Building Power BI semantic models with Direct Lake in Microsoft Fabric

In this video we explore Direct Lake mode in Microsoft Fabric, and how we can use it to create really speedy Power BI reports, even on really large datasets.

We look at what Direct Lake is, the benefits of Direct Lake and how to configure it depending on your requirements.

#powerbi #microsoftfabric #dataanalytics

Timeline:
0:00 Intro
1:18 Exploring the Kaggle dataset
2:54 Default semantic model
4:16 Creating your own semantic model
5:08 What is Direct Lake mode?
7:30 Benefits of Direct Lake mode
7:55 Building a semantic model
9:16 Creating a Power BI report from our model
11:26 Direct Lake Fallback Behaviour
13:30 Semantic model refresh setting
15:12 Fallback behaviour + Semantic Link

[kates]

Great video! Thank you for sharing your knowledge and experience.

[sandrojorgeoliveira]

Will, thank you for share your knowledgement!

[oliveroshea]

these videos are really good. Haven't watched them all yet, but they are on my to do list! Many thanks!

[azwarmzafar]

This is another great one, continuing my journey of going throguh this series. many thanks. I see you as an infuencer in this domain of introducting MS Fabric on YT soon you will be top voice one.

[chrish.]

Hey there, enjoying your Videos on Fabric! In this case I believe it would be important to mention that any user interaction is consuming CU of your Capacity. So depending on the number of users and the interaction direct lake can be quite expensive in number of CU compared to import mode.

[davidrojas]

Thanks for your videos!!! They are all really useful in my job
Just one question, is there a way to create another variables/columns in power query if we use One lake / lake house?

[ThePerkses]

Great videos! Are you able to use a shortcut as a source for a semantic model or do you need to move the data out of a shortcut first? Shortcuts don’t show up as tables when I try to pull them into a new model.

[adatalearner]

can one create semantic model based of the tables from external sources such as on-prem database, azure sql or synapse workspaces?

[kyyxgij]

What persmissions do people need, that want to see the data in the report which is using direct lake?

[mkj-md]

Hello Will, Thanks for the videos, I have two questions:
1- If in the semantic model itself, you created a table using DAX for example a datetable using CALENDAR, would the Direct Lake fall back to Direct Query?
2- I am having a problem when creating a semantic model from 2 large delta tables (Billions of records, yes Billions not millions), I am getting an error of can't add tables, the connection timed out, and i am not sure why? i thought the semantic model will not be creatinga duplication of data that already exist in the lakehouse?
Thanks in advance

[FatherLizard]

Thanks so much for this. Did you make it to Fabcon Will?

[williamstanleyjevons]

Hello really nice video! Do you know if there is a way to transform the data of that semantic model in a power query like method or are we obliged to transform the data upstream ?

[MrThuralion]

Hey Will, how much have you tried Fabric with a paid capacity versus trial? I run into a lot of snafu when I use my paid capacity where it says I need to upgrade to trial, though it feels like that should be a downgrade! Any similar experiences?

[powerbicoach]

What is the benefit of Direct Lake on data imported and transformed once a day?

[MrThuralion]

I can't create a semantic model in a Workspace that doesn't have a lakehouse right? And if I have 2 lakehouses in 1 workspace I can't connect to tables from both when creating a semantic model? I'm guessing I should avoid creating architecture based on this, thoughts?