Dataflows

[PowerRon]

Hai All

can somebody tell me why I should use dataflows if I have a data warehouse where (almost) all transformations and cleansing takes place?

I try to think of requirements of using them, but can't find one

Regards 

Ron

[Anonymous]

Hi @PowerRon, 

Assuming your DWH takes care of all the cleansing and transformation, and there's no real need for existing flows to be enhanced.

I typically position dataflows to a business/technie user to have them prototype something, in case the DWH Team is low on resources to build a new set of facts & dimensions. This helps us get an understanding of what they mean, and if they've done it properly you can refactor it into your existing solution.

Additionally, you might have your business users use dataflows, connected to an external data lake storage, and have them use it as some sort of repo for mapping tables held in Excel. Think SQL MDS, but then without the Silverlight 😂

Then have Azure Data Factory (or Databricks, or whatever ..) pick up those files from the external data lake storage.

[Anonymous · Accepted solution]

Hi @PowerRon .

I think it depends on user's requirement.

https://ssbipolar.com/2018/11/22/positioning-power-bi-dataflows/

This blog talks about the position of Power BI Dataflows, you can take a look if you have time.

Best Regards,

Jay

Community Support Team _ Jay Wang

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

[PowerRon]

Thnx @Anonymous your URL gives a lot information.
Certainly https://ssbipolar.com/2018/11/27/power-bi-dataflows-faq/ makes things more clear.

It is positioned as an analyst / business user help/tool for self-service BI, but in my opinion it is also possible to use it to make for a certain table one version of the truth. Whereby for instance unneccesary columns are deleted, technical names are replaced by logial etc and use this dataflow as input for one or more datasets.

[PhilC]

Hi Ron, agree with where you ended up here.  Also means less calls back to the database where multiple people are building from the same data.