Concepts → Prompt

About a prompt

A prompt is a type of dashboard filter. A dashboard developer creates and configures the properties of one or more prompts for a given dashboard using the Dashboard Filters Manager.

Unlike other types of dashboard filters, a prompt can be simply a filterable column, a partial filter expression, or a complete filter expression. Like a presentation variable, a prompt overrides the default behavior of a dashboard which is to coalesce a list of dimension columns from all insights on all tabs as filterable columns available in the Filter and Search controls of the Filter bar.

A prompt as a pill has various properties. Once defined and configured, a prompt always appears in the Columns panel of the Filter dialog. However, a prompt may not always appear initially as a filter bar pill. There are two properties that influence this behavior, the Default to Display property and the Mandatory property. Here is a brief summary of how these properties affect a prompt's behavior:

• A prompt with the Default to Display property disabled does not initially appear as a filter bar pill and requires a dashboard consumer to select a filter operator and filter value in the Filter dialog.
• A prompt with the Default to Display property enabled with an undefined filter value appears initially as a filter bar pill and requires a dashboard consumer to select a filter value in the Filter dialog.
• A prompt with the Default to Display property enabled with a defined filter value appears initially as a filter bar pill and applies the initial filter expression as a dashboard runtime filter.
• A prompt with the Mandatory property enabled always appears as a filter bar pill. However, the Default to Display related properties for a filter operator and filter value determine if there is an applicable filter expression to apply as a dashboard runtime filter.

A dashboard developer can order prompts alphabetically or manually so as to affect how the prompts appear as filterable columns in the Filter dialog.

Here are some the reasons why a dashboard developer creates one or more prompts for a given dashboard:

• Govern the list of available filterable columns
• Control the display order of filterable columns in the Filter dialog
• Specify a measure or key column as a filterable column
• Specify complex filter expression as a filterable column
• Specify a partial filter expression as filterable column for the dashboard consumer to complete
• Implement a default dashboard runtime filter for a hierarchy level
• Implement a default dashboard runtime filter that appears as removable filter bar
• Implement a mandatory dashboard runtime filter that is a non-removable filter bar pill

A dashboard developer with Edit access rights to a dashboard can create and edit one or more prompts using the Dashboard Filters Manager. Here's how:

• In the Action bar of a dashboard, select More Options (⋮ vertical ellipsis icon).
• In the More Options menu, select Dashboard Filters Manager.
• From the Data panel, drag & drop a column or formula to the Prompts tray.
• Configure the properties of the prompt pill.

Valid references in a filter expression for a prompt pill

As a complete filter expression, a prompt pill consists of at least one filterable column, a filter operator, and an optional filter value. A prompt pill that has a filter expression with an undefined filter value represents a partial filter expression. A partial filter expression results in an incomplete dashboard runtime filter, and as such, does not affect the dashboard.

A filterable column for a prompt pill can be a data-backed column or a formula column from:

• an alias
• a physical schema table
• an Incorta Analyzer table
• a materialized View
• a business Schema View
• an Incorta Analyzer View
• an Incorta SQL View

A formula expression can encapsulate the filterable column or filterable columns within the filter expression.

Certain filter operators afford various options for selecting one or more filter values. For example, the In and Not In operators allow for the selection of none, one or more, or all filter values.

Depending on the data type and filter operator, a filter value for a prompt pill can be a scalar value or array of scalar values. In certain cases, it is possible to reference a variable as follows:

• a filter expression session variable
• an internal session variable
• an external session variable
• a global variable
• a date system variable
• a miscellaneous system variable

You can also define a prompt pill as a formula and createa a filter expression using the Formula Builder.

View access rights

In order to reference a column or variable in a filter expression, you must have View access privileges to the grandparent object such as a physical schema, business schema, or session variable. A global variable is available to all users.

About an applicable insight for a prompt pill

The joins between the physical schema tables in the query plan for an insight measure determine whether or not a filterable column affects the insight. For example, a physical schema column in a parent table serves as a filterable column for a measure in a child table. However, a physical schema column from a child table is not a filterable column for a parent table, even when it is the foreign key reference.

For a filterable column, a prompt meets one of the following conditions:

• The filterable column is a key, dimension, or measure column in the insight.
• The filterable column has a sibling column from the same physical schema table, and that sibling is a dimension or measure for the insight.
• The filterable column is a dimension or key that shares a measure on the insight from a common child table with other dashboard insights.

If an applicable insight already contains an insight filter, an applied filter will complement the insight filter based on a common filterable column as described above.

Properties for a prompt pill as a column

A prompt pill can be a column or a formula column.

A prompt pill that is a column can be simply a filterable column, a filterable column with a partial filter expression with an undefined filter value, or a a filterable column with a complete filter expression that serves as a dashboard runtime filter.

The following are the properties of a prompt pill that is a column:

Properties for a prompt pill as a formula with a partial filter expression

A formula expression can encapsulate the filterable column or filterable columns within the filter expression.

Here is an example of formula for a prompt pill that is a partial filter expression.

Copy
SALES.SALES.AMOUNT_SOLD - SALES.SALES.COST_OF_GOODS

In this example, the two columns in a calculation serve as the filterable columns in the filter expression. A dashboard consumer can select the filter operator and the filter value(s). The data type for the formula result determines the available filter operators. The formula result determines the available filter values.

The following are the properties of the prompt that is a formula pill with a partial filter expression such as a calculation:

Properties for a prompt pill as a formula with a complete filter expression

A formula expression can encapsulate the filterable column or filterable columns within the filter expression. Here is an example of a formula for a prompt pill that is a complete filter expression.

Copy
or(
    SALES.PRODUCTS.PROD_CATEGORY = "Photo",
    startsWith(
        SALES.PRODUCTS.PROD_NAME,
        "17"
    )
)

In this example, there are two filter expressions within the OR boolean built-in function:

Copy
SALES.PRODUCTS.PROD_CATEGORY = "Photo"

and

Copy
startsWith(SALES.PRODUCTS.PROD_NAME,"17")

As complete filter expression, these will evaluate to either True or False. The following are the properties of the prompt that is a formula pill with a complete filter expression: