July 12, 2002 at 9:19 am
I attempted to change the Design storage for the a cube from MOLAP to ROLAP. When I did this and processed the cubes I started running into errors when the aggregation tables were being created in the SQL Server database. The message that I get is as follows:
Error(-2147221409): SQL error (Mandatory statement failed with error 'Disallowed implicit conversion from data type smalldatetime to data type float, table 'Northwind_Orders6.dbo.Order_Cube_Order_Cube_NF6', column 'Year_L6'. Use the CONVERT function to run this query.
There's a lot more to the message like the actual SQL statement that was constructed on the fly. The crux of the problem seems to be that its setting up a Year column in the aggregate table but is not doing a datepart on the source column which is in the format mm/dd/yyyy and is trying to insert it as is into a float column. I also noticed in another case that it was creating the aggregate table with a smallmoney column where the source table had the same column defined as a money column. So obviously when the sum function was applied it became an even bigger number that could not be inserted into the smallmoney column in the aggregate table. It appears to me like there's a bug in Analysis services that is causing it to come up with the wrong table definitions for the aggregate table. I was wondering if any of you had noticed this problem and if there was a workaround or a patch that would fix this problem for me since this is really associated with the internal working of Analysis Services rather than with any sort of code that I'm coming up with.
Regards, John
August 18, 2002 at 11:22 am
I am not sure if SP2 has fixed this or not but it sure sounds like a bug to me. Are you running SP2 on both SQL Server and Analysis Services? If not, you might try installing it (you can download it free from the SQL Server site on microsoft.com).
hth,
Michael
Michael Weiss
Michael Weiss
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply