mainframegurukul.com

Input
====
ABCD sdfsdf klsdfsdf
dfsdfs 12345 dfsdfdf
sdfsdf sdfsdf ******

BDCF dfsdffl dfsdfsdf
sdfsd 67893 dfgdfgd
asdas sdfsd $$$$$

Output
====
ABCD,12345,******
BDCF,67893,$$$$$

Hi All,

I am pretty new to DFsort and been working for just last week. I couldn't find the exact same query, have pasted the required results here.

This is ths sort card, but not giving the desired results


SORT FIELDS=(COPY)
OUTREC FIELDS=(1:5,35,38:30,13,55:35,17)

But i need to use this conditional outrec so that part from different rows is merged into songle row.

OUTREC has no special power. It is the same as INREC (assuming DFSORT), but operates after SORT/MERGE. Without SORT/MERGE, use INREC.

FIELDS is old and overloaded. Use BUILD instead. BUILD is a synonym for FIELDS, so operation is 100% identical, but it is not overloaded: when you see BUILD on INREC, OUTREC and OUTFIL it is less confusing than seeing FIELDS, FIELDS and OUTREC, and also seeing FIELDS on SORT/MERGE, REFORMAT, SUM, etc.

Those are tips. They don't affect what you want.

Assuming that you can guarantee groups of three records (or four if that blank is really there), it is WHEN=GROUP that you want.

You'll need to temporarily add a sequence number to the record. When it is exactly divisible by three (or four) , WHEN=GROUP to identify that and PUSH the data you want from the first record at to an extension of the record, with RECORDS=3.

When it is divisible by three (or four) with remainder one, WHEN=GROUP to identify that and PUSH the data you want from the second record to a separate extension.

When it is divisible with remainder two, use BUILD to format the output that you want from the three pieces of data that you have access to at the time: the third record; the data PUSHed from the first record; the data PUSHed from the second record.

Use OUTFIL INCLUDE=/OMIT= to select the third record of each group. If there is a blank record that you need, you can either INCLUDE=/OMIT= that as well, or generate it on a BUILD in OUTFIL using the slash operator (/).

Look for the DFSORT Getting Started manual, it is aimed exactly at people in your situation and has many useful examples. Full details of everything are in the DFSORT Application Programming Guide.

It is very interesting yet to forward the message.

Thanks William for useful insights and suggestions. I was side tracked untill i came across again same task last week.

I got around the situation using the ICETOOL. Again it was first attempt for me in ICETOOL but it's really useful. I used temp seq no. and ID field before creating Output. And the actual data pos i gave in WITH parameters. I needed a CSV output, so added a , after each block while creating outrec.

//TOOLIN DD *
SPLICE FROM(IN) TO(OUT) ON(120,2,ZD) KEEPNODUPS WITHANY -
WITH(68,20) WITH(89,17) USING(CTL1)
/*
//CTL1CNTL DD *
INREC IFOUTLEN=133,
IFTHEN=(WHEN=GROUP,RECORDS=3,PUSH=(120:ID=2,131:SEQ=1)),
IFTHEN=(WHEN=(131,1,ZD,EQ,2),BUILD=(68:30,20,120:120,2)),
IFTHEN=(WHEN=(131,1,ZD,EQ,3),BUILD=(89:35,17,120:120,2))
OUTFIL FNAMES=OUT,BUILD=(1,87,C',',89,17,C',',107,27)


Thanks,
Ashank