Comparing very large datasets with about 20 million records

morpheus007 · New User Joined: 27 Dec 2005 Posts: 58

We have a requirement to compare 2 very large flat files with about 20 million records in each.If I use the 3.13 TSO standard compare option would there be any problems.Is 3.13 equipped to handle large files.
Kindly let me know.

guptae · Posted: Fri Mar 16, 2007 10:00 am

Hi There,

You can run dfsort for this.

murmohk1 · Posted: Fri Mar 16, 2007 10:02 am

Or use filemaster/fileaid compare option (in batch mode).

morpheus007 · New User Joined: 27 Dec 2005 Posts: 58

Thanks for both the responses.

I do see a compare option in our File handling tool.I will try that out.

With regards to the SORT we use Syncsort.Kindly let me know if there is any Sort card for Syncsort that could be used in this case.

murmohk1 · Posted: Fri Mar 16, 2007 10:22 am

You have a similar post at the link -

ibmmainframes.com/viewtopic.php?t=138&highlight=trick

guptae · Posted: Fri Mar 16, 2007 10:25 am

Hi There,

Please check out this post to find match & unmatch record using syncsort

ibmmainframes.com/viewtopic.php?t=15955&highlight=

raak · Posted: Fri Mar 16, 2007 3:10 pm

hey

If u have COMPAREX in ur shop, u can use that also to compare

Tru this...

//COMPARE EXEC PGM=COMPAREX
//SYSPRINT DD SYSOUT=Q
//OUTPUTDD OUTPUT PRMODE=PAGE,PAGEDEF=ONEUPO,FORMDEF=ONEUPO
//SYSUT1 DD DSN=TEST.PS.FILE1,
// DISP=SHR
//SYSUT2 DD DSN=TEST.PS.FILE2,
// DISP=SHR
//SYSIN DD *
CONTINUE
FORMAT=16
DASH=C'-'
CASE=MIXED
/*

Phrzby Phil · Posted: Fri Mar 16, 2007 5:22 pm

3.13 has a batch option.