IBM Mainframe Forum

by **saurabh pune** » Mon Dec 19, 2011 7:35 pm

Hi,
Needed assistance for Sorting a file based on 3 fields and then removing duplicates on the same file with 2 of the three sorted fields in a single sort step , Is that possible ?

I will further explain it as :

Input as :

Select all

NAAA01AA
NAAA02AA
NAAA03AA
NAAB01AA
NAAB01AB
NAAB03AA

1st Step Sort curently using :

Select all

SORT FIELDS=(1,4,CH,A,5,2,BI,A,7,2,CH,A)

2nd Step Sort currently using :

Select all

SORT FIELDS=(1,4,CH,A,5,2,BI,A),
EQUALS
SUM FIELDS=NONE

Output is :

Select all

NAAA01AA
NAAA02AA
NAAA03AA
NAAB01AA
NAAB03AA

Record eliminated is

Select all

NAAB01AB

as the first 6 bytes were same hence considered a duplicate and removed.

I wish to implement this into a single step, Is that possible.

Thanks a Ton,
Saurabh

by **BillyBoyo** » Tue Dec 20, 2011 5:15 am

I don't think you can do it in one step with SORT, but you might get what you want with SyncTool's SELECT.

Select all

//SYMNAMES DD *
FIRST-SIX-OF-KEY,1,6,CH
FULL-KEY,1,8,CH
//TOOLIN DD *
SELECT FROM(IN) TO(OUT) ON(FIRST-SIX-OF-KEY) FIRST USING(CTL1)
//CTL1CNTL DD *
SORT FIELDS=(FULL-KEY,A)

I have borrowed the above syntax from a DFSORT solution, I dont' have any documentation for SyncSort/SyncTool but if not exact you should be able to find something similar.

I have put your key together, rather than specifying the fields seperately. I think if you look at any recommendations in the SyncSort manul that it would say if they keys are contiguous it is better to specify as one key.

I like the SYMNAMES for flex and self-documentation.

I have no way to try this out. Let us know if it is any use.

by **xknight** » Tue Dec 20, 2011 2:45 pm

Hello,

As bill have pointed out, you can make use of Synctool version if your site has installed some latest PTF level.

Try the below snippet,

Select all

//STEP01 EXEC PGM=ICETOOL
//IN1 DD *
NAAA01AA
NAAA02AA
NAAA03AA
NAAB01AA
NAAB01AB
NAAB03AA
//OUT1 DD DSN=&TMP1,
// RECFM=FB,LRECL=80,BLKSIZE=0,
// DISP=(MOD,PASS,DELETE),
// SPACE=(CYL,(10,10),RLSE),UNIT=SYSDA
//OUT2 DD SYSOUT=*
//TOOLMSG DD SYSOUT=*
//DFSMSG DD SYSOUT=*
//TOOLIN DD *
SELECT FROM(IN1) TO(OUT1) ON(1,4,CH) ON(5,2,BI) FIRST
COPY FROM(OUT1) TO(OUT2) USING(CTL1)
//CTL1CNTL DD *
SORT FIELDS=(1,4,CH,A,5,2,BI,A,7,2,CH,A)
/*

Thanks,
Xavier

by **saurabh pune** » Tue Dec 20, 2011 3:05 pm

Thanks Billyboyo and Xavier for your quick help.
Will positively reply to this by tomorrow (On leave today).

Thanks again.

by **Alissa Margulies** » Wed Dec 21, 2011 12:41 am

Hello Saurabh.

Here is a 1-step Syncsort job that should produce the desired output:

Select all

//SORT1 EXEC PGM=SORT
//SORTIN DD *
NAAA01AA
NAAA02AA
NAAA03AA
NAAB01AA
NAAB01AB
NAAB03AA
//SORTOUT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSIN DD *
SORT FIELDS=(1,8,CH,A),EQUALS
OUTREC IFTHEN=(WHEN=INIT,OVERLAY=(81:SEQNUM,1,ZD,RESTART=(1,6)))
OUTFIL INCLUDE=(81,1,ZD,EQ,1),BUILD=(1,80)
/*

Please let us know if you have any further questions or if this does not produce the desired output.

Thank you.

by **saurabh pune** » Wed Dec 21, 2011 5:20 pm

Thanks a Lot Alissa,Xavier and BillyBoyo

It just worked perfectly fine.
Thanks Again for your Help.

IBM Mainframe Forum

SyncSort: Sorting and Removing Duplicates

SyncSort: Sorting and Removing Duplicates

Re: SyncSort: Sorting and Removing Duplicates

Re: SyncSort: Sorting and Removing Duplicates

Re: SyncSort: Sorting and Removing Duplicates

Re: SyncSort: Sorting and Removing Duplicates

Re: SyncSort: Sorting and Removing Duplicates