Measure Check

The measure check is specifically designed for Cube.js semantic layer validation. It validates pre-defined measures from your Cube.js data model, making it ideal for semantic layer data quality checks.

Configuration

Parameter	Required	Description
name	Yes	Unique name for the check
dataset	Yes	Cube name or SQL query
type	Yes	Must be measure
measure	Yes	Cube.js measure name
condition	Yes	Comparison operator
threshold	Yes	Value to compare against
dimensions	No	Cube.js dimensions for grouping
filter	No	WHERE clause conditions

Prerequisites

• Cube.js instance must be running and accessible
• Datasource must be configured with type: cube
• Measures must be defined in your Cube.js schema

Examples

Basic Measure Check

- name: total_revenue_measure
  dataset: Orders
  type: measure
  measure: totalRevenue
  condition: ge
  threshold: 1000000

This check validates the totalRevenue measure from the Orders cube.

Measure with Dimensions

- name: revenue_by_region
  dataset: Sales
  type: measure
  measure: revenue
  dimensions: [region, quarter]
  condition: gt
  threshold: 50000

This check ensures revenue by region and quarter exceeds $50K.

Complex Measure

- name: customer_lifetime_value
  dataset: Customers
  type: measure
  measure: lifetimeValue
  dimensions: [customerSegment]
  condition: ge
  threshold: 500

This check validates customer lifetime value by segment.

Time-based Measure

- name: monthly_active_users
  dataset: Users
  type: measure
  measure: monthlyActiveUsers
  condition: gt
  threshold: 10000
  time_dimension:
    name: createdAt
    granularity: month

This check ensures monthly active users exceed 10K.

Cube.js Integration

Sample Cube Schema

// schemas/Orders.js
cube(`Orders`, {
  sql: `SELECT * FROM orders`,
  
  measures: {
    totalRevenue: {
      type: `sum`,
      sql: `order_amount`
    },
    
    averageOrderValue: {
      type: `avg`,
      sql: `order_amount`
    },
    
    orderCount: {
      type: `count`
    }
  },
  
  dimensions: {
    region: {
      sql: `region`,
      type: `string`
    },
    
    createdAt: {
      sql: `created_at`,
      type: `time`
    }
  }
});

Datasource Configuration

datasources:
  - name: cube
    type: cube
    uri: http://localhost:4000/cubejs-api/v1

Generated SQL

The measure check generates Cube.js API calls that translate to optimized SQL:

-- Generated by Cube.js for totalRevenue measure
SELECT SUM(order_amount) as totalRevenue
FROM orders

With dimensions:

-- Generated by Cube.js with dimensions
SELECT 
  region,
  quarter,
  SUM(order_amount) as revenue
FROM orders
GROUP BY region, quarter

Use Cases

• Semantic Layer Validation: Ensure business metrics are correct
• KPI Monitoring: Validate key performance indicators
• Business Rules: Check calculated business measures
• Data Consistency: Ensure measures match expectations
• Cube.js Testing: Validate Cube.js schema definitions

Advantages over Raw SQL

1. Pre-calculated: Measures are optimized by Cube.js
2. Business Logic: Includes complex business calculations
3. Caching: Benefits from Cube.js caching layer
4. Consistency: Uses same definitions as dashboards
5. Security: Respects Cube.js security context

Example Results

✓ total_revenue_measure: 1250000 (≥ 1000000)
✗ revenue_by_region_west_q1: 45000 (> 50000)
✓ customer_lifetime_value_premium: 750 (≥ 500)

Common Measure Types

Revenue Measures

- name: monthly_recurring_revenue
  dataset: Subscriptions
  type: measure
  measure: mrr
  condition: gt
  threshold: 100000

User Engagement Measures

- name: daily_active_users
  dataset: UserActivity
  type: measure
  measure: dau
  condition: ge
  threshold: 5000

Conversion Measures

- name: conversion_rate
  dataset: Funnel
  type: measure
  measure: conversionRate
  condition: ge
  threshold: 0.05

Error Handling

Common issues and solutions:

• Cube not found: Verify cube name matches schema
• Measure not found: Check measure exists in cube definition
• API connection: Ensure Cube.js instance is accessible
• Security: Verify API token and permissions

Performance Tips

1. Pre-aggregations: Use Cube.js pre-aggregations for large datasets
2. Caching: Leverage Cube.js caching for repeated checks
3. Partitioning: Use time-based partitioning in cube schema
4. Indexes: Ensure underlying tables have proper indexes

Related Checks

• Numeric - Raw SQL numeric expressions
• Sum - Simple sum aggregations
• Row Count - Basic count validations