|
|
View previous topic :: :: View next topic
|
| Author |
Message |
suresh1624
New User
Joined: 21 Nov 2007
Posts: 28
Location: chennai
|
|
|
|
Hi All
My input file data is as follows.
=COLS> ----+----1----+----2----+
000300 ........001.......
000301 ........026.......
000302 ........025.......
000303 ........016.......
000304 ........003.......
000305 ........017.......
I have a control file based on which the input data is copied to output file.The control file data is as follows
=COLS> ----+----1----+----2-
000306 00101
000307 00201
000308 00301
000309 01502
000310 01602
000311 01702
000312 02403
000313 02503
000314 02603
for example 001 record in the input should be written to first file and 016 to second file and so on; I use only syncsort. the number files is fixed and need not be generated dynamically.Can any one help me please. |
|
| Back to top |
|
 |
|
|
murmohk1
Senior Member
Joined: 29 Jun 2006
Posts: 1439
Location: Bangalore,India
|
|
|
|
Suresh,
You dint say all the i/p file keys are present in control file or ONLY subset is present.
| Quote: |
000311 01702
000312 02403 |
Though you have not given sufficient examples about ur requirement, I assume the value after your key (keys above 017 & 024) says in which o/p file (02 and 03 files respectively) to the i/p record should be written.
If this is your requirement, usign join keys (I believe- as I have not worked with syncsort much).... join the both the records and using simple outfil....include statement, get your job done (with little reformatting). |
|
| Back to top |
|
|
Alissa Margulies
SYNCSORT Support
Joined: 25 Jul 2007
Posts: 500
Location: USA
|
|
|
|
suresh,
Assuming the key field is (6,3) in the input file, and (1,3) for the control file, and the control file output identifier field is (4,2)...
And also assuming that these are both 80 byte FB data sets, try the following:
| Code: |
//SORT1 EXEC PGM=SORT
//SORTJNF1 DD DISP=SHR,DSN=CONTROL.FILE
//SORTJNF2 DD DISP=SHR,DSN=INPUT.FILE
//SORTOF01 DD DSN=OUTFILE1...
//SORTOF02 DD DSN=OUTFILE2...
//SORTOF03 DD DSN=OUTFILE3...
//SYSOUT DD SYSOUT=*
//SYSIN DD *
JOINKEYS FILES=F1,FIELDS=(1,3,A)
JOINKEYS FILES=F2,FIELDS=(6,3,A)
REFORMAT FIELDS=(F2:1,80,F1:4,2)
SORT FIELDS=COPY
OUTFIL FILES=01,INCLUDE=(81,2,CH,EQ,C'01'),OUTREC=(1,80)
OUTFIL FILES=02,INCLUDE=(81,2,CH,EQ,C'02'),OUTREC=(1,80)
OUTFIL FILES=03,INCLUDE=(81,2,CH,EQ,C'03'),OUTREC=(1,80)
/* |
|
|
| Back to top |
|
|
suresh1624
New User
Joined: 21 Nov 2007
Posts: 28
Location: chennai
|
|
|
|
Alissa,
That was gr8 and indeed helpful;But my client doesn't allow the use of join keys. Is there any performance issues with join keys; they rather go for a cobol pgm instead of using joinkeys when processing millions of records.But anyway you helped me to learn abt joinkeys. Thanks for your help. |
|
| Back to top |
|
|
suresh1624
New User
Joined: 21 Nov 2007
Posts: 28
Location: chennai
|
|
|
|
Alissa,
I'm sorry and forgot to mention one more requirement. If the field from 6 to 3 in the input doesnt have a join with the control file record that records should be written to a default dataset. Is there any way to do it?
Thanks,
Suresh |
|
| Back to top |
|
|
dick scherrer
Site Director
Joined: 23 Nov 2006
Posts: 19270
Location: Inside the Matrix
|
|
|
|
Hello,
If the client has directed a cobol solution, how will additional "join" info help implement the code?
Maybe i've misunderstood  |
|
| Back to top |
|
|
Alissa Margulies
SYNCSORT Support
Joined: 25 Jul 2007
Posts: 500
Location: USA
|
|
|
|
| suresh1624 wrote: |
| I'm sorry and forgot to mention one more requirement. If the field from 6 to 3 in the input doesnt have a join with the control file record that records should be written to a default dataset. Is there any way to do it? |
Yes. Modify the jobstream as follows:
| Code: |
//SORT1 EXEC PGM=SORT
//SORTJNF1 DD DISP=SHR,DSN=CONTROL.FILE
//SORTJNF2 DD DISP=SHR,DSN=INPUT.FILE
//SORTOF01 DD DSN=OUTFILE1...
//SORTOF02 DD DSN=OUTFILE2...
//SORTOF03 DD DSN=OUTFILE3...
//SORTOF04 DD DSN=OUTFILE4.NOMATCH...
//SYSOUT DD SYSOUT=*
//SYSIN DD *
JOINKEYS FILES=F1,FIELDS=(1,3,A)
JOINKEYS FILES=F2,FIELDS=(6,3,A)
JOIN UNPAIRED,F2
REFORMAT FIELDS=(F2:1,80,F1:4,2),FILL=C'0'
SORT FIELDS=COPY
OUTFIL FILES=01,INCLUDE=(81,2,CH,EQ,C'01'),OUTREC=(1,80)
OUTFIL FILES=02,INCLUDE=(81,2,CH,EQ,C'02'),OUTREC=(1,80)
OUTFIL FILES=03,INCLUDE=(81,2,CH,EQ,C'03'),OUTREC=(1,80)
OUTFIL FILES=04,INCLUDE=(81,2,CH,EQ,C'00'),OUTREC=(1,80)
/*
|
|
|
| Back to top |
|
|
Alissa Margulies
SYNCSORT Support
Joined: 25 Jul 2007
Posts: 500
Location: USA
|
|
|
|
| suresh1624 wrote: |
| Is there any performance issues with join keys; they rather go for a cobol pgm instead of using joinkeys when processing millions of records. |
I would expect the use of SyncSort's JOIN feature to perform better than a COBOL application doing the same thing... |
|
| Back to top |
|
|
|
|
