IBM Mainframe Forum

by **troubledguy** » Tue Jul 28, 2015 1:22 pm

hi all,
i am new to the forum. i appolozise if i posted a previously asked question or in wrong forum.

i have 2 input files . according to the requiremnt i need only those records present in both files matching on col <5-9> and having col <11-15> as "xxxxx" and col<19-20> as "ee" in file 1 and col<19-20> as "ee" in file 2.

file IN1:

Select all

----+----1----+----2----+----3
111 AAAAA XXXXX EE
111 BBBBB YYYYY TT
111 CCCCC XXXXX EE
111 DDDDD XXXXX EE
111 EEEEE ZZZZZ RR

file IN2:

Select all

----+----1----+----2----+----3
111 AAAAA 11111 EE
111 BBBBB 22222 TT
111 SSSSS 33333 EE
111 CCCCC 44444 EE
111 FFFFF 55555 RR

file OUT1: matched record

Select all

----+----1----+----2----+----3
111 AAAAA 11111 EEA
111 CCCCC 44444 EEA

file OUT2: unmatched from file 1

Select all

----+----1----+----2----+----3
111 DDDDD XXXXX EE

below is the icetool jcl i have written to serve the above purpose.

Select all

//TOOLIN DD *
COPY FROM(IN1) TO(OUT1) USING(CTL1)
/*
//CTL1CNTL DD *
JOINKEYS F1=IN1,FIELDS=(5,5,A),
INCLUDE=(19,2,CH,EQ,C'EA',AND,11,5,CH,EQ,C'xxxxx')
JOINKEYS F2=IN2,FIELDS=(5,5,A),
INCLUDE=(19,2,CH,EQ,C'EA')
REFORMAT FIELDS=(F1:1,10,F2:11,10)
OUTREC BUILD=(1,20,X'C1')
/*

now my new requirement is to get the unmached records from file 1 which has col <11-15> as "xxxxx" and col<19-20> as "ee" in a different output file.
is there any way to modify the jcl to get the unmatched records as well?

by **NicC** » Tue Jul 28, 2015 2:48 pm

In future please use the code tags when presenting code and data. By doing so a fixed font is used and spaces are not trimmed. Yoou can see the code tags button in the editor toolbar when yu use the full editor which is more easily entered by using the Postreply button instead of the QuickReply button.

by **BillyBoyo** » Tue Jul 28, 2015 4:02 pm

Yes,

First, is there a particular reason to use ICETOOL here? JOINKEYS is part of SORT.

Although you can put the INCLUDE/OMIT on the JOINKEYS, the recommended way is to use JNFnCTNL files.

Unless needed, I use INREC in preference to OUTREC. If you are appending something to a record, it will be clearer to use OVERLAY than BUILD.

Select all

//SYSIN DD *
JOINKEYS F1=IN1,FIELDS=(5,5,A)
JOINKEYS F2=IN2,FIELDS=(5,5,A)
REFORMAT FIELDS=(F1:1,10,F2:11,10)
INREC OVERLAY=(21:X'C1')
//JNF1CNTL DD *
INCLUDE=(19,2,CH,EQ,C'EA',
AND,
11,5,CH,EQ,C'xxxxx')
//JNF1CNTL DD *
INCLUDE=(19,2,CH,EQ,C'EA')

Then you'll need a JOIN statement. A JOIN is not needed if you only want matches, but you'll need JOIN UNPAIRED,F1.

In the REFORMAT statement, include the "match marker", a ?, which in your case, due to the JOIN, will be B (on both files, ie matched) or 1 (on File 1 only, unmatched).

Then you'll need two OUTFILs, one for each output file. One should INCLUDE= the Bs and the other should use SAVE.

by **troubledguy** » Tue Jul 28, 2015 4:17 pm

thanks BillyBoyo,

will implement your logic and i hope that works out for me.

the reason i used icetool was because i had 2 pair of files to compare and i wanted to do in a single step. that all.

IBM Mainframe Forum

how to get get unmatched record to a file using icetool?

how to get get unmatched record to a file using icetool?

Re: how to get get unmatched record to a file using icetool?

Re: how to get get unmatched record to a file using icetool?

Re: how to get get unmatched record to a file using icetool?