IBM Mainframe Forum

by **crescencio** » Fri Jun 12, 2009 4:25 pm

Hi,

I need some help two compare two datasets (Day 1 / Day 2).
The files are FB and 10300 bytes record length.

The records can be matched by key but I need to extract only those that are in Day 2 but not in Day 1 file and those that have changed in some part of the record.

I have found some ideas in this forum, but the only way that I have found to expand the comparision to the whole record is to declare th JOINKEYS to 10300.
It works with test files of 30 bytes but not with real files.
I receive a message like this "WER271A JOINKEYS STATEMENT: NUMERIC FIELD ERROR2".

If you can help with this or can suggest me a different tool that fit better with the request I will appreciate.

Thanks a lot for the help.

by **Alissa Margulies** » Fri Jun 12, 2009 8:16 pm

Please post the JCL you submitted when you received that error. Thank you.

by **crescencio** » Tue Jun 16, 2009 3:42 pm

Hi,

First I run a JCL like this to test if the idea could work. Considering the first five numeric characters the real key.

Select all

000005 //STEP1 EXEC PGM=SORT
000006 //SORTJNF1 DD *
000007 12345AAAAABBBBBB
000008 67890CCCCCAAAAAA
000009 13579DDDDDEEEEEE
000010 //SORTJNF2 DD *
000011 12345CCCCCBBBBBB
000012 13579DDDDDEEEEEE
000013 08531DDDDDEEEEEE
000014 //SORTOF01 DD DSN=W#H00I.PRUEBA.TRES,
000015 // DISP=(NEW,CATLG),
000016 // UNIT=SYSDA,SPACE=(TRK,(100,100))
000017 //SYSOUT DD SYSOUT=*
000018 //SYSIN DD *
000019 JOINKEYS FILES=F1,FIELDS=(1,16,A)
000020 JOINKEYS FILES=F2,FIELDS=(1,16,A)
000021 REFORMAT FIELDS=(F2:1,16,F1:1,16),FILL=X'FF'
000022 JOIN UNPAIRED,F2
000023 SORT FIELDS=COPY
000024 OUTFIL FILES=01,INCLUDE=(1,1,BI,NE,X'FF',AND,17,1,BI,EQ,X'FF'),
000025 OUTREC=(1,16),FTOV,VLTRIM=X'FF'

This JCL produce in dataset PRUEBA.TRES the following output

Select all

DSLIST W#H00I.PRUEBA.TRES
Command ===>
*****************************
08531DDDDDEEEEEE
12345CCCCCBBBBBB
*****************************

These two records are Ok because the first one is not in Day 1 file and the second one exits in Day 1 but with different information.

Then I tried with a real JCL like this one.

Select all

//STEP1 EXEC PGM=SORT
//SORTJNF1 DD DISP=SHR,DSN=DAY.ONE
//SORTJNF2 DD DISP=SHR,DSN=DAY.TWO
//SORTOF01 DD DSN=W#H00I.PRUEBA.CUATRO,
// DISP=(NEW,CATLG),
// UNIT=SYSDA,SPACE=(TRK,(100,100))
//SYSOUT DD SYSOUT=*
//SYSIN DD *
JOINKEYS FILES=F1,FIELDS=(1,10300,A)
JOINKEYS FILES=F2,FIELDS=(1,10300,A)
REFORMAT FIELDS=(F2:1,10300,F1:1,10300),FILL=X'FF'
JOIN UNPAIRED,F2
SORT FIELDS=COPY
OUTFIL FILES=01,INCLUDE=(1,1,BI,NE,X'FF',AND,10301,1,CH,NE,X'FF'),
OUTREC=(1,10300),FTOV,VLTRIM=X'FF'

