Відео

😅

Interesting demo. Was interested to learn about how the data sources were being reference and use between the Bronze, Silver and Gold layers.

Measures created in the datamart won't be available in the power bi desktop if connect via import mode.

I'm just here for the battle 😉

I just did it yesterday and used it to authenticate in a Copy Activity in a Data Pipeline from a Sharepoint Online List, but didn't work. Error: invalid credentials 😞 On the Group members, great advice, have to be member not only owner in the Group. 👍

How do i embed my powerbi dashboard in a company intranet

Q Thank you Chris for this wonderful setup of Azure log analytics with power bi, can I customise my Power BI dashboard in the sense take only part of data from Log Analytics and create dashboard? Thank you very much

If you are using shared datasets as source, you would probably need them to create dimensions, groups not in source datasets.

Hi, cant wait for the proof why. I know there are problems if using big tables where there can be better compression. However I would like to see difference in using date table or some dimension. I did not see big differences in that

Isn’t the data already imported? When you schedule the refresh does it have to move data from the data mart to the dataset? I’m hoping I can build one data mart and import everything one time and not have to import it again into different datasets. Ultimate I need 4 different datasets. But they all use a majority of the same tables. I don’t want to refresh 4 datasets when I’m just importing the same data 4 times.

Could you do a video about migrating to an F64 from a P1? Extra credit if they’re in different regions. :)

Thank you all, that was a great session. So useful for the teams.

Thanks Chris, I'm trying to migrate a legacy SSAS setup over to Azure and this is a possible solution. I almost missed the key part of this - tabular editor and it's ability to edit/read a SSAS data source and then export to a AAS data set - the powerbi bit is optional IMO.

bro its fake tho

That cut at -60º below zero, probably couldnt keep a straight face lol

Thanks . Can you share the link ?

thank you so much

😂😂😂😂😂😂

Would this feature help if we have no users? Meaning will this feature help with big ugly refreshes that timeout on occasion?

Congratulations!

Thanks for this, my kids think I am famous now that I was mentioned in a youtube video. I did some more digging including asking David Browne on the CAT team. Here is a summary of whats going on with scale out:, from a technical perspective, behind the scenes. Well at least as much as Microsoft was willing to say. I do find it interesting that there are non-intuitive features in Fabric, that Microsoft won't explain how they work because they are considered proprietry. It comes down to "Trust me". Things are definitely getting complex in Fabric world. Capacity Units (CU) and V-cores: -CU is directly related to v-cores and the size of a premium capacity. -CU is limited and additive within a single capacity. -CU is primarily tied to CPU usage and is shared across the entire capacity. Nodes and Resources: -A capacity is made up of one or more nodes (essentially VMs). -Each node has its own allocated resources, including CPU, memory (RAM), and storage. -While CU is additive across the entire capacity, other resources like storage and some aspects of RAM are specific to each node. Semantic Model Hosting: -Conceptually, each semantic model (dataset) is hosted on a different Analysis Services server with a limited amount of RAM (think docker container or Virtual machine) -This is an approximation; the actual implementation differs but aims to behave similarly. Scale-out Feature: -This feature operates within a single capacity, not by creating new capacities. -It doesn't increase total CU or overall capacity utilization. -Instead, it separates operations (like refresh and query) onto different nodes within the capacity. Benefits of Scale-out: -Reduces resource contention, especially for I/O and memory-intensive operations. -Allows a semantic model to use more resources than available on a single node/server. -Each operation (refresh/query) can utilize the full storage and RAM resources of its respective node without competing. Impact on Performance: -While it doesn't reduce overall CU usage, it can significantly improve performance by reducing bottlenecks (such as I/O rather than CPU). -This is especially beneficial for large models or when dealing with resource-intensive operations. RAM Considerations: -There's still some complexity around how RAM fits in as a "resource". Sounds like this undocumented stuff is the "secret sauce" behind the scenes of Power BI. -While total RAM is capped at the capacity level (e.g., 50 GB for a P2), the scale-out feature allows for more efficient use of available RAM across nodes. -The exact implementation of RAM allocation and usage is not fully documented and may be more nuanced than a simple per-node allocation (again secret sauce). Conceptual Model vs. Actual Implementation: - The conceptual model (separate SSAS servers for each dataset) helps understand the behavior, but the actual implementation is more complex and not fully documented.

