Forum Discussion
Import SSAS Tabular Database: Unexpected Exception
We've updated the SQL Server as follows:
SQL Server: Microsoft SQL Server 2016 (SP1-CU6) (KB4037354) - 13.0.4457.0 (X64) Nov 8 2017 17:32:23
SSAS: 13.0.4457.0
The Size of the source database in SQL Server is: 2.4 GB (MDF File), 4.6 GB (LDF File)
The Size of the database in SSAS Server (the Tabular Cube) is 283 MB. I get this by right clicking on the databse in SSAS then selecting Reports-->Standard Reports-->General
My computer has 8GB of RAM and at the time I try to Import the SSAS cube in Power BI Desktop I have about 4 GB of free RAM.
The error message I am getting now is below. Note that We are already using the 64 bit version of the product and the latest November 2017 update.
Anonymous,
Regarding to the memory settings in SSAS, please post the question in SSAS forum to get dedicated support. And you can refer to this article.
From Power BI side, try to increase maximum allowed memory, and uncheck the following options.
Regards,
Lydia
- Anonymous8 years agoNot applicable
Thanks for the feedback. We've attempted several memory settings but none resolved the issue.
However, after futher experiments we've made some discoveries. So our case was as follows: We have a source database that is modeled as a Star Schema. Furthermore, we've created a SSAS Tabular cube for this database and then tried to Import this SSAS Tabular Cube into Power BI. After creating the connection in Power BI Desktop to this SSAS Tabular cube we go to the Navigator window in Power BI that allows the user to pick from a list of tables. What we have noticted is that as we selected tables from the left hand side of the window, on the right hand side in the preview we can see that the data is displayed as a cross-product of all the rows in all the tables that have been selected. And as we select more tables the error mentioned above comes up. I am not sure why this cross-product takes place but as you can imagine what is in the preview becomes very huge very quickly and this is likely why the error comes up.
So our understanding is that anything imported into Power BI Desktop will be converted into a Tabular Model so it seemed counter intuitive to take an SSAS cube that is already in Tabular mode and have Power BI Desktop convert it to a Tabular Model again. Perhaps it is not best practice to do that in such a scenario.
So instead we did an Import directly from the source database (the star schema instead of our SSAS Tabular) and that worked. Also in that case, as we selected the list of tables to import (from the Navigator window) we did not see any cross-product taking place. The Navigator in this case was simply showing the contents of each table selected.
- Tx_jeremy7 years agoFrequent Visitor
This has to be a bug, right? There is no logical reason for the behavior to be a cross product. Import should either separate the tables selected into distinct tables in Power BI or it should create a results set based on the Tabular model.
- Anonymous7 years agoNot applicable
I can't confirm it's a bug but it sure sounds like one.