|
View previous topic :: View next topic
|
| Author |
Message |
thundercleesed
New User
Joined: 11 Apr 2006
Posts: 4
Location: Athens, OH
|
|
|
|
I need to search 200 fixed block datasets with mixed record lengths. After I submit the job to search, I receive the following after a 30 seconds of execution
| Code: |
| SYSTEM COMPLETION CODE=0C4 REASON CODE=00000011 |
The output dataset does have a few results, but I do not believe all of the datasets were searched before the abend. Not sure how to proceed from here, any suggestions?
| Code: |
//SEARCH EXEC PGM=ISRSUPC,
// PARM=(SRCHCMP,'')
//OUTDD DD DSN=AAA.BBB.SORTED,
// UNIT=DISK,
// DISP=(NEW,CATLG,KEEP),
// DCB=(LRECL=134,RECFM=FB),
// SPACE=(TRK,(10,5),RLSE)
//SYSIN DD *
SRCHFOR 'FIRSTNAME'
SRCHFOR 'MIDDLENAME'
SRCHFOR 'LASTNAME'
/*
//NEWDD DD DSN=YYY.ZZZ.DD001,DISP=SHR
// DD DSN=YYY.ZZZ.DD002,DISP=SHR
// DD DSN=YYY.ZZZ.DD003,DISP=SHR
// DD DSN=YYY.ZZZ.DD004,DISP=SHR
// DD DSN=YYY.ZZZ.DD005,DISP=SHR
// DD DSN=YYY.ZZZ.DD006,DISP=SHR
// DD DSN=YYY.ZZZ.DD007,DISP=SHR
.
.
.
// DD DSN=YYY.ZZZ.DD200,DISP=SHR
|
|
|
| Back to top |
|
 |
Robert Sample
Global Moderator
Joined: 06 Jun 2008
Posts: 8700
Location: Dubuque, Iowa, USA
|
|
|
|
From the JCL Language Reference (link at the top of the page):
| Quote: |
12.1.6.6 Logical Record Lengths for Concatenated Data Sets
Concatenated data sets with format-V records can have different logical record lengths as long as the data set with the largest logical record length appears first in the concatenation. (Note that you can state a value equal to the largest logical record length for LRECL on the first DD statement, regardless of what the actual logical record length of this data set is.) |
|
|
| Back to top |
|
|
thundercleesed
New User
Joined: 11 Apr 2006
Posts: 4
Location: Athens, OH
|
|
|
|
I modified the JCL to include the DCB for each dataset
| Code: |
| DCB=(RECFM=FB,LRECL=####) |
I am still receiving the same errror though, any suggestions?
| Code: |
//SEARCH EXEC PGM=ISRSUPC,
// PARM=(SRCHCMP,'')
//OUTDD DD DSN=AAA.BBB.SORTED,
// UNIT=DISK,
// DISP=(NEW,CATLG,KEEP),
// DCB=(LRECL=134,RECFM=FB),
// SPACE=(TRK,(10,5),RLSE)
//SYSIN DD *
SRCHFOR 'FIRSTNAME'
SRCHFOR 'MIDDLENAME'
SRCHFOR 'LASTNAME'
/*
//NEWDD DD DSN=YYY.ZZZ.DD001,DISP=SHR,
// DCB=(RECFM=FB,LRECL=4096)
// DD DSN=YYY.ZZZ.DD002,DISP=SHR,
// DCB=(RECFM=FB,LRECL=4096)
// DD DSN=YYY.ZZZ.DD003,DISP=SHR,
// DCB=(RECFM=FB,LRECL=1938)
// DD DSN=YYY.ZZZ.DD004,DISP=SHR,
// DCB=(RECFM=FB,LRECL=994)
// DD DSN=YYY.ZZZ.DD005,DISP=SHR,
// DCB=(RECFM=FB,LRECL=850)
// DD DSN=YYY.ZZZ.DD006,DISP=SHR,
// DCB=(RECFM=FB,LRECL=613)
// DD DSN=YYY.ZZZ.DD007,DISP=SHR,
// DCB=(RECFM=FB,LRECL=600)
.
.
.
// DD DSN=YYY.ZZZ.DD200,DISP=SHR,
// DCB=(RECFM=FB,LRECL=13) |
|
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19243
Location: Inside the Matrix
|
|
|
|
Hello,
Are the files fixed or variable length? Fixed, right?
Suggest you use a proc/pend and set up the dsn as a symbolic parameter and process one file per execution. The output can all be written (disp=mod) into the same output file to make use easier after the searches are done. Also, suggest you remove the dcb info from the jcl - it is not needed.
