Skip to Main Content

Analytics Software


For appeals, questions and feedback, please email

Handling metadata errors in ODI/Planning

CLSep 16 2009 — edited Sep 27 2009
I debated between putting this question onto the ODI forum or here but as the world-famous John Goodwin so dominates this board, I thought I’d post it here. :)

How are people handling Planning dimension builds in ODI wrt error handling within packages?

My goal is to only send exception emails when a problem occurs. I am automating this via packages compiled to scenarios.

I have found that a genuine error in an interface will cause a trappable error in a package, e.g., the server name is really and somehow the physical schema gets changed to I can then send a failure email off.

I have also found that if I have a metadata error in a Planning dim build interface I can trap the errors to an error file. A bad record (syntactically correct, but its parent can’t be found, for instance) will generate an entry in the error file if I turn error logging on.

However, that bad record does not generate an ODI error in a package.

I thought about testing for the existence of this error file but the constant generation (from the LOG_ERRORS option in the IKM SQL to Hyperion Planning) of the error file regardless of bad record condition means that I can’t use the ODIFileWait control to test if the file is there or not.

I cannot find a way to test for the length of the file (if it is zero, there is no metadata error) within ODI.

I guess I could do it through an external script (Windows environment) but that is so wrong on so many levels that I can’t believe that is my only option.

Any thoughts? If you come up with an answer I’ll post it in my blog and give you all the credit. As if that were a real incentive…


Cameron Lackpour
This post has been answered by JohnGoodwin on Sep 27 2009
Jump to Answer
Locked Post
New comments cannot be posted to this locked post.
Post Details