Press "Enter" to skip to content

Category: Microsoft Fabric

Cloud Connections in Microsoft Fabric

Published 2024-10-10 by Kevin Feasel

Dennes Torres makes a connection:

I wrote about cloud connections when they were in a very early stage.

Cloud connections evolved and are now sharable. We call the “regular” connection as “personal connection”.

The problem with the “personal connections” is the difficult to make teamwork. The personal connections belong to you and different developers can’t use them. When a different developer needs to work with the same objects, they are required to create their own connection.

Using cloud connections, we can create a single, reusable connection to the data source and share it with all the developers in the team.

Read on to learn more about how they work now that the feature is a bit more mature.

Comments closed

Exploring Semantic Model Relationships with Sempy

Published 2024-10-09 by Kevin Feasel

Prathy Kamasani builds a graph:

Understanding the relationships between datasets is crucial in data analytics, especially in the world of self-service BI. Sempy, a Python library unique to Microsoft Fabric, allows users to visualise these relationships seamlessly. This post explores using Sempy to visualise semantic model relationships and view them in a Power BI Report. Viewing them in Notebook is easy and has been documented on MS Docs.

Click through for a notebook and explanation of the underlying code.

Comments closed

Referencing a Microsoft Fabric ML Model from another Workspace

Published 2024-10-08 by Kevin Feasel

Sandeep Pawar crosses workspaces:

I have written a couple of blogs about working with ML models in Microsoft Fabric. Creating experiments and logging and scoring models in Fabric is very easy, thanks to the built-in MLflow integration. However, the Fabric Data Science experience has one limitation. There are no model endpoints yet, and you cannot load a model from another workspace because the model URI, unlike in Databricks, does not reference a workspace. If you use MLFlowTransformer as shown in this blog, only the model from the workspace where the notebook is hosted is loaded. However, there is a workaround.

Read on for that workaround, as well as the core limitation associated with it.

Comments closed

Announcements from the European Fabric Community Conference

Published 2024-10-08 by Kevin Feasel

James Serra brings tidings:

A TON of new features announcements at the European Microsoft Fabric Community Conference help last week. The full list is here, and I wanted to list my favorite announcements from that list:

  • Access Databricks Unity Catalog tables from Fabric (public preview): You can now access Databricks Unity Catalog tables directly from Fabric. In Fabric, you can now create a new data item called “Mirrored Azure Databricks Catalog”. When creating this item, you simply provide your Azure Databricks workspace URL and select the catalog you want to make available in Fabric. Rather than making a copy of the data, Fabric creates a shortcut for every table in the selected catalog. It also keeps the Fabric data item in sync. So, if a table is added or removed from UC, the change is automatically reflected in Fabric. Once your Azure Databricks Catalog item is created, it behaves the same as any other item in Fabric. Seamlessly access tables through the SQL endpoint, utilize Spark with Fabric notebooks and take full advantage of Direct La

Read on for the rest of what James found exciting.

Comments closed

Spaces in Microsoft Fabric Delta Table Names

Published 2024-10-04 by Kevin Feasel

Sandeep Pawar is looking for a bit more space:

One of the annoying limitations of Direct Lake (rather of the SQL endpoint) was that you could not have spaces in table and column names in the delta table. It was supported in the delta table but the table was not query-able in the SQL endpoint which meant you had to rename all the tables and columns in the semantic model with business friendly names (e.g. rename customer_name to Customer Name). Tabular Editor and Semantic Link/Labs was helpful for that.

But at #FabConEurope, support for spaces in table names was announced and is supported in all Fabric engines. You have to use the backtick to include spaces, as show below.

Read on to learn more about how you can create these, what the limitations are, and then you can decide whether it’s worth it to have spaces in table names.

Comments closed

Recovering Power BI Reports You Cannot Download

Published 2024-10-03 by Kevin Feasel

Kurt Buhler grabs a report:

Below are some reasons why you might not be able to download your Power BI report or model from a workspace:

  • The report was created in the service:
    • Someone created the report manually (using the User Interface) and connects to a model in another workspace.
    • Someone created the report programmatically (for instance, using the REST APIs).
    • Power BI created the report automatically (for instance, it copied the report to a workspace that belongs to a later stage in a deployment pipeline)
  • You used the REST APIs to re-bind a report (changed which model it connects to as a data source).
  • The model has incremental refresh enabled.
  • The model uses automatic aggregations.
  • The model was modified via an XMLA endpoint.
  • Other scenarios described in the limitations in the Microsoft documentation.

When you encounter this scenario, you see something like the following image, which shows the Download this file option greyed out from the File menu of the Power BI report.

Read on to see how you can nonetheless recover these published reports using the semantic-link-labs library.

Comments closed

Plotting the ROC Curve in Microsoft Fabric

Published 2024-10-03 by Kevin Feasel

Tomaz Kastrun gets plotting:

ROC (Receiver Operation Characteristics) – curve is a graph that shows how classifiers performs by plotting the true positive and false positive rates. It is used to evaluate the performance of binary classification models by illustrating the trade-off between True positive rate (TPR) and False positive rate (FPR) at various threshold settings.

Read on to see how you can generate one in a Microsoft Fabric notebook. Tomaz also plots a density function for additional fun.

Comments closed

Mounting Azure Data Factory in Fabric Data Factory

Published 2024-10-03 by Kevin Feasel

Andy Leonard takes up a factory job:

Thanks to the hard work of the Microsoft Fabric Data Factory Team, it’s now possible to mount an Azure Data Factory in Fabric Data Factory. This post describes one way to mount an existing Azure Data Factory in Fabric Data Factory. In this post, we will:

  • Mount an existing Azure Data Factory in Fabric Data Factory
  • Open the Azure Data Factory in Fabric Data Factory
  • Test-execute two ADF pipelines
  • Modify and publish an ADF pipeline

Read on to see how it all works. One of the odd things about Microsoft Fabric—and its predecessor, Azure Synapse Analytics—is the penchant for similar-but-not-quite-the-same services. Yes, we have Data Factory…but it’s not quite the same. Yes, we have Azure Data Explorer (and KQL)…but it’s not quite the same. I get that there are reasons for this (such as not having a resource group with a dozen separate services hanging around), but I’m sure it’s a bit frustrating working on several separate code bases and trying to keep them all approximately in sync.

Comments closed

Programmatic Power BI Report Modification via semantic-link-labs

Published 2024-10-02 by Kevin Feasel

Kurt Buhler makes a change:

Whether building reports in Power BI Desktop or in the web browser via the Power BI service, you have limited options to batch or streamline changes. Put another way; it’s tedious and slow to make many small changes to one or more Power BI reports. It’s also easy to make mistakes

When initially designing or building a report, this is not so much of a problem. Unless you’re using a template, you want to control report layout and formatting, yourself. However, certain changes can be little more than a waste of time. Some examples include:

  • Replacing fields when there’s a broken reference due to i.e. renaming a model measure or column.
  • Swapping one measure or column for another in the report
  • Changing visual container styles, like background, border, and shadow/glow.
  • Changing text or text styles across multiple visuals, pages, or reports.
  • Changing chart formatting (like color) or properties (like edit interactions) across multiple visuals, pages, or reports.

Read on to see how you can make some of these changes in Python code using the semantic-link-labs library.

Comments closed