You would want a step before the PROC statement (iefbr14 or idcams) to catalog a new "output" file, so that all of the executions would append via the "mod". You would have 200 EXECs rather than 200 DDs. . . Suggest the sysin data be placed into a dataset so the search input need not be repeated. |
|
| Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008
Posts: 8700
Location: Dubuque, Iowa, USA
|
|
|
|
| Sorry, I missed the fixed in your post. My solution works for variable but not fixed; use Dick's solution for the fixed length files. |
|
| Back to top |
|
|
thundercleesed
New User
Joined: 11 Apr 2006
Posts: 4
Location: Athens, OH
|
|
|
|
| I am curious as to why Robert's solution will only work with variable length? |
|
| Back to top |
|
|
Robert Sample
Global Moderator
Joined: 06 Jun 2008
Posts: 8700
Location: Dubuque, Iowa, USA
|
|
|
|
From the JCL Language Reference (link at the top of the page):
| Quote: |
There are two types of sequential concatenation:
* Like. The same logical record length (LRECL) value and record format apply to all the data sets. The block size (BLKSIZE) values and device characteristics might differ.
* Unlike. Any of data set characteristics might differ. The application program must have logic to support this capability. |
which goes with
| Quote: |
12.1.6.6 Logical Record Lengths for Concatenated Data Sets
Concatenated data sets with format-V records can have different logical record lengths as long as the data set with the largest logical record length appears first in the concatenation. (Note that you can state a value equal to the largest logical record length for LRECL on the first DD statement, regardless of what the actual logical record length of this data set is.) |
You are explicitly allowed to use variable length records with different LRECL as long as the longest goes first. However, the same does not apply to fixed length records. |
|
| Back to top |
|
|
thundercleesed
New User
Joined: 11 Apr 2006
Posts: 4
Location: Athens, OH
|
|
|
|
Thanks for the help and suggestions. For future reference, here is what I came up with
| Code: |
//**********************************************************************
//* CREATE: Create a new empty dataset to hold search results...
//**********************************************************************
//CREATE EXEC PGM=IEFBR14
//NEWDD DD DSN=AAA.BBB.RESULTS,
// UNIT=DISK,
// DISP=(NEW,CATLG),
// DCB=(LRECL=134,RECFM=FB),
// SPACE=(TRK,(10,5),RLSE)
//**********************************************************************
//* QUERYTXT: Place some search terms into temporary dataset...
//**********************************************************************
//QUERYTXT EXEC PGM=IEBGENER
//SYSABEND DD SYSOUT=*
//SYSTOTAL DD SYSOUT=*
//SYSLIST DD SYSOUT=*
//SYSPRINT DD SYSOUT=*
//SYSUT2 DD DSN=&&QUERY,UNIT=DISK,
// DISP=(NEW,PASS),
// DCB=(LRECL=80,RECFM=FB),
// SPACE=(TRK,(10,5),RLSE)
//SYSIN DD DUMMY
//SYSUT1 DD *
SRCHFOR 'FIRSTNAME'
SRCHFOR 'MIDDLENAME'
SRCHFOR 'LASTNAME'
/*
//**********************************************************************
//* SEARCH: In stream proc for searching given datasets...
//**********************************************************************
//SEARCH PROC
//SEARCHIT EXEC PGM=ISRSUPC,
// PARM=(SRCHCMP,'')
//OUTDD DD DSN=AAA.BBB.RESULTS,
// UNIT=DISK,DISP=(MOD)
//SYSIN DD DSN=&&QUERY,DISP=SHR
//NEWDD DD DSN=&DSNAME,DISP=SHR
// PEND
//**********************************************************************
//* STEPS: Pass a dataset to search proc for processing...
//**********************************************************************
//STEP001 EXEC SEARCH,DSNAME=YYY.ZZZ.DD001
//STEP002 EXEC SEARCH,DSNAME=YYY.ZZZ.DD002
//STEP003 EXEC SEARCH,DSNAME=YYY.ZZZ.DD003
//STEP004 EXEC SEARCH,DSNAME=YYY.ZZZ.DD004
//STEP005 EXEC SEARCH,DSNAME=YYY.ZZZ.DD005
.
.
.
//STEP200 EXEC SEARCH,DSNAME=YYY.ZZZ.DD200 |
|
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19243
Location: Inside the Matrix
|
|
|
|
You're welcome - thanks for the follow up 
d |
|
| Back to top |
|
|
|
|
  |
All times are GMT + 6 Hours |
|
