IBM Mainframe Forum

I cant figure why my duplicates are elimated in the following. I want to create a count of duplicate records on key 'ACCOUNT' (which works ok) and display all records including duplicates. I'm only getting one record for each key. Please someone...
38:11,10, CLIENT NUM
48:126,15, CLIENT NAM
63:43,9, FINANCL INST
72:71,10, AMOUNT
82:X'0000001C')
SUM FIELDS=(82,4,PD)

OUTREC FIELDS=(01:01,18, CUST ACC NUM
21:19,1, CLAIM OFFC
24:20,5, CONTRACT NUM
31:25,10, CERT
43:35,3, SERVICE LINE
48:38,10, CLIENT NUM
60:48,15, CLIENT NAM
77:63,9, FINANCL INST
88:72,10, AMOUNT
100:82,4,PD,EDZ9=(Z,ZZZ,ZZZ))
SORT FIELDS=(01,18,CH,A) POSITION AFT INREC

//SYSOUT DD SYSOUT=*

cant even c&p today !!

//SORT EXEC PGM=SORT
//* PARM='CORE=120000'
//*
//SORTLIB DD DSN=SYS1.SORTLIB,DISP=SHR
//SORTIN DD DSN=ABC.DEF.DAILY.TRANSACT.MARCH,
// DISP=SHR
//*
//SORTOUT DD DSN=ABC.DEF.DAILY.REPORTS.RPT(+1),
// DISP=(NEW,CATLG,DELETE),
// UNIT=TEST,SPACE=(CYL,(1,2),RLSE),
// DCB=(LRECL=108,BLKSIZE=0,BUFNO=30,RECFM=FB)
//*
//SORTWK01 DD UNIT=TEST,
// SPACE=(CYL,(20,10),RLSE),
// DCB=BUFNO=5
//*
//*
//SYSIN DD *
INREC FIELDS=(01:52,18, ACCOUNT
19:22,1,
20:23,5,
25:28,10,
35:38,3,
38:11,10,
48:126,15,
63:43,9,
72:71,10,
82:X'0000001C')
SUM FIELDS=(82,4,PD)
OUTREC FIELDS=(01:01,18,
21:19,1,
24:20,5,
31:25,10,
43:35,3,
48:38,10,
60:48,15,
77:63,9,
88:72,10,
100:82,4,PD,EDZ9=(Z,ZZZ,ZZZ))
SORT FIELDS=(01,18,CH,A)

//SYSOUT DD SYSOUT=*

Still havent found out why duplicates are not displayed....anyone?

The SUM statement deletes records with equal control fields and optionally sums specified numeric fields on those records. If numeric fields are to be summed, the data in the summed fields are added, the sum is placed in one of the records, and the other record is deleted. Provided arithmetic overflow does not occur, the SUM control statement produces only one record per sort key in the output data set.

Instead of coding SUM, you might want to consider coding OUTFIL SECTIONS with a TRAILER3 COUNT or a SEQNUM with RESTART.

thank you for the reply. I'm not very familiar with syncsort, I'm just trying to count the occurances records with the same key and display each record, - is there an easy way to do this?

If you show some sample input records, I can provide you with the control statements.

hi, thank you!!
my jcl is uses a dd statement referencing a sequential file, sample recs like the following:
AAAB 123 lksa ABCD123
AAAC 123 lksa EFGH456
AAAD 123 lksa ABCD123
BBBE 456 kkkl EFGH456
BBBF 456 kkkl ABCD123
CCCG 789 LKDS IJKL789

Key would be 4th field, so sorted, counted file would be like:

Account Nbr Model Serial Occurs
AAAB 123 lksa ABCD123 3
AAAD 123 lksa ABCD123 3
BBBF 456 kkkl ABCD123 3
AAAC 123 lksa EFGH456 2
BBBE 456 kkkl EFGH456 2
CCCG 789 LKDS IJKL789 1

Here is a SyncSort for z/OS 1.3.2 job that will produce the count at the end of every record:

This is the output:

If you want the output in the same order as specified in your last post, then modify the SORT statement in the first step (SORT1) as follows:

If the count is not required on each record, here is an alternate solution that only passes the data once:

This is what the output looks like:

BLESS YOU!
I have understood the example and have applied it to my jcl, its working and I will be able to learn from it!!
Thank you very much!