data-testing

Data Testing

Purpose

Implement automated data testing across the data platform: unit tests for dbt models, regression detection between environments, data quality validation with Soda/Great Expectations, and CI/CD gates for data pipeline changes.

Agent Protocol

Trigger

Exact user phrases: "data testing", "dbt testing", "dbt unit test", "data-diff", "datafold", "Soda", "Great Expectations", "data regression", "data quality test", "data contract test", "data testing framework", "data validation test", "test data pipeline".

Input Context

Before activating, verify:

• Data transformation tool (dbt, SQLMesh, custom)
• Testing infrastructure (Soda Cloud, Great Expectations, datafold, data-diff)
• CI/CD platform (GitHub Actions, GitLab CI, Jenkins)
• Environments (dev, staging, prod) and data sources
• Data contracts or SLAs in place
• Existing test coverage and failure patterns

Output Artifact

Test suite configuration with dbt unit tests, Soda checks, Great Expectations suites, data regression test configs, and CI pipeline integration.

Response Format

# dbt test config
# Soda checks

# Unit test fixtures and assertions
# Regression diff queries

# Great Expectations suite
# CI pipeline

No preamble. No postamble. No explanations. No filler/hedging/transitions. Compress output — why use many token when few do trick.

Completion Criteria

• dbt unit tests written for all critical transformation models
• Regression tests configured between staging and production
• Soda or Great Expectations checks for row count, freshness, null rates, uniqueness
• Data contract tests enforcing schema and shape on every deploy
• CI pipeline runs tests and blocks on failures
• Alerting on test failure with Slack/PagerDuty integration
• Test coverage measured and reported

Max Response Length

300 lines of code and configuration.

dbt Unit Testing

Unit Test Setup

# packages.yml
packages:
  - package: equalexperts/dbt_unit_testing
    version: 1.0.0