MVSFORUMS.com

sandip · Beginner Joined: 28 Jul 2004 Posts: 18 Topics: 3

Hello Board,

I have used following code -

Phantom · Posted: Thu Dec 16, 2004 1:13 am Post subject:

Sandip,

After the WRITE statement is executed, the logical record is no longer available in CRF-RECORD unless the WRITE is un-successful.

Check the following link for more info.
http://publibz.boulder.ibm.com/cgi-bin/bookmgr_OS390/BOOKS/IGY3LR20/6.2.40.5?SHELF=&DT=20040227004926&CASE=

At the same time if you use WRITE logical-record FROM identifer then the data in "identifier" is preserved.

Hope this helps,

Thanks,
Phantom

sandip · Beginner Joined: 28 Jul 2004 Posts: 18 Topics: 3

Hi Phantom,
I understand and thank you for your information.
Can you please suggest how the value changes if we try to display the same? Is it random?

Regards,
Sandip.

Phantom · Posted: Thu Dec 16, 2004 1:45 am Post subject:

Sandip,

In ur first post, You said:

sandip · Beginner Joined: 28 Jul 2004 Posts: 18 Topics: 3

Phantom,
Yes, I can see in some cases values are changing/does not changing after WRITE is issued successfully (I have checked the file status). The output file shows correct value i.e., the value before WRITE.

Regards,
Sandip.

Phantom · Posted: Thu Dec 16, 2004 4:00 am Post subject:

Not quite sure how/why this happens. Probably someone else could provide a solution.

Thanks,
Phantom

s_shivaraj · Posted: Thu Dec 23, 2004 7:30 am Post subject:

FD CRF-FILE.
01 CRF-RECORD.
05 ....
05 ....
.
.
05 CRF-CATEGORY PIC 9(05)
.
.

I am also facing the same type of problem. When I tried to write a record after changing only CRF-CATEGORY, it is writing only the CRF-CATEGORY in the file with some junk in the other fields, which is unchanged.

But when I tried to run the Job by making the output file as SYSOUT, it is working Fine.

slade · Posted: Thu Dec 23, 2004 10:13 pm Post subject:

When a rec is written from the FD rec desription, the rec desc is pointing to a buffer segment (in memory) managed by the access method. When the rec is written the pointer is bumped to the next segment. Since it hasn't been filled by the pgm code, the buffer area probably contains binary zeros (not guaranteed).

The data you moved into the buffer segment before the write is no longer available after the write. The new sement is awaiting new data. The old data may still be in the buffer but the pgm has no way to access it because the poinnter has been adjusted for the next write.
_________________
Regards, Jack.

"A problem well stated is a problem half solved" -- Charles F. Kettering

s_shivaraj · Posted: Fri Dec 24, 2004 12:01 am Post subject:

Jack,
Thanks for your response.

haatvedt · Posted: Fri Dec 24, 2004 3:15 am Post subject:

Sivaraj,

if you direct a file to sysout, then it is NOT blocked. Since JES files are not blocked and do not use buffers, the record pointer is not advanced and the record is available after the write statement.

ps.. this is a common bug when converting a file from sysout / JES to DASD or Tape.

Happy Coding...
_________________
Chuck Haatvedt

email --> clastnameatcharterdotnet

(replace lastname, at, dot with appropriate
characters)

s_shivaraj · Posted: Fri Dec 24, 2004 4:14 am Post subject:

Thanks Chuck . Very Happy

_________________
Cheers
Sivaraj S

'Technical Skill is the Master of complexity, while Creativity is the Master of Simplicity'

Phantom · Posted: Fri Dec 24, 2004 9:44 am Post subject:

Slade,

slade · Posted: Fri Dec 24, 2004 5:02 pm Post subject:

Hi Phantom,

I don't know this for a fact, but I suspect that the last rec written from a buffer can be "seen" because the fetch for a new buffer is delayed until the next write.

Remember that the actual physical write doesn't occur until the buffer is full or at close.
_________________
Regards, Jack.

"A problem well stated is a problem half solved" -- Charles F. Kettering

Phantom · Posted: Sat Dec 25, 2004 6:11 am Post subject:

Hi Slade,