Modelling – Part 7 | Migration from AAS to PPU
In this blog post I am going to show the differences when creating a model/dataset when completing the same task in Azure Analysis Services (AAS) or Power BI Premium Per User (PPU)
Here are the previous 6 blog posts that I have completed in the series.
Modelling – AAS
When creating a new dataset in AAS this is typically done with Visual Studio, there are a lot of similar items between Visual Studio and Power BI desktop, the biggest difference is how you access them are very different as well as their look and feel.
When starting with a new dataset the process is to create a new data source, import some data and then to create the tables.
This will then bring you into the Power Query Editor (which has a different look and feel) but has got the same features in Power BI Desktop’s version of the Power Query Editor
If there are any modifications that need to be made, they are done in the Power Query Editor and then loaded into the dataset.
Below is what a typical Visual Studio project looks like for an AAS dataset, where I can create relationships, Roles, Perspectives and measures.
What I then do is I use Tabular Editor to then create my measures and any other operations related to tables in my dataset.
Below is what it looks like in Tabular Editor for the same Visual Studio Project as shown above.
Below are some of the Pro’s and Con’s when using AAS with the combination of Visual Studio / Tabular Editor
- Integrated Source Control
- Visual Studio is a mature product.
- Only a single user can access the Model.BIM at a time
- Using Power Query Editor is different to Power BI Desktop and can take some getting used to not having an intuitive GUI.
- Has a more developer approach, which requires a different way to navigate and create the dataset which can be a challenge to new users.
- Not a big community of support when running into issues when using AAS with Visual studio.
Updates are infrequent and can take time to get deployed for Visual Studio.
Modelling – PPU
When creating a new dataset for PPU, this is done using Power BI Desktop.
The good news here is that the process of creating the dataset is exactly the same as creating any other dataset using Power BI Desktop.
It starts with getting data via Power Query Editor, where I connect to the data and then either load the data or transform in Power Query Editor.
This data then gets loaded into the dataset where I can create relationships, Roles, Perspectives and measures.
Whilst some of the tasks currently must be done in Power BI Desktop, I also use Tabular Editor to create the measures and other table properties.
As shown below this is what it looks like when I open Tabular Editor and work on my existing dataset.
As I did previously when looking at AAS, below are the Pro’s and Con’s for modelling in Power BI Desktop
- Familiar easy to use GUI
- A lot of support options via the Power BI Community and video content online.
- Frequent updates to Power BI Desktop.
- No integrated source control.
- Only a single user can access the PBIX file at a time.
I hope that you have found the migration comparison from AAS to PPU interesting and informative.
In my opinion I do think it is possible to migrate to use PPU.
The biggest challenge to me, would be the integrated source control that is currently available in AAS/Visual Studio. Other than that, to me everything can be done in either AAS or PPU it just is shown differently.
Thanks for reading of there are any comments or suggestions please let me know.