LookML Sets

Syntax

set: set_name {
  fields: [field_name1, field_name2, view_name.field_name3, ...]
}

Key Features

• Reusability: Define a list of fields once and use it in multiple places.
• Composition:
  • Include All:

    [set_name*]

    includes all fields from another set.
  • Exclude:

    [-field_name]

    excludes a specific field.
  • External Reference:

    view_name.set_name*

    refers to a set in another view (ensure views are joined).

Use Cases

1. Drill Fields (Primary Use Case)

view: orders {
  # ... dimensions ...

  set: order_details {
    fields: [id, created_date, status, user.email]
  }

  measure: count {
    type: count
    drill_fields: [order_details*]
  }
}

2. Controlling Explore Visibility (Field Picker)

explore: orders {
  fields: [ALL_FIELDS*]         # Start with everything (default)
  # OR
  fields: [orders.order_details*, users.user_info*] # Whitelist specific sets
}

3. Excluding Fields

explore: orders {
  fields: [ALL_FIELDS*, -users.password_hash]
}

Best Practices

1. The

detail

Set

detail

drill_set

view: users {
  dimension: id { primary_key: yes ... }
  dimension: name { ... }
  dimension: email { ... }

  # Standard set for drilling
  set: detail {
    fields: [id, name, email]
  }
}

2. Naming Conventions

• Use

  snake_case

  for set names.
• Use descriptive names like

  user_info

  ,

  financial_metrics

  ,

  drill_detail

  .

3. Avoid Cross-View Dependencies in View Sets

other_view.field

other_view

• Better: Define sets local to the view (only its own fields).
• Explore Level: Combine sets from different views at the Explore level (e.g.,

  fields: [orders.detail*, users.detail*]

  ).

4. Cumulative Sets

set: basic_info {
  fields: [id, name]
}

set: extended_info {
  fields: [basic_info*, email, created_date]
}