schemachange
Deploying and managing Snowflake database objects using version control with schemachange. Use this skill when you need to manage database migrations for objects not handled by dbt, implement CI/CD pipelines for schema changes, or coordinate deployments across multiple environments.
When & Why to Use This Skill
This Claude skill facilitates the automated deployment and management of Snowflake database objects using a version-controlled, migration-based approach with schemachange. It solves the challenges of manual schema updates and environment inconsistency by providing a structured framework for database CI/CD pipelines, specifically for objects not managed by tools like dbt.
Use Cases
- Automating Snowflake database migrations within CI/CD pipelines (e.g., GitHub Actions, Azure DevOps) to ensure repeatable and trackable deployments.
- Managing complex Snowflake objects such as Stored Procedures, User-Defined Functions (UDFs), and Stages that require version control outside of standard data modeling tools.
- Ensuring environment parity by coordinating synchronized schema changes across Development, Staging, and Production Snowflake accounts.
- Implementing idempotent database updates using 'Repeatable' and 'Always' scripts to maintain consistent state for views and permissions without data loss.
| name | schemachange |
|---|
Schemachange
Deploy and manage Snowflake database changes using version control and CI/CD pipelines with schemachange's migration-based approach.
Quick Start
Script Types:
- V__ (Versioned) - One-time structural changes (run exactly once)
- R__ (Repeatable) - Objects that can be safely recreated (runs when new/modified)
- A__ (Always) - Scripts that run every deployment (must be idempotent)
Script Naming
Versioned Scripts (V__)
V1.0.0__initial_setup.sql
V1.1.0__create_base_tables.sql
V2.0.0__restructure_schema.sql
Use for: CREATE TABLE, ALTER TABLE, CREATE SCHEMA
Repeatable Scripts (R__)
R__Stage_01_create_views.sql
R__Stage_02_alter_procedures.sql
R__Stage_03_utility_functions.sql
Use for: CREATE OR ALTER VIEW, CREATE OR ALTER PROCEDURE, CREATE OR REPLACE STREAM
Always Scripts (A__)
A__refresh_permissions.sql
A__update_config_values.sql
Use for: Jobs that must run every deployment (idempotent only)
Key Concepts
Execution Order
- Versioned scripts (V__) in numeric order
- Repeatable scripts (R__) in alphabetic order (use naming to control)
- Always scripts (A__) in alphabetic order
CREATE OR ALTER vs CREATE OR REPLACE
- CREATE OR ALTER - Preserves data, tags, policies, grants (preferred)
- CREATE OR REPLACE - Drops and recreates (loses metadata)
See CREATE_OR_ALTER_REFERENCE.md for supported objects.
Configuration
# schemachange-config.yml
root-folder: migrations
create-change-history-table: true
connection-name: default
change-history-table: MY_DB.SCHEMACHANGE.CHANGE_HISTORY
Deployment
# Dry run
schemachange deploy --config-folder . --dry-run
# Deploy
schemachange deploy --config-folder .
# With variables
schemachange deploy --vars '{"env":"prod","schema":"data"}'
Resources
schemachange-config.yml- Complete configuration templateSCRIPT_PATTERNS.md- Examples for V**, R**, A__ scripts (coming soon)CREATE_OR_ALTER_REFERENCE.md- Supported object types (coming soon)CI_CD_EXAMPLES.md- GitHub Actions and Azure DevOps patterns (coming soon)