How to split records into three

farouck · New User Joined: 12 Apr 2006 Posts: 18 Location: France

I must split records of one line to three lines.

I can have x lines but the record key on 1,4 is unique.

My input (fb lrecl 23) is :
1000 ABC DEF KLM
2000 ABB DCC KCC

My output file must be :
VM04 ABC
VM02 DEF
VM0A KLM
DM04 ABB
DM0B DCC
DM0A KCC

i try this jcl but it is wrong
//IN DD DSN=N4GX.TEMPS.TEST.ICETOOL.FILEA,DISP=SHR
//OUT DD DSN=N4G.TEMPS.TEST.ICETOOL.SORTIE,
// DISP=(MOD,CATLG,DELETE),
// UNIT=WORKP,
// SPACE=(23,(1,1),RLSE),AVGREC=K
//TOOLIN DD *
SELECT FROM(IN) TO(OUT) ON(1,4,CH) NODUPS USING(CTL1)
SELECT FROM(IN) TO(OUT) ON(1,4,CH) NODUPS USING(CTL2)
SELECT FROM(IN) TO(OUT) ON(1,4,CH) NODUPS USING(CTL3)
SELECT FROM(IN) TO(OUT) ON(1,4,CH) NODUPS USING(CTL4)
SELECT FROM(IN) TO(OUT) ON(1,4,CH) NODUPS USING(CTL5)
SELECT FROM(IN) TO(OUT) ON(1,4,CH) NODUPS USING(CTL6)
/*
//CTL1CNTL DD *
INREC IFTHEN=(WHEN=(6,3,CH,EQ,C'ABC'),
BUILD=(1:C'VM04 ',6:6,3))
//CTL2CNTL DD *
INREC IFTHEN=(WHEN=(10,3,CH,EQ,C'DEF'),
BUILD=(1:C'VM02 ',6:10,3))
//CTL3CNTL DD *
INREC IFTHEN=(WHEN=(14,3,CH,EQ,C'KLM'),
BUILD=(1:C'VM0A ',6:14,3))
//CTL4CNTL DD *
INREC IFTHEN=(WHEN=(6,3,CH,EQ,C'ABB'),
BUILD=(1:C'DM04 ',6:6,3))
//CTL5CNTL DD *
INREC IFTHEN=(WHEN=(10,3,CH,EQ,C'DCC'),
BUILD=(1:C'DM0B ',6:10,3))
//CTL6CNTL DD *
INREC IFTHEN=(WHEN=(14,3,CH,EQ,C'KCC'),
BUILD=(1:C'DM0A ',6:14,3))
//*

the output is :

VM04 ABC
2000 ABB DCC KCC
VM02 DEF
2000 ABB DCC KCC
VM0A KLM
2000 ABB DCC KCC
DM04 ABB
1000 ABC DEF KLM
DM0B DCC
1000 ABC DEF KLM
DM0A KCC
1000 ABC DEF KLM

thanks for your help

Arun Raj · Posted: Thu Jan 08, 2009 6:48 pm

farouck,

Do you have input records like this? If yes, post the output for that.

farouck · New User Joined: 12 Apr 2006 Posts: 18 Location: France

Hi arcvns,

It is not possible.

But the value are not very important, i search how to code the jcl to split the records

Garry Carroll · Posted: Thu Jan 08, 2009 9:43 pm

Almost. Got the required records, but not in the order you wanted. If it's important, a little more work should get you the right order.

farouck · New User Joined: 12 Apr 2006 Posts: 18 Location: France

thanks gary,

Yes the order is not good, but i'm going to modified it.

Frank Yaeger · Posted: Thu Jan 08, 2009 10:14 pm

farouck,

I doubt that all of those SELECT statements or all of those OUTFIL statements are needed, but I can't tell you how to do what you want in a better way until you explain more clearly what you're trying to do exactly.

If you just want to split each record into three records, you can do it with a DFSORT job like this:

farouck · New User Joined: 12 Apr 2006 Posts: 18 Location: France

frank,

If you agree i give you another sample, it is very different but it is more useful for me:

My iput file (lrecl 18 FB) is :
1000 150,00 140,10
2000 000,00

When on 6,6 the record is equal zero i write VM04 instead of the key, i write the first record on 6,6 and i add U after.
When on 6,6 the record is numeric i write VM04 instead of the key, i write the first record on 6,6 and i add N after.
When on 6,6 the record is not numeric i write nothing

When on 13,6 the record is equal zero i write VM06 instead of the key, i write the first record on 6,6 and i add U after.
When on 13,6 the record is numeric i write VM06 instead of the key, i write the first record on 6,6 and i add N after.
When on 13,6 the record is not numeric i write nothing

The ouput file must be :

VM04 150,00N
VM06 140,10N
VM04 000,00U

In reality I have more columns and records are packed decimal PD.

farouck · New User Joined: 12 Apr 2006 Posts: 18 Location: France

To do that, i wrote this jcl but we can do better?
//S1 EXEC PGM=ICEMAN
//SORTIN DD *
1000 150,00 140,10
2000 000,00
/*
//SYSOUT DD SYSOUT=*
//SORTOUT DD DSN=N4GX.TEMPS.TEST.ICETOOL.SORTIE,
// DISP=(,CATLG,DELETE),
// LRECL=11,
// RECFM=FB,
// UNIT=WORKP,
// SPACE=(11,(1,1),RLSE),AVGREC=K
//SYSIN DD *
OPTION COPY
OUTFIL BUILD=(1:C'VM04 ',6:6,6,/1:C'VM06 ',6:13,6)
//S2 EXEC PGM=ICETOOL
//TOOLMSG DD SYSOUT=*
//DFSMSG DD SYSOUT=*
//SYSPRINT DD SYSOUT=*
//IN1 DD DSN=N4GX.TEMPS.TEST.ICETOOL.SORTIE,
// DISP=SHR
//OUT DD DSN=N4GX.TEMPS.TEST.ICETOOL.SORTIE2,
// DISP=(,CATLG,DELETE),
// LRECL=12,
// RECFM=FB,
// UNIT=WORKP,
// SPACE=(12,(1,1),RLSE),AVGREC=K
//TOOLIN DD *
COPY FROM(IN1) TO(OUT) USING(CTL1)
//CTL1CNTL DD *
INREC IFTHEN=(WHEN=(6,6,CH,EQ,C'000,00'),
BUILD=(1:1,11,12:C'N')),
IFTHEN=(WHEN=(6,6,ZD,GT,0),
BUILD=(1:1,11,12:C'U')),
IFTHEN=(WHEN=(6,6,CH,EQ,C' '),
BUILD=(1:C'BYPASS RECORD'))

Frank Yaeger · Posted: Mon Jan 12, 2009 10:17 pm

Since you say the fields are PD, let's work with PD fields in 6,6 and 13,6.

Here's a DFSORT/ICETOOL job that will do what you asked for:

farouck · New User Joined: 12 Apr 2006 Posts: 18 Location: France

Thanks you Franck,

It is better