Skip to content
This repository has been archived by the owner on Dec 15, 2023. It is now read-only.


Folders and files

Last commit message
Last commit date

Latest commit



43 Commits

Repository files navigation

⚠️ THIS SUITE IS NO LONGER THE RECOMMENDED WAY TO TEST ADAPTERS. If you are building and testing a new dbt adapter, please read instead: "Testing a new adapter"

Installation and use

pip install pytest-dbt-adapter

You'll need to install this package with pip install pytest-dbt-adapter and write a specfile, which is a yaml file ending in .dbtspec. See the included spark/postgres examples in specs. You can also write custom test sequences and override existing default projects.

After installing this package, you should be able to run your spec with pytest path/to/mytest.dbspec. You'll need dbt-core and your adapter plugin installed in the environment as well.

This package also includes a module named dbt_adapter_tests that includes helpers for writing integration tests with Python if necessary. For maintainability purposes, this should only be used as a last resort for scenarios that are impossible to capture with a specfile.


A spec is composed of a minimum of two things:

  • a target block
  • a sequences block
    • The keys are test names. You can select from these names with pytest's -k flag.
    • The values are test sequence definitions.

Optionally, there is also:

  • a projects block


A target block is just like a target block you'd use in dbt core. However, there is one special change: the schema field should include a {{ var('_dbt_random_suffix') }} somewhere that the test suite will insert.


A sequence has a name (the sequence name), a project (the project name to use), and sequence (a collection of test steps). You can declare new sequences inline, or use the name of a builtin sequence. A sequence itself is just a list of steps. You can find examples in the form of the builtin sequences in the sequences/ folder.

You are encouraged to use as many sequences as you can from the built-in list without modification.


The minimum project contains only a name field. The value is the name of the project - sequences include a project name.

A project also has an optional paths block, where the keys are relative file paths (to a dbt_project.yml that will be written), and the values are the contents of those files.

There is a dbt_project_yml block, which should be a dictionary that will be updated into the default dbt_project.yml (which sets name, version, and config-version).

Instead of declaring a name field, a project definition may have an overrides field that names a builtin project. The test suite will update the named builtin project with those overrides, instead of overwriting the full project with a new one.