comma separate the records in a file using DFSORT

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

My input file looks like -

:20:12345
:50:
abc
cde
efg
:20:5678912345
:50:
zaq
xswcde
vfrbgtnhy

using dfsort or icetool i would like to have the output file as

12345,abc,cde,efg
5678912345,zaq,xswcde,vfrbgtnhy

Basically
1) the record starts with :20:---(values employee name)
2) the next record starts with :50: ---(values divided in 3 lines, this contains the address of employee in :20:)
3) the next again starts with :20:

the file is of lrec=80 and i want use dfsort to comma separate the values under :20: and :50: as employee,address

then the next record in output file as employee,address

gylbharat · Posted: Tue May 24, 2011 5:52 pm

Refer this ibmmainframes.com/viewtopic.php?t=54473

Similar to your problem

sqlcode1 · Active Member Joined: 08 Apr 2010 Posts: 577 Location: USA

sandeep badal,

Will there be anything in the :50: line or Is it just an indicator for the address line? Will there be any other values than :20: and :50: probably to indicate other fields than employee name and address? What is the max. field size for Employee Name and Employee Address field?

Also, what is your DFSort PTF level?

Thanks,

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

:50: is just the indicator for address line...

The input file have

1) :20:( indicator to employee name can vary with length from 4 to max 20 char length
2) :50: ( indicator to address can very with length from 4 to 30 char ( inclusive of all spaces)
3) this address is split over 3 lines
ex :

:20: 12345678901234567890
:50:
SANDEEP BADAL
UNIT 10 2-4 4 AVEN
3RD CRSS
:20: 123456789
:50:
XYZ
4TH CROSS
ABCDEFG ZYXW
:20:

4) Each em,ployee name is followed by address
5) there will be no other values other than these 2 indicators only
6) LREC = 80,FB

OUTPUT file required as

12345678901234567890, SANDEEP BADAL, UNIT 10 2-4 4AVEN,3RD CRSS
123456789,XYZ,4TH CROSS,ABCDEFG ZYXW
..............

THE dfSORT USED is v 11.0

Any help is highly appreciated

Frank Yaeger · Posted: Tue May 24, 2011 11:32 pm

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

Hi Frank,

The output of the JCL

//S1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTIN DD *
RECORD
//SORTOUT DD DUMMY
//SYSIN DD *
OPTION COPY
/*

1 - CONTROL STATEMENTS FOR 5694-A01, Z/OS DFSORT V1R10 - 14:25 ON WED MAY 25, 20
OPTION COPY

Is this what you looking for?

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

Hi Frank,

Just saw your sticky notes on the finding the ptf version

ICE201I G RECORD TYPE IS F

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

Hi Frank,

Also to let you know we have installed the PTF UK90025 for z/OS DFSORT V1R10

Frank Yaeger · Posted: Wed May 25, 2011 11:48 pm

Here's a DFSORT/ICETOOL job that will give you what I think you're asking for:

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

WOW....Thanks much Frank...This is exactly the way i want it...

Could you please give an explanation on the steps taken in IF then clause

Skolusu · Posted: Thu May 26, 2011 9:16 pm

sandeep badal,

If you're not familiar with DFSORT and DFSORT's ICETOOL, I'd suggest reading through "z/OS DFSORT: Getting Started". It's an excellent tutorial, with lots of examples, that will show you how to use DFSORT, DFSORT's ICETOOL and DFSORT Symbols. You can access it online, along with all of the other DFSORT books, from:

www.ibm.com/support/docview.wss?rs=114&uid=isg3T7000080

Frank Yaeger · Posted: Fri May 27, 2011 9:15 pm

Sandeep,

I'm glad to hear that's what you wanted.

I think you will learn a lot more about DFSORT if you figure out what the control statements do yourself using the references mentioned by Kolusu.

If you still can't figure it out, let me know and I'll go through it for you.

