May 12, 2015 at 2:50 pm
I ran into an issue this morning processing cubes. I processed the entire project (all cubes) and got a failure each time related to an attribute key problem. After looking through the data, which was fine, I processed all of the dimensions first and then processed cubes and the error was gone. I was under the assumptions that processing the entire project would process dimensions first but I guess I was wrong. Is there an option, setting, etc. that will make it so that dimensions are always processed before cubes?
May 12, 2015 at 5:26 pm
When you say "entire project", does that mean you processed the database? Or did you individually process each cube within the database? What was the process setting (i.e. ProcessFull, ProcessDefault, ProcessUpdate)?
May 12, 2015 at 5:30 pm
Using SSDT - Analysis Services project - Right clicked the entire project which contains all cubes and selected process. So I believe it would be default as I made no changes.
May 12, 2015 at 5:34 pm
JoshDBGuy (5/12/2015)
Using SSDT - Analysis Services project - Right clicked the entire project which contains all cubes and selected process. So I believe it would be default as I made no changes.
If you didn't change anything, it would most likely have used the "ProcessDefault" setting which does not reprocess dimensions...unless there was a structural change.
The following blog post describes all the processing options pretty well: http://blogs.msdn.com/b/karang/archive/2011/01/25/kind-of-ssas-processing-in-simple-words.aspx
May 15, 2015 at 7:35 am
when you run the project, it first processes Dimensions, then Cube.
what is the data source you are using. If its other than SQL Server, you will be getting this kind of errors.
And also there might be some problem with data also. Check whether your column have nulls.
Are you directly importing tables or using Named Queries. better to use Named Query and get the data what you need. I did get this issue when I work with my Oracle source tables. I used Named Query and take the data only I need like
select column1, column2...columnN from fact_table
where date_pk_id > 20091231
select column1, column2,...columnN from dim_table
where dim_pk_id in (select dim_fk_id from fact_table where date_pk_id > 20091231)
If you want to control the processing of SSAS Cube then you need to use SSIS package.
you have a Analysis Services Processing Task. you can connect to cube database, then select the required dimensions then put another Analysis Services Processing Task, select only Cubes there.
Hope this helps.
Thank You
May 15, 2015 at 9:23 am
VRT (5/15/2015)
what is the data source you are using. If its other than SQL Server, you will be getting this kind of errors.
Not true...the issue experienced by the OP has nothing to do with the type of data source. Missing or unmatched attribute key errors mean that a fact table contains a key that could not be found in the dimension. The error has more to do with what is processed and how, and not the type of data source.
May 15, 2015 at 12:35 pm
Martin Schoombee (5/15/2015)
VRT (5/15/2015)
what is the data source you are using. If its other than SQL Server, you will be getting this kind of errors.Not true...the issue experienced by the OP has nothing to do with the type of data source. Missing or unmatched attribute key errors mean that a fact table contains a key that could not be found in the dimension. The error has more to do with what is processed and how, and not the type of data source.
You are right. But Key errors meaning some times we also get duplicate key errors. I dont see any thing wrong with my key columns. But it is also hard to understand with huge tables. yes I do work on more than 50 million records.
I also get some time duplicate key error for non-key columns...I use queries instead of tables.
I thought in that way, If its not really make sense I can delete that post.
Thank You
Viewing 7 posts - 1 through 6 (of 6 total)
You must be logged in to reply to this topic. Login to reply