February 12, 2010 at 4:00 pm
I'm using SSAS 2008. I have a fact table that could contain NULL values that are associated with a dimension. Instead of ignoring dimension errors, I set the Unknown Member attribute in the dimension to visible. My understanding is that after setting that attribute, during cube processing, if a null is encountered no error is thrown, but rather that fact is thrown in the unknown bucket for that dimension.
I am still getting the dimension error :
Errors in the OLAP storage engine: The attribute key was converted to an unknown member because the attribute key was not found. Attribute Branch Id of Dimension: Lending Branch from Database: Rapid2, Cube: Rapid2Cube, Measure Group: Request Queue Trans Query, Partition: Request Queue Trans Query, Record: 1.
Am I missing something?
Thanks for the help, in advance 🙂
February 23, 2010 at 4:19 am
I think the problem lies in the fact that when the cube is processing your fact table, he will link the surrogate keys (or attribute key, whatever you use) to your dimension. If the SK is null he can not map the relationship.
For example:
FACTS (SK_Customer, amount, date_transaction)
1 500 2010/02/23
null 600 2010/02/23
2 1000 2010/02/22
...
CUSTOMER DIMENSION (SK_Customer, customer_ID, first_name, last_name)
1 customer1 firstname1 lastname2
2 customer2 firstname2 lastname2
...
In this case, the surrogate key of the second fact is missing. During processing, the cube can not find the SK with the value 'null' in your dimension, causing an error. Therefore you should map the null values in your fact table to surrogate key -1 and insert the following line in your dimension (following the previous example):
-1 N/A NOT APPLICABLE NOT APPLICABLE (N/A for codes, NOT APPLICABLE for text items)
Now the cube will find the NOT APPLICABLE member instead of nothing and therefore there is no error.
Anologue you can use a SK with value 0 for UNKNOWN members (and thus creating your own unknown member).
You can do this mapping in the ETL fase or in the DSV.
Need an answer? No, you need a question
My blog at https://sqlkover.com.
MCSE Business Intelligence - Microsoft Data Platform MVP
September 25, 2011 at 10:48 pm
Perfect explanation by the Koen. Just to add my two cents...
Make an effort to assign some default values to such dimension entities.
Raunak J
Viewing 3 posts - 1 through 2 (of 2 total)
You must be logged in to reply to this topic. Login to reply