jeremyyeo/dbt-sandcastles

dbt-sandcastles

For testing dbt with various environments...

.
├── terraform/
│   ├── aws        # For managing AWS resources.
│   └── gcp        # For managing GCP resources.
├── dbt/
│   ├── bigquery   # dbt project using BigQuery.
│   ├── redshift   # dbt project using Redshift.
│   └── snowflake  # dbt project using Snowflake.
└── ...

Setup

AWS

Google Cloud Platform

Requirements:

• A valid GCP project.
• The following GCP APIs enabled:
  • BigQuery API.
  • BigQuery Connection API.
  • Cloud SQL Admin API.
• A service account and it's key (a json file) downloaded to your machine that has access to the above project.

Terraform

Terraform is used for standing up and tearing down cloud resources. To install it, run the following:

curl -sS https://releases.hashicorp.com/terraform/1.0.7/terraform_1.0.7_darwin_amd64.zip > terraform.zip &&
unzip terraform.zip &&
mv terraform /usr/local/bin/

Within the terraform folder, create a terraform.tfvars file with the following variables (replace the values with your actual ones):

gcp_project              = "my-gcp-project-id"
gcp_service_account_file = "/my-local/path/to/the/gcp_service_account_file.json"

And then initialize terraform with terraform init. After initialization, you can now provision a Cloud SQL PostgreSQL instance and a BigQuery dataset with:

terraform plan    # To check what Terraform will attempt to do.
terraform apply   # To actually apply the plan above.
terraform output  # To print some variables out such as the public ip address of the postgres database.

Note that the PostgreSQL database will have some dummy data on it, see terraform/scripts/database_setup.sql for more information or to customize it yourself.

Once you are done doing what you need to do, simply destroy everything with terraform destroy. There is a chance that terraform will fail to destroy the stack due to an inability to actually delete the google_sql_user resource, in this case, simply remove the user from the state file with terraform state rm module.gcp.google_sql_user.root_user before attempting to destroy again.

dbt

Install dbt into a Python virtual environment:

python3 -m venv venv
source venv/bin/activate
pip install dbt
dbt --version

Setup your ~/.dbt/profiles.yml file and navigate to the bigquery folder to start running dbt commands on your BigQuery dataset.