handling the empty file condition

jzhardy · Posted: Thu Aug 09, 2007 7:33 am

I have a requirement to process an input file and produce a comma separated list of numbers.

My input data set could contain 0, 1 or more rows, and for each I want a different output format.

If input is empty, I want one output row containing "0"
if input contains one row, i want just the relevant value from that row.
if input contains more than one row, i want a comma separated list of the relevant values.

So, to make this clearer, suppose my input file had the relevant value at (34,9). My copy step could look something like :

//CPY1CNTL DD *
INREC OVERLAY=(1:SEQNUM,8,ZD)
OUTREC IFTHEN=(WHEN=(1,8,CH,EQ,C'00000001'),
BUILD=(1:C' ',2:34,9,11:70X)),
IFTHEN=(WHEN=NONE,
BUILD=(1:C',',2:34,9,11:70X))
/*

but this only works when the input file is non-empty.

SO...

I could either throw an RC4 on the empty input file condition, and handle this in a separate IF step

or

use HEADER1/TRAILER1 with a constant value of '0'. But there are two problems in doing this :

1. I would get the '0' for all input states, and so I would have to strip this for the non-null condition.
2. I am writing to FB (not FBA) ... i don't know how to do this without getting an ICE224A error ..

is there a elegant one (or two) step solution to give me what I want ? two steps would be great .. one would be really flash.

Frank Yaeger · Posted: Thu Aug 09, 2007 10:54 pm

It's not clear what you want to do. Please show an example of the input record(s) and the expected output record(s) for one row, two rows and five rows.

What is the RECFM and LRECL of the input file?

What is the RECFM and LRECL of the output file?

jzhardy · Posted: Fri Aug 10, 2007 9:00 am

Input file is : RECFM=FB,LRECL=300,BLKSIZE=0

Output is a PDS of type FB 80 (but in some environments this same pds is defined as FBA 80)

For empty input I would like output of :

-1

For input of one row I want just the key value.

eg -

input :
111 xxx

output:
111

For input of more than one row, I want a comma separate list of values

eg -

input
111 xxx
222 yyy

output
111
,222

The following sort job does what I want, but imho does not look that elegant.

//SORT01 EXEC PGM=ICETOOL
//*
//TOOLMSG DD SYSOUT=*
//DFSMSG DD SYSOUT=*
//TOOLIN DD *
*
* dedup and extract key value :
*
SELECT FROM(SORTIN) TO(WORK01) ON(34,9,CH) FIRST USING(SEL1)
*
* REMOVE ASA CHARACTER :
*
COPY FROM(WORK01) TO(PRMLIB) USING(CPY1)
*
/*
//SEL1CNTL DD *
OUTFIL FNAMES=WORK01, REMOVECC,
HEADER1=(1:'-1'),
BUILD=(1:C',',2:34,9,11:69X)
//*
//CPY1CNTL DD *
OUTFIL FNAMES=PRMLIB,
BUILD=(1:2,79,X)
//*
//SORTIN DD DSN=%%HLQSOUT..%%JOBNAME.D%%ODATE.O%%ORDERID.DAT,
// DISP=SHR
//*
//WORK01 DD DSN=&&WORK01,
// DISP=(,PASS),
// SPACE=(%%ALUNIT,(%%PRIS,%%SECS),RLSE),
// RECFM=FBA,LRECL=80,BLKSIZE=0
//*
//PRMLIB DD DSN=TRAILSD4.PARMLIB(TCASIDS),DISP=SHR
//*

also ...

When I looked up ICE224A in my manual i didn't get anything about the REMOVCC option :

-'DFSORT messages, codes and diagnosis guide' :

ICE224A ddname CANNOT BE USED FOR A
REPORT
Explanation: Critical. The OUTFIL report specified for
the OUTFIL data set associated with ddname cannot be
produced because either:

v the OUTFIL data set is VSAM, or
v the record format specified or retrieved for the
OUTFIL data set does not contain â€™Aâ€™ (ASA carriage
control characters).

System Action: The program terminates.

Programmer Response: If the OUTFIL data set is
VSAM, it cannot be used for a report. If the OUTFIL
data set is non-VSAM, either let DFSORT set the
RECFM appropriately or specify a RECFM that includes
â€™Aâ€™.

jzhardy · Posted: Fri Aug 10, 2007 9:01 am

.. but not quite ...

the problem is that it puts the '-1' on all outputs. This is something that I can live with, but was not exactly what i wanted ..

Frank Yaeger · Posted: Fri Aug 10, 2007 7:56 pm

Frank Yaeger · Posted: Fri Aug 10, 2007 8:47 pm

Still looking at this, but I noticed you have a blank before REMOVECC in:

Frank Yaeger · Posted: Fri Aug 10, 2007 9:42 pm

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