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

Hi Frank,

Sorry to bother again...but i got fixed with the solution provided...the input file is not consistent in the example i gave

The input file have

1) :20:( indicator to employee name can vary with length from 4 to max 20 char length
2) :50: ( indicator to address can very with length from 4 to 30 char ( inclusive of all spaces)
3) this address is split over 3 lines and in some cases 4 lines of address

ex :

:20: 12345678901234567890
:50:
SANDEEP BADAL
UNIT 10 2-4 4 AVEN
3RD CRSS
:20: 123456789
:50:
XYZ
4TH CROSS
ABCDEFG ZYXW
8TH AVENUE ROAD
:20: 1234567
:50:
12XYZ
4TH CROSS
ABCDEFG ZYXW
:20:

4) Each em,ployee name is followed by address
5) there will be no other values other than these 2 indicators only
6) LREC = 80,FB

TAKING THIS RECORD AS EXAMPLE

:20: 123456789
:50:
XYZ
4TH CROSS
ABCDEFG ZYXW
8TH AVENUE ROAD

All the above records having 3 lines of address worked fine but when we have a 4th line it gets truncated when we group records = 4 ...this gets selected and populated as a new record in the next line causing the mess for the rest of the records

INREC IFTHEN=(WHEN=GROUP,RECORDS=4,PUSH=(114:ID=8,122:SEQ=1))

Is there a way we can format based on the tags ...

that is :20: indicates name
:50: indicates address

read the records till the next tag:20: is encountered and then treat it as a new line

OUTPUT file required as

12345678901234567890, SANDEEP BADAL, UNIT 10 2-4 4AVEN,3RD CRSS
123456789,XYZ,4TH CROSS,ABCDEFG ZYXW
123456789,XYZ ,4TH CROSS ,ABCDEFG ZYXW ,8TH AVENUE ROAD

sandeep badal · New User Joined: 24 May 2011 Posts: 8 Location: India

Hi Frank,

i was to figure a solution for it

step 1 ) i introduced a dummy tag :70 at the end of each address

then used the begin and end condition to group the records

step 2)

//TOOLIN DD *
SPLICE FROM(IN) TO(OUT) ON(165,8,CH) WITHEACH -
WITH(22,36) WITH(58,36) WITH(94,36) WITH(130,36) USING(CTL1)
//CTL1CNTL DD *
OMIT COND=(1,3,CH,EQ,C':59',OR,1,3,CH,EQ,C':70')
INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,3,CH,EQ,C':21'),
END=(1,3,CH,EQ,C':70'),PUSH=(165:ID=8,122:SEQ=1)),
IFTHEN=(WHEN=(122,1,ZD,EQ,1),
BUILD=(5,20,JFY=(SHIFT=LEFT,TRAIL=C',',LENGTH=21),
165:165,8)),
IFTHEN=(WHEN=(122,1,ZD,EQ,2),
BUILD=(22:1,35,JFY=(SHIFT=LEFT,TRAIL=C',',LENGTH=36),
165:165,8)),
IFTHEN=(WHEN=(122,1,ZD,EQ,3),
BUILD=(58:1,35,JFY=(SHIFT=LEFT,TRAIL=C',',LENGTH=36),
165:165,8)),
IFTHEN=(WHEN=(122,1,ZD,EQ,4),
BUILD=(94:1,35,JFY=(SHIFT=LEFT,TRAIL=C',',LENGTH=36),
165:165,8)),
IFTHEN=(WHEN=(122,1,ZD,EQ,5),BUILD=(130:1,35,165:165,8))
OUTFIL FNAMES=OUT,
BUILD=(1,164,SQZ=(SHIFT=LEFT,MID=C' ',LENGTH=180))
/*

Frank Yaeger · Posted: Tue May 31, 2011 11:06 pm

Garbage in, garbage out. You said each address was 3 lines so I gave you a solution based on that.

If you can have groups with varying number of lines, that requires a different solution.

Note that your solution is more complicated than it needs to be. You don't have to add a dummy :70 tag or use END. You can just use BEGIN for :20 to group the records.

sqlcode1 · Active Member Joined: 08 Apr 2010 Posts: 577 Location: USA

sandeep badal,
You don't need to create dummy record but rather use :20: to mark beginning of a group of lines representing "Address". Because now you have variable address lines(2 or 3 or 4 lines per address),you need to code for maximum address lines. Below solution assumes you could have max. of 4 lines per address. However, you can change it accordingly to handle more lines. I have also increased final LRECL to be 100 bytes.

Frank had almost given you everything to handle this situation. See if below works...