IBM MAINFRAME HELP & SUPPORT FORUMS

vasco ferreira · Joined: 12 Jul 2008 Posts: 3 Location: sintra - portugal

Hi. I am running a copy of 6634978 records with idcams. It takes about 8 minutes.
Only have the input (disp=shr,sequential dsn) the output (disp=old,vsam dsn) and finally the usual repro/replace.
I think is possible less time of execution, could you help me find a best solution?

dick scherrer · Joined: 23 Nov 2006 Posts: 8733 Location: 221 B Baker St

Hello Vasco and welcome to the forums,

Adding more than 6.5 million records to a vsam file in about 8 minutes does not seem like a lot of time. . .

You might try to accomplish the copy using your sort product rather than idcams.

vasco ferreira · Joined: 12 Jul 2008 Posts: 3 Location: sintra - portugal

Thanks for your answer. I already have a previous sort step. Will you show me what kind of sort product are you suggesting? Instead of idcams should use other thing?

vicky10001 · Joined: 13 Jul 2005 Posts: 109

You can use DFSORT(IBM product) OR Syncsort

expat · Posted: Mon Aug 25, 2008 7:34 pm Post subject:

You can sort out directly to the VSAM cluster.

dick scherrer · Joined: 23 Nov 2006 Posts: 8733 Location: 221 B Baker St

Hello,

Quote: I already have a previous sort step. If you are already sorting the data, pointing the sorted output to the vsam file rather than a qsam file should save 2 full passes of the data (one to write it out of the sort and one to read it back into the idcams).

If you write the vsam directly out of the sort, it should save considerable time.

vasco ferreira · Joined: 12 Jul 2008 Posts: 3 Location: sintra - portugal

thanks to you all. I'll try Dfsort and Syncsort
In the mean time i show what exactly i'm doing.

//* ___________________________________________________________________

//* ORDENA LOGAIX01D (NO JOB ORIGINAL LEVA 1 SEGUNDO)

//* ___________________________________________________________________

//SORTX01D EXEC PROC=SORT

//SYSOUT DD SYSOUT=*

//SORTIN DD DSN=PRDCME.LGNIX01D.SEPA.VCF,DISP=SHR

//SORTOUT DD DSN=PRDGR1.TEMP.VCF,DISP=(,PASS),

// LRECL=28,RECFM=FB,DSORG=PS,

// SPACE=(CYL,(50,50),RLSE)

//SYSIN DD *

SORT FIELDS=(1,28,A),FORMAT=BI

SUM FIELDS=NONE

//*

//*

//* ___________________________________________________________________

//* RPAIX01D (NO JOB ORIGINAL LEVA 8 MINUTOS E 46 SEGUNDOS)

//* ___________________________________________________________________

//RPAIX01D EXEC PGM=IDCAMS

//SYSPRINT DD SYSOUT=*

//FINX01D DD DSN=PRDGR1.TEMP.VCF,DISP=(OLD,DELETE)

//FOUTX01D DD DSN=PRDQUE01.LNAIX01D.VCF.CL,

// AMP=('BUFNI=010,BUFND=010'),DISP=SHR

//SYSOUT DD SYSOUT=*

//SYSIN DD *

REPRO INFILE(FINX01D) OUTFILE(FOUTX01D) REPLACE

//*

dick scherrer · Joined: 23 Nov 2006 Posts: 8733 Location: 221 B Baker St

Hello,

As i mentioned before, i believe you will run faster if you write the output from the sort directly to the vsam file rather than the intermediate write to the qsam file.

Good luck - we are interested to hear your results :)

d