View previous topic :: View next topic |
Author |
Message |
sbpmishra Beginner
Joined: 13 Aug 2007 Posts: 4 Topics: 1
|
Posted: Wed Aug 15, 2007 3:06 am Post subject: COBOL programm skipping records |
|
|
I've a cobol program which while writing records to a VB file, for some reason did not write certain records. This is a simple program where the input is VB & based on certain condition the program writes output to different files. Though from SYSOUTs I can makeout that the program did process all the input records but for some reason few records were not written to output. This happened only once.
Every other time I tried with same set of input files, it gave me proper result.
Could someone please tell me if this can be an issue with program or environment. |
|
Back to top |
|
|
CICS Guy Intermediate
Joined: 30 Apr 2007 Posts: 292 Topics: 3
|
Posted: Wed Aug 15, 2007 3:41 am Post subject: Re: COBOL programm skipping records |
|
|
sbpmishra wrote: | Could someone please tell me if this can be an issue with program or environment. | I'd bet on the program as the cause.....
Can't really tell without looking at the code and reproducing the error.... |
|
Back to top |
|
|
sbpmishra Beginner
Joined: 13 Aug 2007 Posts: 4 Topics: 1
|
Posted: Wed Aug 15, 2007 4:36 am Post subject: |
|
|
I could not recreate the error. I ran the program 100 times after that everytime with the same file & never faced any issue. I was just one time that the program didn't write into the output file.
More interesting is in the same paragraph just before writing the program accumulated counts, they look perfect. Its as if the records were lost. |
|
Back to top |
|
|
dbzTHEdinosauer Supermod
Joined: 20 Oct 2006 Posts: 1411 Topics: 26 Location: germany
|
Posted: Wed Aug 15, 2007 4:54 am Post subject: |
|
|
what differences were there between the first time it was run and each subsequent run.
one nice thing about computers, they do the same thing each time.
if your program was not changed, then the only thing that could change would be output file dispositions. something left over from the first run, caused different results on the second run(s).
don't blame it on the computer. something changed after the first run............... _________________ Dick Brenholtz
American living in Varel, Germany |
|
Back to top |
|
|
jsharon1248 Intermediate
Joined: 08 Aug 2007 Posts: 291 Topics: 2 Location: Chicago
|
Posted: Wed Aug 15, 2007 8:09 am Post subject: |
|
|
How do you know the records were not written? Were there other records on the file or was it just empty? More details please. |
|
Back to top |
|
|
sbpmishra Beginner
Joined: 13 Aug 2007 Posts: 4 Topics: 1
|
Posted: Wed Aug 15, 2007 11:51 am Post subject: |
|
|
I ran the program with same set of input files. The output file names were different but the properties were same. I ran it hundred times but nver got the problem again.
The way I came to know of missign records was afer my program there is a sort step & the sysout in sort shows me less records. The file was not empty. Just skipped some records.
Something else I should look at. Please advise |
|
Back to top |
|
|
jsharon1248 Intermediate
Joined: 08 Aug 2007 Posts: 291 Topics: 2 Location: Chicago
|
Posted: Wed Aug 15, 2007 1:39 pm Post subject: |
|
|
You're not providing enough details. Please reply with more details. Show us a sample of the input data and identify the records that are not written to the output file. Show us some of the relevent code. Show us the sysout from the program and the sort step. Show us the JCL for this program and the sort step. Show us compare results proving that the input files are identical.
One more thing. What do you think happened? It's your application in your shop. Sometimes intuition points us in the right direction. Not always, but I'd be interested to know what you think caused this. |
|
Back to top |
|
|
sbpmishra Beginner
Joined: 13 Aug 2007 Posts: 4 Topics: 1
|
Posted: Wed Aug 15, 2007 2:26 pm Post subject: |
|
|
I'm sorry.
I can't show sample data.
I can't provide the code/JCL/SYSOUT. These are highly confidential for my organization.
I know it sounds rude but I can't help it either.
My intutuion is problem with dasd volume where the file was written.
I can assure you though, the input files are identical. I was hoping if anyone has ever encountered such situation so that I can use their experience.
Input is VB file, output is VB. There is no File Status in program. The output is written directly from inpout record, not from working storage variable. |
|
Back to top |
|
|
jsharon1248 Intermediate
Joined: 08 Aug 2007 Posts: 291 Topics: 2 Location: Chicago
|
Posted: Wed Aug 15, 2007 2:50 pm Post subject: |
|
|
I, and I'll bet most others, have experienced a situation similar to the one you described; the actual outcome is different from what we expected. The only way to determine the cause is to review the code, JCL, sysout, data, etc. But I won't ask for more details again. I won't ask you to provide samples and 'x' out sensitive details. I won't ask what kind of problem you suspect there is with the dasd volume. Speaking only for myself, I will tell you that given the minimal information you supplied, your expectation for a resolution has exceeded my ability to provide one. I'm sorry too. Good luck. |
|
Back to top |
|
|
dbzTHEdinosauer Supermod
Joined: 20 Oct 2006 Posts: 1411 Topics: 26 Location: germany
|
Posted: Thu Aug 16, 2007 6:12 am Post subject: |
|
|
Quote: |
My intutuion is problem with dasd volume where the file was written.
|
If there was a problem with DASD, it will have been documented in your jes output.
Quote: |
I can assure you though, the input files are identical. I was hoping if anyone has ever encountered such situation so that I can use their experience.
|
In my experience, the machine has always done the same thing, as long as the environment was the same. And, in my experience no one has been able to prove that statement false.
If there was a difference is results, and the pgm and jcl were the same, there was a problem. IBM makes notification and you have not looked hard enough to find the 'notification'.
you have overlooked something. simple as that. the machine did not burp. _________________ Dick Brenholtz
American living in Varel, Germany |
|
Back to top |
|
|
jyoung Beginner
Joined: 10 Nov 2005 Posts: 36 Topics: 2 Location: Flint, MI
|
Posted: Fri Aug 17, 2007 11:22 am Post subject: |
|
|
On VSE we use to have this problem when the blocking factor for the file was not the same as what was specified in the program. |
|
Back to top |
|
|
dbzTHEdinosauer Supermod
Joined: 20 Oct 2006 Posts: 1411 Topics: 26 Location: germany
|
Posted: Fri Aug 17, 2007 12:56 pm Post subject: |
|
|
jyoung,
but the problem was constant? i.e. each time you ran with the same input, same jcl and same program, your results were the same. _________________ Dick Brenholtz
American living in Varel, Germany |
|
Back to top |
|
|
jyoung Beginner
Joined: 10 Nov 2005 Posts: 36 Topics: 2 Location: Flint, MI
|
Posted: Fri Aug 17, 2007 1:21 pm Post subject: |
|
|
Yes, that is true. I am having a hard time understanding how with the same exact input each time you could have different results. |
|
Back to top |
|
|
dbzTHEdinosauer Supermod
Joined: 20 Oct 2006 Posts: 1411 Topics: 26 Location: germany
|
Posted: Fri Aug 17, 2007 2:43 pm Post subject: |
|
|
you can't. _________________ Dick Brenholtz
American living in Varel, Germany |
|
Back to top |
|
|
JanM Beginner
Joined: 29 May 2006 Posts: 9 Topics: 0
|
Posted: Mon Aug 20, 2007 6:29 am Post subject: |
|
|
There is one reason I can think of.
Is it possible that an older version of the program is still present on a LINKLIST-library.
If during the first time the program was executed, no steplib was provided, the old program could accidently be called.
If a steplib is used for all susbsequent executions, the 'new' program could be called explaining whyyou got the error only once. |
|
Back to top |
|
|
|
|