View previous topic :: View next topic
|
Author |
Message |
swaroopnaga
New User
Joined: 15 Oct 2005 Posts: 8 Location: US
|
|
|
|
Hi,
I am trying to compare two files of 20 million rows using the batch ISRSUPC, the summary is shown below:
?104345 NUMBER OF LINE MATCHES 10000 TOTAL CHANGES (PAIRED+NONPAIRED
0 REFORMATTED LINES 10000 PAIRED CHANGES (REFM+PAIRED INS
10000 NEW FILE LINE INSERTIONS 0 NON-PAIRED INSERTS
10000 OLD FILE LINE DELETIONS 0 NON-PAIRED DELETES
?114850 NEW FILE LINES PROCESSED
?114850 OLD FILE LINES PROCESSED
Can anyone provide some input as what "?" stands for? |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
most probably some kind of <number> overflow |
|
Back to top |
|
|
Bill Woodger
Moderator Emeritus
Joined: 09 Mar 2011 Posts: 7309 Location: Inside the Matrix
|
|
|
|
How long did it take to run, out of interest?
The length of those number fields, including the ?, would be able to hold 9,999,999 - probably usually enough for a file comparison.
I agree with enrico. When the 10millionth arrived, the program cleverly noted that it couldn't be displayed, so when producing the totals it plopped the ? down to bring your attention to it. Probably somewhere deep in the manual it mentions file limits, and I guess you are lucky that this is the only apparent effect.
I'd not feel lucky. I'd split the files into three equal pieces and do the compare(s) again. I wouldn't want to "trust to luck" that half my file compared when I know "something" is "wrong" with the program for that half of the file. |
|
Back to top |
|
|
Escapa
Senior Member
Joined: 16 Feb 2007 Posts: 1399 Location: IL, USA
|
|
|
|
Or try with DFSORT compare and check how many new records and how many old records try to match the number with this number..
you will get to know what that ? means... |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
using isrsupc for <application> dataset comparison is a bit of overshooting
application data is usually sorted on some key, most probably a <SORT> approach might be faster |
|
Back to top |
|
|
swaroopnaga
New User
Joined: 15 Oct 2005 Posts: 8 Location: US
|
|
|
|
It took 16 hours for this comparison....
thanks for your inputs..
Naga |
|
Back to top |
|
|
Escapa
Senior Member
Joined: 16 Feb 2007 Posts: 1399 Location: IL, USA
|
|
|
|
swaroopnaga wrote: |
It took 16 hours for this comparison....
thanks for your inputs..
Naga |
Your job was running for 16 long hours
For doing just comparision?
You are lucky .. Your job is not thrown away by system management team.. |
|
Back to top |
|
|
swaroopnaga
New User
Joined: 15 Oct 2005 Posts: 8 Location: US
|
|
|
|
Thats the reason I don't want to run again without perfect solution |
|
Back to top |
|
|
Escapa
Senior Member
Joined: 16 Feb 2007 Posts: 1399 Location: IL, USA
|
|
|
|
Give your input file layouts....
and Tell us what is the sort product installed at your shop... |
|
Back to top |
|
|
swaroopnaga
New User
Joined: 15 Oct 2005 Posts: 8 Location: US
|
|
|
|
the file length is 250.I used SUPERC program. |
|
Back to top |
|
|
Escapa
Senior Member
Joined: 16 Feb 2007 Posts: 1399 Location: IL, USA
|
|
|
|
swaroopnaga wrote: |
the file length is 250.I used SUPERC program. |
Read my post again and answer what I asked (completely) |
|
Back to top |
|
|
|