Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Earn the coveted Fabric Analytics Engineer certification. 100% off your exam for a limited time only!

Reply
schwinnen
Helper V
Helper V

Create Dataflow with dataflow as source

I created some dataflows and now I am attempting to create some "sub-dataflows" from those dataflows.  When I select dataflow as the source, however, it asks for the gateway information.  I provide the gateway.  Under Authentication kind, I cannot select anything and I cannot select "Next" to continue.  The other dataflows are using the same gateway so I don't know what else I need to do.  

 

In case anyone has other suggestions, my goal is this:

I have created dataflows with a ton of data.  I've set up incremental refesh on this data.  Several reports can now use these dataflows, but they don't need all the data.  Therefore, I load this data into Power BI desktop and set filters there.  I'm new to this method so I could be incorrect, but I believe when I save the file, the file size is based on the filtered data.  This is good.  However, when I refresh the data in desktop, it takes forever, which leads me to believe it is still pulling in all of the data before filtering.  Either that or it just takes that long to filter the data.  So the plan was to create these "sub-dataflows" which would be filtered versions of the original dataflows.  

So, overall I'm looking to:

1. Utilize incremental refresh

2. Keep desktop file size as small as possible

3. Keep desktop refresh as fast as possible 

1 ACCEPTED SOLUTION
GilbertQ
Super User
Super User

Hi there

What happens is when you use a dataflow in Powr BI Desktop it will import ALL the data from the dataflow. This is by design.

There is good news.

What you can do is to create 1 dataset which has got all the data that you need modeled as a Star Schema.

You can then upload this dataset to the Power BI Service.

Next you can then use Power BI desktop and connect to the shared dataset in the Power BI Desktop.

This will mean that your reporting files will be small because they are connecting to the dataset in the Power BI Service.




Did I answer your question? Mark my post as a solution!

Proud to be a Super User!







Power BI Blog

View solution in original post

4 REPLIES 4
StanleyBlack
Helper I
Helper I

Would you not be better off building your dataflow structure from smallest granularity to the greatest rather than complete to subset? By that I mean say you were looking at sales data and wanted to subfilter by year in your example. Would it not be feasable to create the following:

 

Dataflow A with last years sales

Dataflow B with current years data

Then either in dataflow C through a linked entity you append A and B or have the combine be done in the Power BI file.

 

This would mean you meerly change the source in the dataset depending on your requirements.

GilbertQ
Super User
Super User

Hi there

What happens is when you use a dataflow in Powr BI Desktop it will import ALL the data from the dataflow. This is by design.

There is good news.

What you can do is to create 1 dataset which has got all the data that you need modeled as a Star Schema.

You can then upload this dataset to the Power BI Service.

Next you can then use Power BI desktop and connect to the shared dataset in the Power BI Desktop.

This will mean that your reporting files will be small because they are connecting to the dataset in the Power BI Service.




Did I answer your question? Mark my post as a solution!

Proud to be a Super User!







Power BI Blog

@GilbertQ- what do you mean by "Star Schema"?  

I do use the method you described for some other reports - connecting desktop reports to datasets in Power BI Service.  The downside, from what I can tell, is that I am limited as to what I can do with these desktop reports.  I can create measures, but I cannot create calculated columns.  

HI there

Yes that is correct in that you can only create new measures and not calculated columns.

If that is required then it would have to be added in the dataflow, which means that all others can then also use those new columns




Did I answer your question? Mark my post as a solution!

Proud to be a Super User!







Power BI Blog

Helpful resources

Announcements
April AMA free

Microsoft Fabric AMA Livestream

Join us Tuesday, April 09, 9:00 – 10:00 AM PST for a live, expert-led Q&A session on all things Microsoft Fabric!

March Fabric Community Update

Fabric Community Update - March 2024

Find out what's new and trending in the Fabric Community.

Top Solution Authors
Top Kudoed Authors