When publishing a report with data from PBI Dataflow, am I storing twice the data?

[Anonymous]

Hi,

I have created a dataflow with tables A, B, C, then I have imported those tables in Power BI Desktop to create a report; once I publish the report and a new dataset is created in the Service, am I storing twice A, B, C?

In other words, suppose the tables A, B, C have a size of 20MB, am I "duplicating" this size?

Thanks,

Roberto

[v-lili6-msft]

hi  @Anonymous 

Yes, it will load the data from dataflow and then transform and modeling in power bi desktop, when you publish it to power bi service, it will create a dataset.

https://ssbipolar.com/2019/10/14/power-bite-dataflows-vs-datasets/

Regards,

Lin

[Anonymous]

Hi @v-lili6-msft,

thanks for your reply. 

So if I have to use table A for n reports, in relation with different tables for each report, and then publish all of them, am I storing table A n times?

If yes, I suppose that's not the right way to proceed. 

Maybe it could make more sense to create a unique dataset with all relations, store it only once and always use that dataset for the n reports? 

The problem with this approach could be that if the other tables are very large I could have performance problems by importing all of them (even if not used).

I hope I've been clear.

Thanks

Regards,
Roberto

[v-lili6-msft · Accepted solution]

hi  @Anonymous 

Yes,  it will store table A n times.

You connect dataflow, it has the same logic for you connect other datasource, then you transform and modeling and create a dataset in power bi service, dataflow is a datasource  not a dataset.

For your requrement, you could create a dataset then share/use it to create reports, now it will not store data anymore in power bi service, it is a live connection model for these reports.

https://docs.microsoft.com/en-us/power-bi/service-datasets-share

Regards,

Lin