Cross-filtering is available only for Looker’s new dashboard experience; it is not available for legacy dashboards. To determine which type of dashboard you are using, compare the screenshots on the Viewing dashboards and Viewing legacy dashboards documentation pages.

Cross-filtering makes it easier and more intuitive for viewers of dashboards to interact with a dashboard’s data and understand how one metric affects another. With cross-filtering, users can click a data point in one dashboard tile to have all dashboard tiles automatically filter on that value.

Multiple cross-filters can be created at one time, and cross-filters can be used in conjunction with standard dashboard filters.

Enabling cross-filtering on dashboards

For cross-filtering to be turned on successfully, all dashboard tiles must be created from the same Explore.

You can turn on cross-filtering on a dashboard with the following steps:

1. Make sure you have the Manage Access, Edit access level for the dashboard.

2. Enter the dashboard’s edit mode by selecting Edit dashboard from the three-dot dashboard menu.

3. Once in edit mode, click Filters in the blue toolbar at the top of the dashboard, and then enable the Cross-filtering switch.
   
   

1. If the dashboard contains any unsupported visualization types, a pop-up window appears to indicate that the dashboard contains tiles that cannot initiate cross-filtering:
   
   
   Click OK to continue enabling cross-filters.
2. Click Save in the toolbar.

Once cross-filtering is enabled on an instance and a dashboard, you can create cross-filters.

LookML developers may be able to use aggregate awareness to optimize cross-filtering performance. To learn more about using aggregate awareness with dashboards, visit the Creating user-defined dashboards documentation page.

Creating cross-filters on dashboards

To create a cross-filter:

1. Click a data point within a tile containing a supported visualization.
   
   For bar, column, line, area, scatterplot, and pie charts, you can also click the chart legend to initiate cross-filtering.
2. All data tiles will be filtered by the value of that data point.
   
3. The cross-filter field and value appear at the top of the dashboard for reference.
   
4. To add an additional value to an existing cross-filter, hold down the Command (Mac) or Control (Windows) key on your keyboard while clicking on additional data points or legends. For example, if you are already cross-filtering on the Traffic Source field for a value of Search and you want to add Organic as an additional traffic source, hold down the Command or Control key and then click the Organic region of a supported visualization.
   
5. The tiles will filter on the additional value as well.
   
6. The additional value for the cross-filter will appear at the top of the dashboard.
   
7. To add a new cross-filter, click another data point on any tile containing a supported visualization type.
   
8. All tiles will be filtered by that cross-filter as well, and an additional field and value appear at the top of the dashboard.
   

Supported and unsupported visualization types

Each data tile on a dashboard contains a visualization, and the type of visualization determines whether you can use the data points on that tile to create cross-filters or not.

Supported visualization types

The following visualization types are supported and can create cross-filters:

• Table charts
• Table (legacy) charts
• Column charts
• Bar charts
• Line charts
• Area charts
• Scatterplot charts
• Pie charts
• Map charts that use the Points option for data of type: location and map charts that use data of type: zipcode
• Static map (regions) charts
• Static map (points) charts
• Word cloud charts
• Donut multiples charts
• Funnel charts
• Timeline charts
• Waterfall charts
• Boxplot charts
• Custom visualizations

For information on how to enable cross-filtering on custom visualizations, visit the API 2.0 Reference documentation in the custom_visualizations_v2 GitHub repo.

Unsupported visualization types

The following visualization types are not supported and cannot create cross-filters; however, they can be filtered by cross-filters:

• Single value charts
• Single record charts
• Map charts that use Plot options other than Points for data of type: location, or map charts that use map layers
• Visualizations that use grid layout
• Visualizations based on merged results when all queries in the merged results are from the same Explore as the rest of the dashboard tiles (if a dashboard contains a merged results tile with queries from different Explores, cross-filters cannot be enabled for that dashboard)

Regardless of visualization type, you cannot create cross-filters (or standard dashboard filters) on a custom field.

Using range selection

You can use range selection to cross-filter a dashboard on an entire range. In order to use range selection:

• A dashboard tile must use a line, area, or scatterplot visualization.
• The visualization must use continuous data in the x-axis, which is usually numeric data, data using timeframes, or data of a time-based type, although some timeframes and time types are not continuous.
• The visualization’s x-axis Scale Type must not be Ordinal. If it is Ordinal, range selection creates separate cross-filters for each data point selected.

To use range selection, left click on one end of the range you want to select, and hold and drag to the other end of the range. When you release your click, the cross-filter will appear at the top of the dashboard and the other tiles will filter on that range. You cannot add additional values to range selection cross-filters.

In the screenshot below, a dashboard viewer has selected the weeks of November 25th and December 2nd. You can see the range selected because it is shaded. The cross-filter appears at the top of the dashboard.

In the screenshot above, each data point encompasses a week. While the shaded area on the visualization appears to end on December 2nd, the cross-filter actually ends on December 9th, the last day of that week. The cross-filter at the top of the dashboard displays the full range, November 25th through December 9th.

Non-continuous time types

Some timeframes and time-based types are not considered continuous, and if you use range selection with them, a separate cross-filter is created for each data point you select. Those time types are:

• date_time_of_day
• date_day_of_week
• date_month_name
• date_quarter_of_year
• date_fiscal_quarter_of_year

Using cross-filters together with standard dashboard filters

A dashboard can contain both cross-filters and standard dashboard filters. Both types of filters appear at the top of the dashboard. Cross-filters appear after standard filters and are a darker blue. For example, in the image below, the filter on Status is a standard filter and the filters on Created Month and Traffic Source are cross-filters:

Cross-filters differ from standard dashboard filters in a few ways:

• Cross-filtering needs to be enabled separately for your Looker instance and dashboard(s).
• You do not need edit access to a dashboard to create or change cross-filters (although you do need edit access to turn on cross-filtering for a dashboard).
• Cross-filters cannot be configured with the same dashboard controls as standard filters.
• Specific cross-filters are not visible to other users unless the cross-filtered dashboard is shared through a link or embedded into a website.
• Cross-filters are not saved to the dashboard; once you navigate away from a dashboard, cross-filters are cleared.
• Cross-filters are not applied to dashboard downloads, schedules, or alerts.

While using cross-filters, you can still make temporary changes to a dashboard’s filter values. However, if you create a cross-filter on the same field that a standard filter is using, the standard filter value cannot be changed until the cross-filter is removed.

If you have edit permissions for a dashboard, you can continue to edit and create standard dashboard filters while cross-filters are enabled. However, entering edit mode to edit or add standard filters will clear any current cross-filters.

Sharing cross-filtered dashboards

Cross-filters are not saved to a dashboard, but they can be shared in the following ways:

• By sharing a link to the dashboard
• By embedding a dashboard that contains cross-filters

Drilling when cross-filtering is enabled

When cross-filtering is enabled on a dashboard, you can drill by right-clicking to drill into the visualization’s data.

Removing cross-filters

There are two options to remove an individual cross-filter:

1. Click the data point being cross-filtered again.
2. Click the X next to the cross-filter value at the top of the dashboard.

If there are multiple cross-filters, you can clear them all by clicking the X to the right of the filter values at the top of the dashboard.

Entering edit mode also clears all cross-filters.

Optimizing cross-filtering performance using aggregate awareness

LookML developers may be able to optimize query performance of cross-filtered dashboards by using aggregate awareness. With aggregate awareness, developers can create aggregate tables under LookML explore parameters, which can optimize queries for one or more tiles in a dashboard. The first step is to get the aggregate table LookML, which appears under the Get LookML option in the dashboard three-dot menu. For more information, see the aggregate_table parameter documentation page.