Power BI Update October 2024

Combo charts are now supported

You can now use visual calculations in combo charts, such as the line and clustered column chart, just like you could in the other chart types. Here’s an example of a visual calculation returning the moving average over three quarters:

Field parameters are now supported

This month we have enabled the use of visual calculations with field parameters, you can add a visual calculation to a visual that contains a field parameter and vice versa.

Field parameters can be used to quickly switch around what’s shown in a visual. For example, you can create a field parameter to enable your users to decide which attributes of a dimension to show. In this example a field parameter called Product Attribute can be used to determine what the Percent of grand total visual calculation returns:

The Percent of grand total visual calculation is defined using the template as:

Percent of grand total = DIVIDE([Sales Amount], COLLAPSEALL([Sales Amount], ROWS))

Since the Percent of grand total visual calculation used here refers to ROWS as its axis, it will update and reflect the correct values when another product attribute is picked:

Please try out combining field parameters and visual calculations and let us know what would make the combined experience better for you by commenting below or at feedback.

Faster way to add a templated visual calculation

You can now add a templated visual calculation with fewer clicks by clicking on the bottom part of the New visual calculation ribbon button to see a menu that includes the templates. Clicking on a template will open the visual calculation mode where you can fill in the template and add your visual calculation.

If you want to create a new visual calculation without using a template, either select the top part of the New visual calculation button or choose Custom from the visual calculation template menu shown above.

This month, the Azure Maps visual brings a powerful enhancement to its functionality with data bound reference layers.

In previous releases, the Azure Maps reference layer was limited to static shapes without the ability to conditionally format or bind geometries to customers’ business data. Additionally, the static nature of the reference layers limited user interaction, preventing actions such as selecting, filtering, clicking, or accessing tooltips for polygons and points, unlike other visual components.

With the data bound reference layer, this limitation is addressed by allowing seamless integration between the reference layers and customer business data. Reference layers can now be dynamically bound with the spatial fields used, empowering customers to visualize their business data in context with geographic or spatial elements. Users can now update their visuals in real time, interact with their data through Power BI’s standard features such as filtering, cross-highlighting, and tooltips—greatly enhancing the flexibility and interactivity of the Azure Map visual.

Making your reference layers data bound is easy to do. We’ll automatically map the shapes in your reference layer to values of the field in the Location bucket in the Build pane based on the name property you provide in your reference layer file.

This update also allows you to customize the colors of your shapes as well, using features like conditional formatting or through tying their color to a legend color.

Shapes that aren’t tied to a value in your model are considered unmapped. You can format them to use custom colors or hide them completely from your map. As cross-highlighting is a temporary filter on the map, the treatment you apply here is also what will happen to unselected shapes when cross-highlighting from another visual.

Marker enhancements

Revamping the rendering of columns, bars, ribbons, lines, area charts, and markers is a top priority. These elements form the foundation of our core visuals and will eventually impact other areas. By providing more control, our report creators can enhance their storytelling and help users easily interpret data.

In the October 2024 update, markers for line charts, scatter charts, and anomalies are improved with this revamp. This update introduces new options that offer greater customization and flexibility, explore these new options and maximize their potential.

Markers for line and scatter charts can be customized in two ways:

• Categories: When the chart has no series, the dropdown menu displays categories. You can customize each data point’s marker based on the selected x-axis category.
• Series: When the chart displays a legend, the dropdown menu displays series. You can customize the markers for the complete set of data points within the selected series.

You can hide or show markers for a specific data point category by toggling the ‘Show for this category’ option. Please note that the ‘Markers’ toggle has been moved under ‘Show for all series.’

New format settings have been added to markers for line charts, scatter charts, and anomalies, including:

• Shape: Shape markers continue to offer control over their type and size. Additionally, rotation is now available for all shape types, except for the circle shape. Rotating shapes enhance the variety of shape types at your disposal, which is particularly convenient when multiple lines require unique shapes.

• Color: Changing the color of markers has always been a convenient control. Now, you can also modify the transparency of markers for a specific category, series, or all markers.

• Border (New): Borders for markers have been introduced, allowing you to add borders to a specific marker category, series, or all markers. Additionally, you can fully customize the marker borders by adjusting their transparency and width.

New List Slicer (Preview)

In this update, we would like to introduce the new List Slicer. To try it, navigate to Options and settings > Options > Preview features > List slicer visual. Major enhancements are coming, including image support, labels, additional conditional formatting options, and improved default styles specifically designed for hierarchical layouts.

Please note, this new visual is in its early development stage, we don’t recommend using it in production currently. However, this is an excellent opportunity to experience the capabilities of this new slicer and provide us with feedback for future improvements.

The new List Slicer can become hierarchical when more than one field is dragged into the field data well. This action will activate additional format settings specific to hierarchical data.

Another advantage of the new slicer is the level of customization it offers, including:

• Selection: Customize how items are selected within the slicer.
• Shape: Adjust the shape of the slicer to fit your design needs.
• Layout: Modify the layout to better organize the slicer elements.
• Overflow: Manage how the slicer handles overflow content (e.g., Continuous scroll, paginated).
• State styles: Define styles for different states (e.g., selected, unselected, on hover).
• Selection icon: Choose an icon to represent selected items. Available for the ‘Tile slicer’ too.
• Expand/Collapse icon: Select icons for expanding or collapsing hierarchical data.
• Button styles: Customize the appearance of buttons within the slicer.