Never a DAX calculated column! 💯💯💯

This is affirming to hear! My journey into data took me from a CRM to a no-code GUI to query our database, to Power BI, Power Query, and now SQL. Next will be Python and then who knows?!

Like your videos, how can we load table in dbo schema in fabric. Table is available but not able to query due to missing in dbo schema. can you help me

Learning basic SQL even helps with concepts in Power Query and other languages. You may not write SQL, but the concepts are very transferable.

Power Query of course! 😊 or start with PQ and move up to more code it you need to.

Almost getting boring

HALLELUJAH! Sir, this is the SINGLE most helpful tutorial I have found for autorefresh. I have struggled with this for a solid three months now battling with the gateways. THANK YOU!

#BDE #ftw

Can I create a custom sunburst visual from this?

Great video! One of my main concerns recently has been migrating from my existing semantic models using import mode to new one's using directlake. I was worried I would need to rewrite all my relationships, measures, roles, etc. Looks like Tabular Editor could save me the headache.

Excellent video, thank you. The issues we are having seem to be related to the limitations of service principal when using Fabric API. For example, we can run GET .../v1/workspaces/{id}/items/{id} no problem but then trying to list tables in a Lakehouse (GET .../v1/workspaces/{id}/lakehouses/{id}/tables) we get "The operation is not supported for the principal type" message. Do you have any experience with that issue?

Thanks !

Amazing 42 minutes of my time! Thank you for putting this video up. I am also struggling in trying to understand on how to integrate all the latest features in PBI. This end to end project of yours is just an eye opener on what MS Fabric can do.

I'll give it a go thanks

You are trolling Satya come on he is great for Microsoft

RE gIt integration. I just want a few buttons in Desktop..clone commit, pull, merge etc. Current implementation pathway feels way too convoluted..

Like so many Fabric features, I don't understand how this feature works. As I interprete it, this "scale out " is happening on a single capacity. But a single capacity has a set amount of CPU (CU) resources. This is measured as Interactive CU- as it is linked to DAX query execution. Interactive CU is additive within a smoothing window, so if one DAX query consumes 10% of a capacity, two similar queries would consume 20% of the capacity. So I don't see how this feature actually reduces capacity utilization, and potential throttling. Can anyone explain the mechanics of how this is working, or where my logic is incorrect.

Good one

"It's the end of the trial as I know it, and I feel fine" -- REMish

Lovely. I prefer XP.

Is there a way to have incremental refresh on dataflow Gen2? I'm using dataflow gen2s to load tables in the warehouse but can't find a way to incrementally refresh them

A second question. With dataflows would one still require Analysis Services?

If in a small organization there is a single data analyst publish reports, would the old method of creating reports using Power BI Desktop be better than using dataflows. I guess once you have teams this is where dataflows shine. Awesome explanation. Thank you

Is the only way to refresh the data by clicking refresh it would not refresh automatically like power bi does?

It's a pity this is only for Premium workspaces - also is there a way to I.D. users that have downloaded a semantic model from the service?

BDE!!!!! Congrats.

Congratulations!! 🎉

I actually do like them but for a specific purpose: Scatter Plots

my usecase is that power bi needs to access a restapi with currently a separate token required but that expires so it needs to be resetted manually. Could a service principle be used to do the refresh for the dataset via rest api? Note the restapi can use entra id as identification. I am not sure whether that is possible and if so how to set it up. Any ideas or hints here? Thx and br