materializeinc/dbt-action

A GitHub Action for deploying to Materialize with dbt.

Materialize dbt action

A GitHub action that runs dbt commands against a Materialize instance.

Required secrets

The following secrets are required:

Name	Description	Example
MZ_PASS	The password for the Materialize user.	MZ_PASS: ${{ secrets.MZ_PASS }}
MZ_USER	The username for the Materialize user.	MZ_USER: ${{ secrets.MZ_USER }}
MZ_HOST	The host for the Materialize instance.	MZ_HOST: ${{ secrets.MZ_HOST }}

Variables

You can override the default values for the following variables:

Name	Description	Default
DBT_MZ_VERSION	dbt-materialize adapter version to install	latest
MZ_DATABASE	The database for the Materialize instance	materialize
MZ_SCHEMA	The schema for the Materialize instance	public
MZ_CLUSTER	The cluster for the Materialize instance	default
DBT_PROJECT_FOLDER	The dbt project folder	.

Usage

Create a new workflow file in your repository, for example .github/workflows/materialize-dbt-action.yml and add the following content:

name: Materialize dbt action
on:
  push:
    branches:
      - main

jobs:
  build:
    runs-on: ubuntu-latest
    steps:
    - uses: actions/checkout@v2
      with:
        fetch-depth: 0
    - name: Materialize dbt action
      uses: MaterializeInc/dbt-action@main
      with:
        dbt_command: "dbt --version --profiles-dir /"
      env:
        DBT_MZ_VERSION: 1.3.4
        MZ_PASS: ${{ secrets.MZ_PASS }}
        MZ_USER: ${{ secrets.MZ_USER }}
        MZ_HOST: ${{ secrets.MZ_HOST }}
        MZ_DATABASE: materialize
        MZ_SCHEMA: public
        MZ_CLUSTER: default
        DBT_PROJECT_FOLDER: ${{ env.DBT_PROJECT_FOLDER }}

The above will run the dbt --version command against your Materialize instance on every push to the main branch.

To run on a pull request, change the on section to:

on:
  pull_request:
    branches:
      - main

Inputs

dbt_command

Required The dbt command to run. Default "dbt debug". You can change this to run any dbt command.

Overriding the default profiles.yml

You can override the default profiles.yml by creating a new file in your repository called profiles.yml and add the following content:

default:
  outputs:
    default:
      type: materialize
      threads: 1
      user: "{{ env_var('MZ_USER')}}"
      pass: "{{ env_var('MZ_PASS')}}"
      host: "{{ env_var('MZ_HOST')}}"
      port: 6875
      database: materialize
      schema: public
      cluster: default
      sslmode: require

  target: default

Then change the dbt_command input to dbt --version --profiles-dir . if you have a profiles.yml file in your repository. The ${DBT_PROJECT_FOLDER} will be added to the --profiles-dir argument.

This will override the default profiles.yml file and use the one from your repository.

DBT_PROJECT_FOLDER

You can specify a custom dbt project folder by setting the DBT_PROJECT_FOLDER environment variable. This is useful if you have multiple dbt projects in the same repository.

dbt-materialize version

By default the action will install the latest version of the dbt-materialize adapter. You can specify a specific version by setting the DBT_MZ_VERSION environment variable. For example:

env:
  DBT_MZ_VERSION: 1.3.2

dbt-action's Issues

Support pulling dbt version from dependency management config files

dbt is distributed as a Python package, which means that it's common for dbt users to have a Python dependency management system in place, like Conda or Poetry. We should support pulling the dbt version from a configuration file (e.g. poetry.lock), if one is provided in the repo.

This feature was requested by a user in this thread. For an example of a real dbt deployment using Poetry, see our internal analytics pipeline.

Publish to Github Actions Marketplace

Once ready, we should begin publishing to the Github Actions marketplace.