Skip to main content

I am having issues with decimal numbers in my Power BI model created with TimeXtender XMLA endpoint (version 7026.1).
No matter the precision of my datatype on SQL server it seems like all values are forced into a rounded 4 decimal size. Is there a setting hiding somewhere to prevent this behavior?

Below are examples or test data to demonstrate the behavior.

Inspecting the model with tabular editor it seems all decimal fields are stored as fixed decimal rather than the expected double


 

Test data

 

Result in Power BI

 

Hi,

if you look into your Power BI model is the data type a Fixed Decimal or a Decimal? Fixed decimals are rounded at 4 decimals: https://learn.microsoft.com/en-us/power-bi/connect-data/desktop-data-types


It’s shown as fixed decimal (see my first screenshot), but I don’t see any way to override that datatype through TimeXtender, directly or indirectly.

Usually 4 decimals would be sufficient, but in this case I actually need precision to be at the 6th decimal


Hi,

yeah - you can only choose a decimal formatting which just adds a presentation on top of the datapoint. It lt looks like the type chosen is DataType.Decimal (which is Fixed Decimal) instead of DataType.Double which would allow more precise values.

Good find - I never noticed this. I am not 100% sure this was always the case, but it looks like there is no way to generate more precise data in PBI.


Note that the type naming is rather confusing: DataType.Double is Decimal in PBI and the equivalent of float in SQL Server. DataType.Decimal is Fixed Decimal Number / Currency in PBI and the equivalent of decimal(19,4) in SQL Server.


Hi ​@nmo does Rory’s comment above answer your question? If so please help us by marking a best answer. If you have follow up questions please let us know


Hi ​@Christian Hauggaard 

I don’t really see any solution to the problem other than some change in TimeXtender to allow users to either explicitly select which datatype to use in the power bi model or to have TimeXtender determine the appropriate datatype based on the datatype in the DWH

numeric(38,20) → Double

numeric(19,4) → Fixed decimal


Hi nmo,

I suspect a SQL Server float will become a Power BI double.


I can confirm that changing my SQL datatype to float triggers a Power BI double.

However, I feel like this is just replacing one uncertainty with another, so as a work-around I have just multiplied my SQL values by 1.000.000 and then I divide by the same value in my frontend measures.