lookml-explore

Instructions

1. Core Standards

1. Naming Convention: snake_case for the Explore name.
2. Required Parameters:
   • description: 100% Coverage. Every Explore MUST have a description.
   • label: A user-friendly name for the Explore in the UI.
   • view_name: Defaults to explore name, but explicit definition is safer.
3. Joins:
   • relationship: Required (one_to_one, many_to_one).
   • sql_on: Required. Use ${left.id} = ${right.id} syntax.
   • type: defaults to left_outer. Use inner or full_outer explicitly if needed.
4. Formatting:
   • Do NOT use from to rename views just for aesthetics. Use view_label instead.
   • Exception: Polymorphic joins, Self-joins, Rescoping extensions.

2. Advanced Configuration

• always_filter: specific filters that users can change but cannot remove.
• sql_always_where: specific restrictions that users cannot change.
• persist_with: Link explore cache to datagroups (e.g., default_datagroup).
• fields: Use inclusive lists to strictly control content when necessary (ALL_FIELDS*, -view.field).

3. Performance Optimization (Aggregate Tables)

Aggregate Tables (Aggregate Awareness) allow Looker to query smaller, pre-aggregated tables instead of the raw granular data, drastically improving query performance.

Anatomy of an Aggregate Table

explore: orders {
  aggregate_table: rollup_name {
    query: {
      dimensions: [created_date, status]
      measures: [total_revenue, count]
      filters: [orders.created_date: "6 months"]
    }
    materialization: {
      datagroup_trigger: ecommerce_etl
      # partition_keys: ["created_date"] # BigQuery/Presto optimization
      # increment_key: "created_date"    # Incremental builds
      # increment_offset: 3              # Rebuild last 3 periods
    }
  }
}

Key Parameters

1. Query: Defines the "shape" of the rollup.
   • Dimensions: Include all dimensions commonly used in dashboards (including filters).
   • Measures: Include base measures (sum, count). Looker can derive averages from sum+count.
   • Filters: Optional. Restricts the rollup to a subset of data (e.g., "last 6 months").