tech_subscriptions_data_contract.yml

dataset: datasource/database/schema/tech_subscriptions

variables:
  FRESHNESS_HOURS:
    default: 24

checks:
  - schema:
      allow_extra_columns: false
      allow_other_column_order: false
  - row_count:
      threshold:
        must_be_greater_than: 0
  - freshness:
      column: start_date
      threshold:
        unit: hour
        must_be_less_than_or_equal: ${var.FRESHNESS_HOURS}
  - failed_rows:
      name: "Subscription start_date must not be after end_date"
      expression: end_date IS NOT NULL AND start_date > end_date
  - failed_rows:
      name: "Cancelled or expired subscriptions must have an end_date"
      expression: (billing_status IN ('CANCELLED', 'EXPIRED') AND end_date IS NULL)
  - failed_rows:
      name: "Active subscriptions should not have an end_date in the past"
      expression: (billing_status = 'ACTIVE' AND end_date IS NOT NULL AND end_date < ${soda.NOW})

columns:
  - name: subscription_id
    data_type: varchar
    checks:
      - missing:
      - duplicate:
  - name: customer_id
    data_type: varchar
    checks:
      - missing:
  - name: plan_id
    data_type: varchar
    checks:
      - missing:
  - name: start_date
    data_type: date
    checks:
      - missing:
  - name: end_date
    data_type: date
  - name: billing_status
    data_type: varchar
    checks:
      - missing:
      - invalid:
          name: "Allowed billing statuses"
          valid_values:
            - TRIALING
            - ACTIVE
            - PAST_DUE
            - CANCELLED
            - EXPIRED
  - name: currency
    data_type: varchar
    checks:
      - missing:
      - invalid:
          name: "Currency must be ISO-4217-like (3 uppercase letters)"
          valid_format:
            name: ISO-4217 code
            regex: "^[A-Z]{3}$"