Connect Soda to GCP BigQuery

Last modified on 01-Feb-24

For Soda to run quality scans on your data, you must configure it to connect to your data source.
To learn how to set up Soda and configure it to connect to your data sources, see Get started.

Connection configuration
Authentication methods
Test the datasource connection
Supported data types
Use a file reference for a BigQuery data source connection
Troubleshoot

A note about BigQuery datasets: Google uses the term dataset slightly differently than Soda (and many others) do.

• In the context of Soda, a dataset is a representation of a tabular data structure with rows and columns. A dataset can take the form of a table in PostgreSQL or Snowflake, or a DataFrame in a Spark application.
• In the context of BigQuery, a dataset is “a top-level container that is used to organize and control access to your tables and views. A table or view must belong to a dataset…”

Instances of “dataset” in Soda documentation always reference the former.

Connection configuration reference

Install package: soda-bigquery

# Service Account Key authentication method
# See Authentication methods below for more config options
data_source my_datasource_name:
  type: bigquery
  account_info_json: '{
      "type": "service_account",
      "project_id": "gold-platform-67883",
      "private_key_id": "d0121d000000870xxx",
      "private_key": "-----BEGIN PRIVATE KEY-----\n...\n-----END PRIVATE KEY-----\n",
      "client_email": "abc333@project.iam.gserviceaccount.com",
      "client_id": "XXXXXXXXXXXXXXXXXXXX.apps.googleusercontent.com",
      "auth_uri": "https://accounts.google.com/o/oauth2/auth",
      "token_uri": "https://accounts.google.com/o/oauth2/token",
      "auth_provider_x509_cert_url": "https://www.googleapis.com/oauth2/v1/certs",
      "client_x509_cert_url": "https://www.googleapis.com/robot/v1/metadata/x509/..."
    }'
  auth_scopes:
  - https://www.googleapis.com/auth/bigquery
  - https://www.googleapis.com/auth/cloud-platform
  - https://www.googleapis.com/auth/drive
  project_id: "platinum-platform-67883"
  dataset: sodacore

Authentication methods

Using GCP BigQuery, you have the option of using one of several methods to authenticate the connection.

1. Application Default Credentials
2. Application Default Credentials with Service Account impersonation
3. Service Account Key (see connection configuration above)
4. Service Account Key with Service Account Impersonation

Application Default Credentials

Add the use_context_auth property to your connection configuration, as per the following example.

data_source my_datasource:
  type: bigquery
  ...
  use_context_auth: True

Application Default Credentials with Service Account impersonation

Add the use_context_auth and impersonation_account properties to your connection configuration, as per the following example.

data_source my_datasource:
  type: bigquery
  ...
  use_context_auth: True
  impersonation_account: <SA_EMAIL>

Service Account Key with Service Account impersonation

Add the impersonation_account property to your connection configuration, as per the following example.

data_source my_database_name:
  type: bigquery
  ...
  account_info_json: '{
      "type": "service_account",
      "project_id": "...",
      "private_key_id": "...",
    ...}'
  impersonation_account: <SA_EMAIL>

Test the data source connection

To confirm that you have correctly configured the connection details for the data source(s) in your configuration YAML file, use the test-connection command. If you wish, add a -V option to the command to returns results in verbose mode in the CLI.

soda test-connection -d my_datasource -c configuration.yml -V

Supported data types

Use a file reference for a BigQuery data source connection

If you already store information about your data source in a JSON file in a secure location, you can configure your BigQuery data source connection details in Soda Cloud to refer to the JSON file for service account information. To do so, you must add two elements:

• volumes and volumeMounts parameters in the values.yml file that your Soda Agent helm chart uses
• the account_info_json_path in your data source connection configuration

You, or an IT Admin in your organization, can add the following scanlauncher parameters to the existing values.yml that your Soda Agent uses for deployment and redployment in your Kubernetes cluster. Refer to Deploy using a values YAML file for details.

soda:
  scanlauncher:
    volumeMounts:
      - name: gcloud-credentials
        mountPath: /opt/soda/etc
    volumes:
      - name: gcloud-credentials
        secret:
          secretName: gcloud-credentials
          items:
            - key: serviceaccount.json
              path: serviceaccount.json

Use the following command to add the service account information to a Kubernetes secret that the Soda Agent consumes according to the configuration above; replace the angle brackets and the values in them with your own values.

kubectl create secret generic -n <soda-agent-namespace> gcloud-credentials --from-file=serviceaccount.json=<local path to the serviceccount.json>

After you make both of these changes, you must redeploy the Soda Agent. Refer to Deploy using a values YAML file for details.

Adjust the data source connection configuration to include the account_info_json_path configuration, as per the following example.

my_datasource_name:
  type: bigquery
  account_info_json_path: /opt/soda/etc/serviceaccount.json
  auth_scopes:
  - https://www.googleapis.com/auth/bigquery
  - https://www.googleapis.com/auth/cloud-platform
  - https://www.googleapis.com/auth/drive
  project_id: ***
  dataset: sodalibrary

Troubleshoot

Problem: When running a scan, you encounter an error that reads, 400 Cannot query over table 'event_logs' without a filter over column(s) 'serverTimestamp' that can be used for partition elimination.

Workaround: The error occurs because the table in BigQuery is configured to require partitioning.

• If the error occurs when you are profiling your data with Soda, you must disable profiling.
• If the error occurs when the scan is executing regular SodaCL checks, be sure you always apply a filter on serverTimestamp. See Dataset filters

Was this documentation helpful?

What could we do to improve this page?

• Suggest a docs change in GitHub.
• Share feedback in the Soda community on Slack.

Documentation always applies to the latest version of Soda products
Last modified on 01-Feb-24

Sign up for Soda