|
View previous topic :: View next topic
|
| Author |
Message |
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
But, can we sort that huge number of input records that are residing in the tape?
The job is utilizing huge amount of CPU time and at the end its abending with SE37. |
|
| Back to top |
|
 |
gcicchet
Senior Member
Joined: 28 Jul 2006
Posts: 1702
Location: Australia
|
|
|
|
Hi,
it makes no difference whether you are sorting from disk or tape.
The SE37 is on which file ?
Why are you sorting the file if you are extracting data ?
Extract the data and then sort it. From my understanding, you are not selecting many records.
Gerry |
|
| Back to top |
|
|
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
The runjcl that I am using is as under:
| Code: |
//SORT01 EXEC PGM=SORT
//*
//SYSPRINT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//*
//SORTIN01 DD DISP=SHR,DSN=&NAME..SSET2.MITDPFI.DEC2005.CMI
//SORTIN02 DD DISP=SHR,DSN=&NAME..SSET3.MITDPFI.DEC2005.CMI
//SORTIN03 DD DISP=SHR,DSN=&NAME..SSET6.MITDPFI.DEC2005.CMI
//SORTOUT DD DSN=&NAME..CLAIMDEC.OUT,
// DISP=(NEW,CATLG,CATLG),
// DCB=(RECFM=FB,LRECL=5200,BLKSIZE=0,DSORG=PS),
// UNIT=3390,SPACE=(TRK,(3335,3335),RLSE)
//SYSIN DD *
MERGE FIELDS=(24,7,CH,A)
INCLUDE COND=(24,7,CH,EQ,C'0125032',OR,
24,7,CH,EQ,C'0175002')
SUM FIELDS=NONE
/*
//* |
|
|
| Back to top |
|
|
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
| I am getting the SE37 abend on the SORTOUT file. |
|
| Back to top |
|
|
gcicchet
Senior Member
Joined: 28 Jul 2006
Posts: 1702
Location: Australia
|
|
|
|
Hi,
like it has been mentioned earlier, run it 1 file at a time and merge at the end.
Output can be written to tape if space is an issue.
Can you also post the joblog output including messages.
Gerry |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
|
|
I might be wrong, but there is something not clear in the whole approach
( the record number does not look that big )
if the need is simply to extract some records from some tape files
a merge application is overkill..
the main drawback is that it will need as many tape units ( real or virtual )
as the sortin statements
whats' s wrong in running ( a purist might object - but I am no syncsort xpert ) two steps...
the first one to extract the records ...
something like
| Code: |
option copy
include ..... |
and the other one to sort as needed
or using the proper control statement include/discard the wanted/extraneous records
in the input stage ( before sorting ) ??? |
|
| Back to top |
|
|
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
Hi Enrico,
The method that you have suggested can be tried as the last option as there are 30 such tape files for a particular month and I have to extract the 1800 records for 9 months. So, the total number of records increased to 270 in total.
That was the reason I was asking whether we can directly extract the data from the tapes without copying them into DASD.
I thought of a method. In that, I would extract the data from each individual file from tape (if there is any method to do so). After that for the next month, I would proceed to append the data into the same file. Thus this would help me to save one more merge step. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10873
Location: italy
|
|
|
|
I might be dumb ..
but my suggestion just does that it extracts the "include ..." things from tape to dasd
or am I missing something else 
here is a possible flow...
