Page 1 of 1

Posted: Mon Jan 11, 2010 12:02 am
by chulett
Since the summations that you are doing are in your source SQL and not something you're doing in the job itself (say, in an Aggregator) then it seems to me that this is strictly an Oracle materialized view issue.

Posted: Mon Jan 11, 2010 12:24 am
by CoolTiranga
Thanks for your input, Craig!

I tried a few leads in that respect, like,

1. The job and the snapshot running at the sametime.
2. For a few records in the target table for which the values were computed incorrectly, I've noticed that the incorrect value is the sum of the first few records.

The snapshot is a manual fast refresh cycled at every 15 mins. But, the records of impact have no updates and have been the same since the insert which is about a few months ago.

I am still researching the issue and hopefully find someting. :-)

Posted: Mon Jan 11, 2010 7:30 am
by chulett
Have you involved your DBA? I sure as heck would be banging on their door if I was in your shoes. :wink:

Posted: Mon Jan 18, 2010 4:32 pm
by CoolTiranga
Thanks Craig! Cannot thank the DSXchangers enough for the valuable contribution of time in sharing the knowledge.

With my issue, luckily with the Datasets I saved before and after the correct run we've realized that the sql query is returing duplicates by using the differece stage of comparing the two. After searching for any known bugs in Oracle, I stumbled upon this one post. I hope I know the real issue as to why there are duplicates soon! I can say that there is no data reason why a duplicate should occur and not the immediate next minute.

http://dbaspot.com/forums/oracle-server ... p-bug.html

Thank you!

Posted: Mon Jan 18, 2010 4:40 pm
by chulett
Yup, from your linked post it sounds like it is indeed an (gasp!) Oracle bug. At least there's a patch for it. Good luck.