Max Check

The max check validates the maximum value of a numeric column. This check helps ensure data quality by validating that the largest values in your dataset meet expectations and don't exceed reasonable bounds.

Configuration

Parameter	Required	Description
name	Yes	Unique name for the check
dataset	Yes	Table name or SQL query
type	Yes	Must be max
measure	Yes	Column name to find maximum
condition	Yes	Comparison operator
threshold	Yes	Value to compare against
dimensions	No	Group by columns
filter	No	WHERE clause conditions

Examples

Basic Max Check

- name: maximum_order_amount
  dataset: orders
  type: max
  measure: order_amount
  condition: le
  threshold: 10000

This check ensures no order exceeds $10,000.

Max with Filter

- name: max_discount_percentage
  dataset: promotions
  type: max
  measure: discount_percentage
  condition: le
  threshold: 0.5
  filter: status = 'active'

This check validates that active promotions don't exceed 50% discount.

Max by Dimensions

- name: max_salary_by_level
  dataset: employees
  type: max
  measure: salary
  dimensions: [job_level]
  condition: le
  threshold: 200000

This check ensures maximum salary per job level doesn't exceed $200K.

Max with Time Dimension

- name: daily_max_temperature
  dataset: sensor_readings
  type: max
  measure: temperature
  condition: le
  threshold: 45
  time_dimension:
    name: reading_date
    granularity: day

This check validates daily maximum temperature stays below 45°.

Use Cases

• Data Quality: Ensure no unreasonably high values
• Business Rules: Validate maximum acceptable limits
• Range Validation: Check upper bounds of data ranges
• Outlier Detection: Identify potentially problematic maximum values
• Security: Prevent values that might indicate data issues or attacks
• Compliance: Ensure maximum limits are not exceeded

Generated SQL

The max check generates SQL like:

SELECT MAX(order_amount)
FROM orders
WHERE status = 'completed'

With dimensions:

SELECT job_level, MAX(salary)
FROM employees
GROUP BY job_level

Example Results

✓ maximum_order_amount: 8500.00 (≤ 10000)
✗ max_discount_percentage: 0.75 (≤ 0.5)
✓ max_salary_by_level_senior: 180000 (≤ 200000)
✗ max_salary_by_level_executive: 250000 (≤ 200000)

Data Types

The measure column should be numeric or comparable. Common types include:

• INTEGER
• DECIMAL / NUMERIC
• FLOAT / DOUBLE
• DATE / TIMESTAMP
• TIME

NULL Handling

• MAX() ignores NULL values automatically
• If all values are NULL, MAX() returns NULL
• The check will fail if MAX() returns NULL (unless specifically handled)

Common Patterns

Fraud Detection

- name: suspicious_transaction_amount
  dataset: transactions
  type: max
  measure: amount
  condition: le
  threshold: 100000
  dimensions: [user_id]

Performance Monitoring

- name: max_response_time
  dataset: api_logs
  type: max
  measure: response_time_ms
  condition: le
  threshold: 30000  # 30 seconds max

Data Validation

- name: age_upper_bound
  dataset: users
  type: max
  measure: age
  condition: le
  threshold: 120  # Reasonable max age

System Limits

- name: max_file_size
  dataset: uploads
  type: max
  measure: file_size_bytes
  condition: le
  threshold: 104857600  # 100MB max

Performance Tips

1. Indexes: Ensure the measure column has appropriate indexes
2. Partitioning: Use partitioning for time-based queries
3. Filters: Apply filters to reduce data scanning
4. Statistics: Keep table statistics updated for query optimization

Alerting Scenarios

Max checks are particularly useful for alerting on:

• Anomalous spikes in transaction amounts
• Performance degradation (response times, queue lengths)
• Resource exhaustion (disk usage, memory consumption)
• Data quality issues (unrealistic values)

Related Checks

• Min - Minimum value validation
• Sum - Sum aggregation validation
• Numeric - Custom numeric expressions