MVSFORUMS.com

[bauer]

Hi *,

I'm trying to read the directory of a PDS (or PDSE) dataset using PL/1 (or Cobol).

This works for membernames, the structure for a PDS directory is documented for exemple here: OSDFSMS: Using Data Sets, SC26-7410-05.

My problem is now the "last modified" date (like ISPF output). Which bytes are the relevant bytes in the PDS directory records ? How to interpret this bytes?

There are some solutions in the internet available using REXX and the LMMFIND and ZLM4DATE command. But this doesn't help.

I need to do this task in a PL/1 and/or Cobol program.

So can anybody give me some advise, about the relevant bytes and the interpretation of the bytes ?

kind regards,
bauer

[William Collins]

http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/ispzdg90/C.4.4

[bauer]

William Collins,

YES, this looks really pretty good and helpful.

Thank you very much for your very fast answer.

Now, let's do the coding.

Kind regards,
bauer

[bauer]

Ok, I did the coding. It works, thank you again.

[ironponygrl]

[bauer]

ironponygrl,

well I'm, sorry. It is a PL/1 solution, not COBOL.

But here is the coding (only the main parts of it ....), may be, it's helpful for anybody.

The complete routine "ReadPDSDirectory" is a submodule with a callback option to provide each entry of the PDS Directory to the calling (main-) routine.

This is the datastructure which provides information for each entry.

[kolusu]

[kolusu]

[ironponygrl]

thanks fellows, that's awesome. now I can see it clearly and if i ever need to do that, it's here for all perpetuity.

[William Collins]

I always thought that RECFM=U was required for reading a PDS directory. I recently discovered that RECFM works equally. Others have always used RECFM=F and were surprised that RECFM=U worked As important is the LRECL, which everyone agrees is 256.

[bauer]

kolusu,

according your questions:

1) see line DCL PDS FILE RECORD SEQUENTIAL INPUT ENV(U) ; in the PL/1 coding. Here is RECFM = U But, yes you are right, the JCL provides a different RECFM parameter.

2) yes, this is handled. See code line IF SUBSTR(ISPFStatisticsEntry.Misc,1,3) = '404040'X , according the IBM dokumentation: If this bytes are Blank => statistics data available, otherwise not available.

Remark: My here provided coding are only the most important lines. The real coding contains some additional logic for wildcardsearch, descision for reading SSI Data in case of loadmodule libs (type of libaray is evaluated).

kind regards,
bauer

[bauer]

kolusu,

add 2),

In your coding you evaluate this:

IF WS-PDS-INDC = X'00'
DISPLAY 'NO STATS EXIST FOR MEMBER : '
WS-PDS-MEMBER-NAME

This is not handled in my provided coding. Thank you for asking.

My check IF SUBSTR(ISPFStatisticsEntry.Misc,1,3) = '404040'X checks, if the available statistics bytes are filled, content is set.

In our environment we have PDS datasets with available statistics bytes (your checked byte not equal zero) but without a useful content of this bytes.

So my understanding:

I should add your evaluation of the single byte (your field WS-PDS-INDC, my field Entry.INDIC), if statistics data bytes are in general available.

You should add my evaluation, if the statistics bytes have a really filled content.

kind regards,
bauer

[William Collins]

I'm pretty sure that it makes no difference whether RECFM=F or RECFM=U in the JCL or records are Fixed or Undefined in the program. They can even be mixed.

The index consists of individual records each of 256 bytes. Whether F or U, the results (256 bytes of data for every read) are the same. Both F and U work, even in a mixture.

If the index records were blocked, it would be a different thing.

[kolusu]