Power BI bookmarks: separate presentation state from data filters

A bookmark can capture more than visual visibility. Depending on its settings, it can restore data selections and other report state. If a button is intended only to switch a chart into a table, it should not unexpectedly reset the reader's date or region selection.

Define the intended state change before creating the bookmark. “Show table” and “Reset filters” are different actions and deserve different behaviour and labels.

Create a controlled view-switching example

Using the retail lab model, build a Paid category bar chart and an equivalent category table. Both should use the same measure and support the same page filters.

Add buttons labelled Chart view and Table view. Use the Selection pane and bookmarks to control which visual is visible. Review the bookmark's Data, Display and visual-scope settings so it captures only the state required for that action.

Microsoft documents those choices in its report-bookmark guidance. Inspect current application options rather than assuming a bookmark records only what is visible on the canvas.

Use a filter that exposes accidental resets

Select January 2026 before switching views. The Paid total should remain 47,500 paise, with Software 29,000 and Training 18,500. Switching presentation should not restore all-period Paid value of 69,500 unless the button explicitly promises a reset.

Then select February and repeat. Expected Paid value is 22,000. Test both directions, because one bookmark may have been created before the intended settings were applied to the other.

This exercise's expected values come from the fixture. Bookmark behaviour itself must be checked in Power BI; the supplied scripts do not execute the report interface.

Scope the bookmark to the intended visuals

When a bookmark should affect only a pair of presentation visuals, use the relevant visual scope and verify the selection before creating or updating it. Accidentally capturing a page-wide state can affect unrelated elements.

Give visuals and bookmarks meaningful names so the author can identify what is being controlled. Names such as CategoryChart and CategoryTable are easier to review than a series of default object names.

After adding a new visual or changing a slicer, revisit the bookmark configuration. An old bookmark can preserve obsolete assumptions even when the page's current appearance looks correct.

Build a separate reset action

A Reset filters button may intentionally restore a documented default period and clear region or product selections. Label it accordingly and state the resulting default where it matters.

Test whether reset should preserve the current page, visual mode or user-selected presentation. A reader may want to clear analytical filters without losing the table view they chose for accessibility or detail inspection.

Do not use a visually identical button for two different state changes. The label should explain whether the action changes layout, data selection or navigation.

Test interactions beyond slicers

Bookmarks can interact with visual selections, cross-highlighting and other captured report states. Select a category bar, switch views and observe whether the intended selection remains or clears.

Check drill-through and return navigation if the page participates in that journey. A restored bookmark should not leave a stale hidden filter that changes the next detail page unexpectedly.

Also distinguish author-created report bookmarks from a consumer's personal saved views. Test the actual distribution context used by the report, because reader state can affect how the initial page appears.

Preserve accessible navigation

Use clear button text and logical keyboard order. Verify that hidden visuals do not create a confusing reading or navigation experience in the final configuration. Provide an obvious indication of the selected view where useful.

A screenshot of each state cannot establish interaction correctness. Record the sequence of actions and the expected measure values after each step, then run that sequence in the application.

Exercise: intentionally create a view bookmark with a saved all-period data state. Select January, activate it and observe the unwanted reset. Correct the configuration and verify that both view-switching buttons preserve January while the separate reset button behaves as documented.

NeuraPath's Data Analytics with Generative AI course connects Power BI interactions with reliable analysis. A polished report makes state changes predictable and keeps presentation controls from silently changing the question being answered.

Continue learning

This article is part of the Power BI data models and reporting sequence. Use the neighbouring tasks when you need the prerequisite or the next application.

• Review the prerequisite or neighbouring task in Design dashboard tooltips that explain a metric.
• Continue with Create a dashboard freshness indicator.