This is a proof of concept implementing schemachange in Terraform. Having managed Snowflake using both, I thought it would be an interesting exercise to try and bring the two together.

The meat of the implementation is in terraform/sql.tf. The schemachange-esque change scripts can be found under sql/.

Not looking to spend a lot of time to match the functionality exactly. Do not use this in production.

1. Set up the Snowflake CLI.

   1. Create a virtual environment.

      python -m venv .venv

   2. Activate the virtual environment.

      source ./.venv/bin/activate

   3. Install the Snowflake CLI.

   4. Configure the Snowflake CLI. Call the connection default.

2. Run Terraform. (-parallelism=1 is set to match schemachange's behavior of only running a single script at a time.)

   cd terraform
   terraform init
   terraform apply -parallelism=1

• When files change, it shows a diff in the plan.
• You see the result of each query.
• It can be integrated with other Terraform resources.
• You could add dependencies between scripts/resources through depends_on.

• At time of writing, the Snowflake CLI doesn't support MFA caching.

There's definitely more that should be in this list.

• [x] Support versioned scripts
  • [ ] Ignore changes to the filenames after the version number
  • [ ] Limit to scripts with a V followed by digit(s) — currently considers any file with a V followed by anything as a versioned script
  • [ ] Run in numeric order
• [x] Support repeatable scripts
• [x] Support always scripts
• [x] Match the order between script types
• [x] Ignore script folders