define an accumulation dataset an write an end of file
for each tape group
.... extract the needed records to
copy the accumulation dataset to a temporary dataset
( I do not like to use as sortout a sortin ! )
sort/merge the extracted records to the accumulation dataset,
cleanup as needed |
|
| Back to top |
|
|
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
hi Gerry,
Sorry for the late response. The Job log and the output message are as under:
Job Log:
| Code: |
TMS009 IEC501A M 7D3D,W90523,SL,COMP,XXXXXXT1,SORT01.K3709T01,&NAME..SSET2.MITDPFI.DEC2005.CMI
+WER068A XXXXXXT1,K3709T01,SORT01 -OUT OF SEQ SORTIN01 , BLOCK 393,345
IEF450I XXXXXXT1 SORT01 K3709T01 - ABEND=S000 U0016 REASON=00000000 075
TIME=08.19.21
|
Message:
| Code: |
N48FF9T1 SORT01 K3709T01 - COMPLETION CODE - SYSTEM=000 USER=0016 REASON=00000000
XXXXXX.XXXXXXT1.J0043298.D0000102.? SYSOUT
XXXXXX.XXXXXXT1.J0043298.D0000103.? SYSOUT
&NAME..SSET2.MITDPFI.DEC2005.CMI KEPT
VOL SER NOS= W90161,W90281,W90523,W91164.
&NAME..SSET3.MITDPFI.DEC2005.CMI KEPT
VOL SER NOS= W90238,W90507,W90642,W90765.
&NAME..SSET6.MITDPFI.DEC2005.CMI KEPT
VOL SER NOS= W90428,W90606,W90759,W90915.
&NAME..CLAIMDEC.OUT RETAINED, DDNAME=SORTOUT
XXXXXX.TEST.BED.PULL.SRUNJCL RETAINED, DDNAME=SYSIN
STEP/SORT01 /START 2008185.0741
STEP/SORT01 /STOP 2008185.0819 CPU 0MIN 06.53SEC SRB 0MIN 01.11S
** PRGMNAME CORU CORA CLKTME 3033CPU CORTME DSKTME
** SORT 3096K 6144K 38.0 234.9 11764.9 0.0
JOB/XXXXXXT1/START 2008185.0741
JOB/XXXXXXT1/STOP 2008185.0819 CPU 0MIN 06.53SEC SRB 0MIN 01.11S
** KIND PR ABND AVCR CLKTME 3033CPU CORTME DSKTME
** 8200 00 16 3079K 38.2 241.0 11765.1 0.0
|
|
|
| Back to top |
|
|
gcicchet
Senior Member
Joined: 28 Jul 2006
Posts: 1702
Location: Australia
|
|
|
|
The error I can see is +WER068A XXXXXXT1,K3709T01,SORT01 -OUT OF SEQ SORTIN01 , BLOCK 393,345
Nothing about E37
Gerry |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
| Quote: |
| That was the reason I was asking whether we can directly extract the data from the tapes without copying them into DASD. |
There is 100% no reason to copy all the data from tape to dasd.
My earlier suggestion of splitting the process nto multiple jobs would be to select the required records from the tape files directly and copy them to multiple dasd files. These could later be combined or left in the separate files and concatenated into the process that needs the selected data.
What you need to do will take some wall time, but should not be a dasd space problem. Again, it would be good to coordinate with your storage management people. |
|
| Back to top |
|
|
gcicchet
Senior Member
Joined: 28 Jul 2006
Posts: 1702
Location: Australia
|
|
|
|
WER068A OUT OF SEQ SORTINxx[, BLOCK y]
EXPLANATION: A record in the SORTIN data set indicated by xx is
out of sequence according to the FIELDS specification on the MERGE
statement. The number y of the block containing the out-of-sequence |
|
| Back to top |
|
|
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
| So what exactly does this mean? |
|
| Back to top |
|
|
gcicchet
Senior Member
Joined: 28 Jul 2006
Posts: 1702
Location: Australia
|
|
|
|
It means you cannot merge the files
Gerry |
|
| Back to top |
|
|
swapnadeep.ganguly
Active User
Joined: 21 Mar 2007
Posts: 203
Location: India
|
|
|
|
So, its final now that I should follow Dick's method to extract the data from the Tape datasets individually and finally merge the output datasets.
Thank you all for your time and knowledge enhancing suggestions. |
|
| Back to top |
|
|
gcicchet
Senior Member
Joined: 28 Jul 2006
Posts: 1702
Location: Australia
|
|
|
|
Hi,
if each dataset spans across more than 1 volume ie. multi volume, you can extract off individual tapes.
Gerry |
|
| Back to top |
|
|
|
|
