Single PBIX vs multiple attaching to a common Power BI data source

[cklopfen]

I'm working on a multi-developer project that will result in 6 fairly substantial Power BI reports connected to a single data model. I had planned to deploy 7 separate pbix's: one for each report and a stand-alone file for the data model. This was mostly so that developers could work on reports in parallel without conflicting with other coders.

We'll be embedding the reports into an existing web site. We thought there was the possibility that granting access to each report separately might cause performance issues and be overly complex.

So I'm trying to balance potential performance lags from security resources and the dev challenges of having multiple people working on reports simultaneously. Having individual development pbix files that we merge into a single deployment report is certainly an option, but it seems hacky. Also, should we have the full data model and the reports in a single file if we go that route? Or is it better to keep the data model in one pbix and let all the reports access the data source?

[v-deddai1-msft · Accepted solution]

Hi @cklopfen ,

You can create your perfect data model in Power BI Desktop and publish it to the Power BI service. Users canestablish a live connection to a shared dataset in the Power BI service and create report depend on it.

About Power BI Embedded performance, you can follow the advice in ：https://docs.microsoft.com/en-us/power-bi/developer/embedded/embedded-performance-best-practices#:~:....

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

Best Regards,

Dedmon Dai

[Greg_Deckler]

@cklopfen Generally I have seen people go with a Golden Dataset approach and the dataset is in a separate PBIX than the reports and the other reports connect live to the dataset. Check out Power BI Usage Models In Pictures - https://community.powerbi.com/t5/Community-Blog/Power-BI-Usage-Models-in-Pictures/ba-p/1342820

There are other things you can do for development like Azure Dev Ops and Power BI Pipelines.

[cklopfen]

Thanks! That's certainly the approach I've taken in the past. But I'm wondering what kind of performance hit that approach will take with embedding each report separately and having to resolve tokens, etc for 6 reports instead of 1.