case

Usage

view: view_name {
  dimension: field_name {
    case: {
      when: {
        sql: SQL condition ;;
        label: "value"
      }
      # Possibly more when statements
      else: "value"
    }
    alpha_sort:  yes
  }
}

Hierarchy View File view Field Type case

Possible Field Types Dimension Accepts A SQL condition and a string Special Rules Use an alpha_sort parameter if you want the values alphabeticalized

Definition

case lets you bucket results with case logic. While you can write raw SQL CASE statements instead, using case will create a drop-down menu for your users in the Looker UI. A SQL CASE statement will not create such a menu.

The general form of case is:

dimension: status {
  case: {
    when: {
      sql: condition ;;
      label: "Label of Condition"
    }
    # possibly more when statements
    else: "Label If No Condition Met"
  }
}

These parameters work as follows:

• when — You may use as many when statements as you would like to represent each condition for which you want to supply a label. The when statements are evaluated in order from the first one listed to the last one listed, and the first when statement that is evaluated to true will assign the associated label.
• sql — The sql parameter accepts a SQL condition that evaluates to true or false.
• label — If the SQL condition is true, this is the label that will be assigned. The assigned label has a data type of string. The value of each label in a case statement must be unique. If you use the same label value for multiple SQL conditions, only the last SQL condition in the case statement is assigned the label value. See Examples on this page.
• else — If none of your conditions are met, this is the label that will be used.

Caution: The suggestions presented to a business user for a case dimension will not be restricted when using full_suggestions:yes. Any option that you include in a case parameter will be visible to all users. The label values from the case parameter will also take precedence over suggestions from the suggestions parameter.

Choosing the sort order of labels with alpha_sort

Typically, case values appear in the order you write them. If you prefer an alphabetical sort, you can use alpha_sort: yes like this:

dimension: status {
  alpha_sort: yes
  case: { ... }
}

Examples

Assign several human-readable labels to different status numbers:

dimension: status {
  case: {
    when: {
      sql: ${TABLE}.status = 0 ;;
      label: "pending"
    }
    when: {
      sql: ${TABLE}.status = 1 ;;
      label: "complete"
    }
    when: {
      sql: ${TABLE}.status = 2 ;;
      label: "returned"
    }
    else: "unknown"
  }
}

When the same condition is repeated and evaluates to different labels, LookML uses the first condition that evaluates to true. In the following example, ${TABLE}.status = 0 evaluates to pending and not returned, since the pending condition is evaluated first.

dimension: status {
  case: {
    when: {
      sql: ${TABLE}.status = 0 ;;
      label: "pending"
    }
    when: {
      sql: ${TABLE}.status = 1 ;;
      label: "complete"
    }
    when: {
      sql: ${TABLE}.status = 0 ;;
      label: "returned"
    }
    else: "unknown"
  }
}

When multiple conditions evaluate to the same label, LookML uses only the first condition. In the following example, Looker will use ${TABLE}.status = 0 instead of ${TABLE}.status = 2 to generate the SQL CASE statement that evaluates to pending. When ${TABLE}.status = 2, the CASE statement evaluates to unknown.

view: orders
dimension: status {
  case: {
    when: {
      sql: ${TABLE}.status = 0 ;;
      label: "pending"
    }
    when: {
      sql: ${TABLE}.status = 1 ;;
      label: "complete"
    }
    when: {
      sql: ${TABLE}.status = 2 ;;
      label: "pending"
    }
    else: "unknown"
  }
}

Things to know

When a dimension contains a case parameter that references another field, that additional field may be added to the underlying SQL of a query that the dimension is used in. If the referenced field is not present in the query's visualization, and the visualization is a table chart with manually rearranged columns, the column order in some downloaded formats may be affected.