Page 1 of 1

Merge Stage -Differnence in row count

Posted: Wed Jul 07, 2010 1:18 am
by jyothisdasms
Hi,

I have migrated one server job which contain merge stage from 7.5 to 8.1. Now when i am trying to run the job ,i am getting more number of rows in 8.1.When i checked the target file , some blank rows are present at the end of the file which are not in 7.5.

The size of the two input files in merge stage is same in both 7.5 and 8.1.

Please help me to solve this .

Thanks in Advance

Posted: Wed Jul 07, 2010 1:20 am
by ArndW
Could you FTP the input file from one machine to the other and do a "diff" command - perhaps the size is the same but the contents different. If you can't copy the file over, perhaps you have "md5sum" installed and can compare hashes.

Posted: Wed Jul 07, 2010 9:31 am
by jyothisdasms
I checked with diff command .Thers is no difference in the datas .

One more thing I noticed is,If i am using left outer join in 8.1 i am getting the same number of rows as in 7.5.But if we use complete set option differnce in rows will come.

Thanks in advance

Posted: Thu Jul 08, 2010 7:17 am
by ArndW
There is no easy answer. I would pare down the data to a few rows which demonstrate the problem, i.e. which deliver different results, and then see if there is some sort of explanation.

Posted: Tue Aug 03, 2010 2:08 pm
by arunkumarmm
We had similar kinda problem when migrating from 7.5.2 to 8.1

But a little different issue. 7.5 gave more rows than 8.1

As per IBM's suggestion, we set the NLS of each column in the merge stage to be UTF-8 and it worked fine! :D

But I really dont understand about it yet, because our NLS project defult is UTF-8 :x