Malloy Documentation

Currently, BigQuery, PostgreSQL, and DuckDB are supported. These instructions assume you have already installed the Malloy Extension in VSCode.

Adding the Connection in VS Code

NOTE: DuckDB is natively supported, allowing you to skip these initial steps.

  1. Open or create a file that ends in .malloy. In the explorer view (the top left icon that looks like files), you'll see a section called "MALLOY CONNECTIONS".

  2. In the "MALLOY CONNECTIONS" panel, select "Edit Connections". This opens the connection manager page.

  3. Click "Add Connection" and fill out the relevant details. See below for database-specific instructions.

  4. Press "Test" on the connection to confirm that you have successfully connected to the database

  5. Hit "Save," then dive into writing Malloy! We recommend starting with one of our Samples, which can be found here


Authenticating to BigQuery can be done either via OAuth (using your Google Cloud Account) or with a Service Account Key downloaded from Google Cloud

Option 1: OAuth using gcloud

To access BigQuery with the Malloy Extension, you will need to have a Google Cloud Account, access to BigQuery, and the gcloud CLI installed. Once the gcloud CLI is installed, open a terminal and type the following:

gcloud auth login --update-adc
gcloud config set project {my_project_id} --installation

Replace {my_project_id} with the ID of the BigQuery project you want to use & bill to. If you're not sure what this ID is, open Cloud Console, and click on the dropdown at the top (just to the right of the "Google Cloud Platform" text) to view projects you have access to. If you don't already have a project, create one.

When creating the connection in the VS Code Plugin, you can leave the optional fields blank as it will connect using your gcloud project configuration.

NOTE: The Malloy Extension used the BigQuery Node SDK, which does its best guess at finding credentials stored on your device - it looks in environment variables, and also in places that gcloud is known to store application credentials. There is not necessarily a 1:1 mapping of how gcloud is authenticated and how the BigQuery SDK will authenticate.

Option 2: Service Account

Add the relevant account information to the new connection, and include the path to the service account key.


Add the relevant database connection information. Once you click save, the password (if you have entered one) will be stored in your system keychain.

Parquet and CSV Files (via DuckDB)

Parquet and CSV files are queryable via DuckDB is available without needing to explicitly configure a connection. Local files can be referenced in a source. This example has the CSV in the same directory as the .malloy model file: source: baby_names is duckdb.table('babynames.csv')

There are a number of examples on public data using DuckDB available here.