Modeling

Dynamic format strings for measures (generally available)

Dynamic format strings for measures are now generally available, giving you ultimate flexibility in how measures are displayed in visuals. These format strings can be conditionally applied using Data Analysis Expressions (DAX) based on the measures value, filters applied, and/or to add additional information, such as units.

If you haven’t already, check out the Guy in a Cube video, the deep dive blog, and feature documentation to learn more and get ready-to-use examples of how to apply them today. If you know of a community post or video about these, please share it in the comments!

Introducing INFO.VIEW Data Analysis Expressions (DAX) functions

DAX query view introduced new DAX functions to get metadata about your semantic model with the INFO DAX functions, and now four of these functions are also available as INFO.VIEW DAX functions, which convert IDs to friendly names, and can be used in calculated tables of the semantic model in addition to being able to run them in DAX query view.

INFO.VIEW.TABLES() shows information about the tables in your model, including what storage mode each table is in. You can also quickly identify tables marked as a date table by the Data Category of Time.

INFO.VIEW.RELATIONSHIPS() shows information about all the relationships in your model, including a relationship column giving a quick summary of the to and from columns with direction and cardinality.

INFO.VIEW.MEASURES() shows information about the model measures, including if it’s in a valid or error state.
INFO.VIEW.COLUMNS() shows information about the columns, including data category and data type.

Value filter behavior (preview)

DAX has an automatic filtering mechanism that occurs when multiple columns from the same table are filtered. This behavior is informally called ‘auto-exist’. DAX understands that not all combinations of values across these columns are valid and as a result it automatically excludes invalid combinations. The DAX Engine generated a coalesced value filter that not only returns valid combinations but also affects measured calculations. This month we are giving you more control over whether you want this behavior in your semantic model. You can decide whether you want to turn off coalesced values filters and turn on independent value filters instead. Turning on independent value filters by setting the ‘Value filter behavior’ setting to Independent (see below) results in multiple filters on the same table being kept separate instead of the DAX engine combining these into one.

Understanding current value filter behavior

When you are filtering multiple columns on the same table, the current default value filter behavior takes these filters and combines them into one, considering only the combinations that exist. Consider the following two columns on the same table:

• Year, which contains values like ‘2023’.
• Month, which contains values like ‘January 2024’.

If you filter on both Year and Month, since these columns are on the same table, the value filter behavior combines the filters into one, but only the combinations that exist are considered. Of course, the combination of the month January 2024 with year 2023 does not exist and would not be included in the filter. There are, however, situations in which the results are surprising.

Let’s look at an example, where we have a catalog showing availability of colors for products by year. The manufacturer of these products has experimented with making products in various colors throughout the years:

We have three products that have been available in various colors over the years. Notice how there are no red products offered in 2024. That is going to be important a little later.

Now, let’s count the number of products by adding the following measure:

Number of Products = COUNTROWS( 'Catalog' )

The following matrix shows the number of products that are available in various colors per year:

Now, let’s add another measure to calculate the total number of products for all years:

Number of Products All Years = CALCULATE ( [Number of Products], ALL ( 'Catalog'[Year] ) )

Let’s put these measures side-by-side and filter to year 2023 and just the blue and red colors (no black). As you can see the number of products is 4 and the number of products across all years for these two colors is 6:

If we switch the Year to 2024, we expect the ‘Number of Products’ measure to return 2, as there are just two products that are blue in 2024 and there are no red products in that year.

On top of that, we would expect that the number of products for all years will not change, because, after all, it is supposed to be calculated across all years. However, the ‘Number of Products for All Years’ changes from 6 to 5:

The number of products across all years should still be 6, not 5. What we are seeing here is the value filter behavior in action: it is combining filters on the same table, removing combinations that did not exist. The filters are Year = 2024 and Color = Blue or Red. Since these two filters are on the same table, these filters are combined into one filter that only filters for the combinations that exist. Since there are no red products in 2024, the applied filter is Year = 2024 and Color = Blue.

Therefore, the number of products for all years now counts just the number of blue products, not the blue or red products. This returns 5, as you can confirm in the table.

Influencing the value filter behavior

This month we are giving you control over whether you want this behavior in your semantic model, by using the ‘Value filter behavior’ setting on your semantic model in the properties pane in the model view:

Three options are available:

• Automatic – This is the default setting and currently turns on the Coalesced behavior. When we wrap up this preview, new models set to ‘Automatic’ will use Independent, there will be an announcement at that time.
• Independent – This forces filters on the same table to be kept separate. After setting the ‘Value filter behavior’ setting to ‘Independent’, the total number of products for all years returns 6 as expected (see below).
• Coalesced – This forces the value filter behavior to be enabled for the semantic model and will result in combining the filters on the same table into one. The number of products for all years in our example will continue to return to 5.

The table below shows the impact of this setting to our example:

Note that setting the ‘Value filter behavior’ to Automatic, means it is equal to Coalesced for now, but will be switched to Independent for new semantic models in the future.

If you set the ‘Value filter behavior’ to Independent, the number of products for all returns 6, as expected, since the filters are Year = 2024 and Color = Blue or Red and are no longer combined: