I think I am at least a fair quality programmer, but I want to learn
more and I need to read other people's code to do that.
I figure once I get more experience I'll be good at it. I've been a
notary public for 24 months and a computer programmer for 24 years.
Thank you for any assistance you may be able to offer. You may send
replies to the list or to me directly if you prefer.
Paul Robinson
"A computer programmer and Notary Public
in and for the Commonwealth of Virginia, at large."
--
Joel C. Ewing, Fort Smith, AR [EMAIL PROTECTED]
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
