Forum Discussion
Cannot create Calculated column in shared dataset
Hi all,
I published a dataset to workspace to share with others, then find out shared users cannot create calculated column, only possible with measure.
The goal is to let user have more autonomy in building the report with customized columns, then what should be the best practice?
Sharing dataflow seems to be unfavorable as the failure downstream in a refresh causing issue all upstream.
Kindly advise!
Hi, rdnguyen
If you have existing Power BI Premium Gen 2, Premium per capacity (PPC) or Premium Per User (PPU) workspace, you might consider creating a PowerBI datamart.
After you load the data into the datamart, you can model the data in the datamart such as creating relationships, creating mesures.
Best Regards,
Community Support Team _ Eason
4 Replies
Hi rdnguyen
Currently it is not possible for users to create calculated columns in the dataset, because for a calculated column the entire table needs to be reprocessed. Whilst when creating a measure it is only executed when it is put into a visual.
If you would want this to work please find the idea at https://ideas.powerbi.com
Thanks Gilbert, I will wait for the Datamart then 🙂
Hi, rdnguyen
If you have existing Power BI Premium Gen 2, Premium per capacity (PPC) or Premium Per User (PPU) workspace, you might consider creating a PowerBI datamart.
After you load the data into the datamart, you can model the data in the datamart such as creating relationships, creating mesures.
Best Regards,
Community Support Team _ EasonHello v-easonf-msft,
not sure if I understand correctly. I have the same use case as the original poster where another department has built a dataset and in order to map it with our data I connected to it and used calculated columns and tables based on the direct query table. Now I can´t refresh my dataset anymore via the service, which is a no go for the users.
I tried to move the ETL and data model steps inside datamarts, but you can´t get data from a shared dataset in datamarts? Or do you mean that the original source data, in my case from the other department, uses datamarts instead of a dataset? That would not work in my case... Or do you see any other workaround?
Best regards
Lennart
