How field mapping for custom actions works in Insights

Automated field mapping is based on identical fields. Fields with the same name and data type map automatically across datasets. Their field names and data types must be an exact match. This works similar to a join, except that it is automatically generated based on names and data types for every matching field. If you are missing fields, you can create them by using calculated fields in the dataset that's missing a field.

It's important to make sure that all target fields are mapped if they are enabled for use with a filter action (in the Filter scope). Doing this allows filtering to apply automatically. If some target fields aren't mapped, the automatic filtering doesn't work.

Mapping is generated only when you create or save a custom action. So after every change that affects the mapping, make sure to return to it and save it again. When you create an action, mapping is based on the fields as they exist at that point. When you save an action, any mapped fields that you renamed since you created the custom action stay mapped. However, if you alter the data type of a mapped field, the mapping is removed.

If your mapping is missing some fields, you can do one of the following to fix it:

Only target the mapped fields, by removing the unmapped fields from the Filter scope.
Remove the visual in question from the target visuals.
Create calculated fields to supply the missing fields for the mapping, and then save your custom action.

NOTE The information that displays on the mapping screen shows the configuration from the most recent time you saved it. To refresh or update the view, save the action again.

If you remove a parameterized field or any other targeted field from the source visual, the action that uses it breaks. The action for the missing field either doesn't work when you select a data point, or it's hidden from the context menu.