Microsoft Fabric, first impressions

This Week, Microsoft released the new Fabric resource on Azure. It’s been hard to miss as many MVP’s were tweeting, Mastodoning and sending messages on LinkedIn that Build should not be missed this year. So yeah, it was clear something huge was going to happen.

When the announcement came, I was outside. My kid had to leave for his training on his racing bike and that took precedence. But I could follow along via Twitter. Of course, the first messages were ‘amazing’, ‘fantastic’ etc. As it’s new technology and, as my coworkers describe it, my species are immediately jumping on it. New shiny toys are always welcomed! But I had to give them credit, from what I read on the tiny screen outside in the sun, it was looking good!

After coming home, first thing I did was dig into all the links that were shared. Let’s see what people are blogging about and what Microsoft is writing about. Because the last few years, I’ve seen massive improvements in the docs. There were a number of blogs and whole set of links to be found. I’ve included my list at the bottom of this blog for readability.

Learn path

The first thing I decided to check out was the learn path. Granted, I activated my trial license for Fabric before that but then the learn path. As chance would have it, I was working my way through the Azure Data Engineer learning path to prepare for renewal of that certification. And a lot of the things I had read that day came back. But there were changes, updates and new things as well. So after about two hours I completed the theory and felt happy to try an end-to-end tutorial.


Microsoft has created a number of so-called end-to-end tutorials where you are guided from scratch to a small working data solution. I took on the datalake one as it is closest to my daily work.

The things I struggled most with was getting a CSV file uploaded into the lake and processed. The tutorial expected me to have OneDrive storage available, but in my tenant this isn’t the case. So I had to create a work-around with a file storage, open it up to Fabric and get the data from that one.

If anything, it did show some flexibility and you should be able to put a file on your OneDrive, start a pipeline and load the data into your OneLake. Pretty cool. But, as it is CSV, pretty slow as well. The CSV file with a few records took 55 seconds to process into the datalake.

The next part of the tutorial made me get data from parquet files and load that into the OneLake. Obviously this went much faster. Let’s start with the fact table:

Followed by the dimensions:

This went quite quick, less than a millisecond to process over a thousand rows. Eat that CSV ;).
But yeah, this can be done in Synapse Analytics as well. And that is true. But, Synapse won’t provide you with a data set, ready for PowerBI. You have to do extra work to get the lineage done (think Purview).

More than a mashup

And this is where, in my current opinion, Fabric stands out compared to the different components that created it. It is more than putting Data Factory, Synapse, PowerBI and Purview together. It’s a seamless integration that could even make die-hard back-enders like me create reports. And, because you can auto create reports from the PowerBI interface, even prevent me from using pie charts. Yes, I’ve listened to you Benny de Jagere and Alexander Arvidsson. Still, I’d strongly advise anyone to keep me away from reporting ;).

What’s next?

Well, for my database performance tests, I’ve created a TPC-H dataset. I’m really curious to see what happens when I try and load multiple gigabytes of data, create a models and auto-create a report. The tech will work, no doubt there. But the performance is more interesting.

Another big question is the pricing model. I haven’t seen anything from that yet at the time of writing (may 26th 2023), and I hope there will be some sort of training on that. My guess is that it will be in some way connected to the PowerBI licensing and I can’t understand that one. What I’m hoping for is a fixed license per user and a pay per use model like the Sql Serverless databases.


Finally, I’ll leave you with a number of links that helped me out getting started. No doubt I missed a lot of them, feel free to add them in the comments.

Microsoft Build 2023 Book of News

Click here to open

Microsoft Documentation

Microsoft Blogs

Community Blogs (alphabetically)–business.aspx?m=1

Youtube Videos

What is Microsoft Fabric? Is it an analytics platform for everyone?
Webinar Series: Introduction to Microsoft Fabric
Can cross-cloud data analytics be easy? | Microsoft Fabric
Advancing Fabric – What is Microsoft Fabric?
Advancing Fabric – Lakehouse vs Warehouse

Thanks for reading, leave your links in the comments and happy Fabricing(?)!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s