Fileaid compare layout concatenation problem

niks_jude · Posted: Sat Oct 27, 2012 2:49 am

Following is how my fileaid step looks like which is part of a PROC. In some cases I do not pass the member name and pass the library name in XXPY symbolic in DD01RL. Now I have a requirement to pass the exact copybook member here which it is able to take but without the second library in concatenation. So i tried dummy but it abends (even passed DCB of LRECL=80, RECFM=FB with the dummy but did not work).

dbzTHEdinosauer · Posted: Sat Oct 27, 2012 3:47 am

niks_jude · Posted: Sat Oct 27, 2012 4:22 am

Sorry again, it can be seen as // without comment (in my test i ran commenting it, so that I can get rid of the problem.)

dbzTHEdinosauer · Posted: Sat Oct 27, 2012 5:23 am

well, remove the comment
and provide us with the jes expansion of the PROC
with your override problem
so we can help you.

niks_jude · Posted: Sat Oct 27, 2012 8:16 am

This is the error I am getting -

dick scherrer · Posted: Sat Oct 27, 2012 8:31 am

Hello,

Why have you not referred to the Sysgtem messages for the 013 and then to the manual for the IRC141I (that the 013 abend mentions)? Both the z/OS System Codes and the z/OS System Messages (you want VOL VII, i believe) are available via the IBM Manuals link at the top of the page.

Read the explanaton of the -64 in the documentation and post back here if something is not clear.

You really need to spend some time learning how to do the basic research for abends. All of the new people in the organizations i support learn this in less than a year. Actually, by the time they are working for 2 years, they are helping the new people learn.

niks_jude · Posted: Sat Oct 27, 2012 10:30 am

This is what I had got but did not make sense to me why it is going to open a dummy dataset. I thought dummy would be treated as an empty file in concatenation.

PeterHolland · Posted: Sat Oct 27, 2012 1:20 pm

From the JCL manual :

Use the DUMMY parameter to specify that:
No device or external storage space is to be allocated to the data set.
No disposition processing is to be performed on the data set.
For BSAM and QSAM, no input or output operations are to be performed on the data set.
Code the DCB parameter, if needed. If the program does not supply all the data control block information, make sure that the DCB parameter supplies the missing information.

You are using a sequential dataset, concatenated with a PDS. So there a not matching DCB values, and a PDS will be processed by BPAM (not BSAM or QSAM).

niks_jude · Posted: Sat Oct 27, 2012 1:42 pm

dick scherrer · Posted: Sun Oct 28, 2012 9:14 am

Hello,

Sort may be more forgiving than fileaid. . . What you have coded is documented is not valid per the manuals.

You might copy the member to a temporary qsam dataset and then concatenate the 2 . . .

niks_jude · Posted: Sun Oct 28, 2012 1:00 pm

Fine i was assuming that PDS member would be treated as a PS file and i can just add a dummy PS to it, so only the copybook would be left. A little strange to know this limitation though.

I cant copy to QSAM as I want to use this PROC for dual purposes. I will think of something else.

PeterHolland · Posted: Sun Oct 28, 2012 2:13 pm

Are you using the member name in your DD01RL (lay out library) also
in your control statements? Like LAYOUT=mbr or MAP=mbr?

niks_jude · Posted: Mon Oct 29, 2012 12:14 am

Yes it is there. I do not understand this much but in some cases they are passing the exact copybooks and in some cases they are just passing library as override. Looking at the layouts, the layout has the copybook name defined. I cannot understand what they are achieving by passing a copybook name, looks like you mean to say it will pick it up from the library concatenation. But since I did not want to take any chances as this is a very important job but still wanted to change it have symbolics and concatenations as part of my project, was trying to keep it the same way as current job.

dick scherrer · Posted: Mon Oct 29, 2012 4:09 am

Hello,

niks_jude · Posted: Mon Oct 29, 2012 10:04 am

I cannot copy because in one case I do not have any member at all.

One case

PeterHolland · Posted: Mon Oct 29, 2012 11:21 am

My guess is, the moment FILEAID sees a control statement with a lay-out member, it is going to do BPAM access. And that is not possible on a sequential dataset (e.g. pds(mbr)). So your 013 abend is not created by your dummy dd statement.

niks_jude · Posted: Mon Oct 29, 2012 12:07 pm

I guess that is the case, i will not do this in production but will try removing the member names from control card and see what happens. I guess they originally decided to pass the member name because they did not want to mention it in the control card - not sure why they did this kind of different coding?

PeterHolland · Posted: Mon Oct 29, 2012 12:15 pm

Im almost 100% sure FILEAID is not going to work without giving the layout names in its control statements.

niks_jude · Posted: Mon Oct 29, 2012 12:42 pm

Then i am curious what is the use of passing a member name instead of whole library if control card is going to have it? It is going to search the copybook in the library and use the same.

PeterHolland · Posted: Mon Oct 29, 2012 12:51 pm

I don't see any use for it, and also FILEAID doesnt. Its not designed to work with member names. What if there are 10 control statements with different layout names?
So you have to find another solution, maybe your sysin using a symbolic dsname for different sets of control statements.

niks_jude · Posted: Tue Oct 30, 2012 1:04 pm

Fine I will try as it looks like but I tried once without the copybook member name and it did work. I was just worried if I am not messing up with anything as its a very important job.

dick scherrer · Posted: Tue Oct 30, 2012 9:10 pm

Hello,

Then it needs to be tested for "everything" (actually everything needs thorough testing - not just that it runs once) to ensure no production problems later.

Have co-workers/team lead suggest ways to try to "break" it . . .

niks_jude · Posted: Wed Oct 31, 2012 2:34 am

Not many people are aware of it, so not sure how to break it, and this is the reason I was looking for the turn around to still have the concatenation and make it work as earlier.