And I got the error

Select all

JOINKEYS FILES=F1,FIELDS=(1,10300,A)
*
JOINKEYS FILES=F2,FIELDS=(1,10300,A)
*
REFORMAT FIELDS=(F2:1,10300,F1:1,10300),FILL=X'FF'
JOIN UNPAIRED,F2
SORT FIELDS=COPY
OUTFIL FILES=01,INCLUDE=(1,1,BI,NE,X'FF',AND,10301,1,CH,NE,X'FF'),
OUTREC=(1,10300),FTOV,VLTRIM=X'FF'
WER271A JOINKEYS STATEMENT: NUMERIC FIELD ERROR
WER271A JOINKEYS STATEMENT: NUMERIC FIELD ERROR

Thanks a lot for your interest.

by **Alissa Margulies** » Tue Jun 16, 2009 9:18 pm

Try this SyncSort for z/OS job:

Select all

//STEP1 EXEC PGM=SORT
//SYSOUT DD SYSOUT=*
//SORTJNF1 DD DSN=DAY.ONE,...
//SORTJNF1 DD DSN=DAY.TWO,...
//SORTOUT DD DSN=NEW.CHANGED.RECORDS,...
//SYSIN DD *
JOINKEYS FILES=F1,FIELDS=(1,5,A)
JOINKEYS FILES=F2,FIELDS=(1,5,A)
JOIN UNPAIRED,F2
REFORMAT FIELDS=(F2:1,10300,F1:1,10300)
SORT FIELDS=COPY
OUTFIL INCLUDE=(1,256,CH,NE,10301,256,CH,OR,
257,256,CH,NE,10557,256,CH,OR,
513,256,CH,NE,10813,256,CH,OR,
769,256,CH,NE,11069,256,CH,OR,
1025,256,CH,NE,11325,256,CH,OR,
1281,256,CH,NE,11581,256,CH,OR,
1537,256,CH,NE,11837,256,CH,OR,
1793,256,CH,NE,12093,256,CH,OR,
2049,256,CH,NE,12349,256,CH,OR,
2305,256,CH,NE,12605,256,CH,OR,
2561,256,CH,NE,12861,256,CH,OR,
2817,256,CH,NE,13117,256,CH,OR,
3073,256,CH,NE,13373,256,CH,OR,
3329,256,CH,NE,13629,256,CH,OR,
3585,256,CH,NE,13885,256,CH,OR,
3841,256,CH,NE,14141,256,CH,OR,
4097,256,CH,NE,14397,256,CH,OR,
4353,256,CH,NE,14653,256,CH,OR,
4609,256,CH,NE,14909,256,CH,OR,
4865,256,CH,NE,15165,256,CH,OR,
5121,256,CH,NE,15421,256,CH,OR,
5377,256,CH,NE,15677,256,CH,OR,
5633,256,CH,NE,15933,256,CH,OR,
5889,256,CH,NE,16189,256,CH,OR,
6145,256,CH,NE,16445,256,CH,OR,
6401,256,CH,NE,16701,256,CH,OR,
6657,256,CH,NE,16957,256,CH,OR,
6913,256,CH,NE,17213,256,CH,OR,
7169,256,CH,NE,17469,256,CH,OR,
7425,256,CH,NE,17725,256,CH,OR,
7681,256,CH,NE,17981,256,CH,OR,
7937,256,CH,NE,18237,256,CH,OR,
8193,256,CH,NE,18493,256,CH,OR,
8449,256,CH,NE,18749,256,CH,OR,
8705,256,CH,NE,19005,256,CH,OR,
8961,256,CH,NE,19261,256,CH,OR,
9217,256,CH,NE,19517,256,CH,OR,
9473,256,CH,NE,19773,256,CH,OR,
9729,256,CH,NE,20029,256,CH,OR,
9985,256,CH,NE,20285,256,CH,OR,
10241,59,CH,NE,20541,59,CH),
BUILD=(1,10300)
/*

by **crescencio** » Wed Jun 17, 2009 1:53 pm

yes, it works perfectly.
It was a very good help.

Thanks a lot for your time Alissa.

IBM Mainframe Forum

Compare two datasets. Compare the whole record

Compare two datasets. Compare the whole record

Re: Compare two datasets. Compare the whole record

Re: Compare two datasets. Compare the whole record

Re: Compare two datasets. Compare the whole record

Re: Compare two datasets. Compare the whole record