Deployments – Part 8 Migrating from AAS to PPU
Welcome to part 8, where in this blog post, I am going compare deploying datasets.
For those people who are not exactly sure what deployments are, what this means is when you are using Power BI Desktop and you click on Publish, you are effectively deploying your changes to the Power BI Service (Which could also be a server in the cloud).
In this blog post I will show the differences when completing a deployment from AAS and then PPU.
There is a slight difference when deploying a project/dataset for the first time. Due to this fact when I am going to demonstrate below is when deploying changes or updates for AAS or PPU. This is where the majority of my time will be spent is making deployment changes which could include adding or updating measures. As well as adding or updating expressions, tables and potentially new columns.
Here are the previous 7 blog posts that I have completed in the series.
2 – Scalability – Part 2 Migrating AAS to PPU – Reporting/Analytics Made easy with FourMoo and Power BI
AAS – Deployments
There are multiple options to deploy changes for AAS.
Deployment using Visual Studio
What I am showing below is when I making a deployment using the native Deploy option from Visual Studio
Please note that I always default the Processing Option to “Do Not Process” so that it will not try and process the data when deploying the changes.
Another big thing to note with doing it this way is that it will overwrite everything that has been deployed.
A good example is if I had partitions in my table in production, and I did not have those tables in my Visual Studio Project. If I then had to deploy from Visual Studio I would overwrite all of those partitions and loose all the data in those partitions!
And this is what it looks like once the changes have been deployed
Deployment with Visual Studio and BISM Normalizer
A much better option is to use the BISM Normalizer extension which can be used with Visual Studio.
As shown below when using BISM Normalizer it gives me the control to not only decide what options to deploy it also allows me to decide which options I do or do not want to deploy.
This allows me to be able to granularly control changes and not overwrite something that could create an issue.
As shown below is when using BISM Normalizer it allows me to compare the differences between my Visual studio dataset and the AAS dataset.
I changed the “Select Actions” to only hide Skip objects, leaving me with objects that have Update/Create/Delete
I then validate the selections, which ensures that everything is valid and will succeed.
I then deploy the changes as shown below.
PIC
Tabular Editor Deployments
I can also use Tabular Editor to complete deployments for AAS.
As shown below I am using Tabular Editor 3 and making a deployment change.
Below are my deployment option settings, I have selected “Deploy Model Structure” so that it will not overwrite partitions etc
Then Deployment summary and then deploy
Once done I got the message
As I have shown there are quite a few ways to make deployment changes to AAS datasets.
PPU
Below I go over the steps on how to make deployment to PPU.
Deployment using Power BI Desktop
The default option or what people are most comfortable with is using Power BI Desktop to publish updates to the Power BI Service.
To do this using Power BI Desktop I would click on Publish.
And then I get prompted to Overwrite the existing dataset if it already exists.
Some things to be aware of when completing this deployment:
- This will overwrite all the tables, data, sand partitions.
- Once this is deployed, there is no way to undo the changes.
- If you have configured incremental refreshing, on the next refresh it will re-process all the partitions again (this could take some time and potentially appear to be missing data)
Deployment with ALM Toolkit
Another option that is available us to use the ALM Toolkit when deploying to a PPU dataset.
ALM Toolkit is the updated version of BISM Normalizer that is used in Visual Studio.
You can download ALM Toolkit here: Home Page – ALM Toolkit (alm-toolkit.com)
I then go to the External Tools menu and click on ALM Toolkit.
This then opened the ALM Toolkit, and I then selected my PBIX file and the associated Workspace and dataset as shown below.
As shown below I now can selectively choose what I want to deploy.
The advantages of this are the following:
- I can selectively choose which items to deploy.
- I can configure to not overwrite existing partitions
- I can also save the report differences to know what was changed.
- Finally, I can also view which changes are about to made and if that is what I expect.
Deployment using Tabular Editor
As shown previously, I can also use Tabular Editor to complete deployments for PPU.
As shown below I am using Tabular Editor 3 and making a deployment change.
Below are my deployment option settings, I have selected “Deploy Model Structure” so that it will not overwrite partitions etc
Then Deployment summary and then deploy
Once done I got the message
Summary
As I have demonstrated in this blog post when completing a deployment to either AAS or PPU, they are quite similar.
And I personally feel at this point that it is just as easy to complete granular deployments to AAS as it is to PPU.
Thanks for reading and if you have any comments, please leave them in the section below.