Skip to Content
SourcesBigQueryOverview

Google BigQuery

Source (Direct Connection) — Vendo reads directly from your BigQuery datasets using a shared service account. No data is copied — queries run in your project.

What Vendo Accesses

Vendo connects to your BigQuery project and queries tables directly for use in attribution models, audience segments, and data models.

Supported Data

DataDescription
Any TableRead from any table or view in authorized datasets
Partitioned TablesSupports date-partitioned and ingestion-time partitioned tables
Nested/Repeated FieldsFull support for STRUCT and ARRAY columns
ViewsAuthorized views and materialized views

Connection Behavior

BehaviorDescription
Direct QueryNo data import — Vendo queries your tables in place
Service AccountUses a shared service account with read-only access
No Storage CostNo data duplication — queries run against your existing tables
Schema Auto-DetectionVendo discovers your table schemas automatically

Platform Details

SettingValue
Auth MethodService account key (JSON)
Connection TypeDirect connection (no sync)
APIBigQuery Storage Read API
BillingQueries billed to your GCP project

Common Use Cases

  • GA4 data — Query GA4 BigQuery exports for attribution and analytics
  • Custom datasets — Use existing warehouse tables as sources for models and segments
  • Cross-platform joins — Join BigQuery data with other Vendo sources
  • BQML models — Run predictions on existing tables

Setup

  1. Navigate to App Connections > Add App Connection
  2. Select Google BigQuery from the data warehouse category
  3. Upload your service account JSON key or paste it inline
  4. Grant the service account BigQuery Data Viewer and BigQuery Job User roles on your project
  5. Select which datasets Vendo should have access to
  6. Click Save — Vendo will auto-discover your tables

Permissions Required

The service account needs these IAM roles on your GCP project:

RolePurpose
roles/bigquery.dataViewerRead table data and metadata
roles/bigquery.jobUserExecute queries

Troubleshooting

  • “Permission denied” errors — Ensure the service account has the correct IAM roles on the specific datasets
  • Tables not appearing — Check that the service account has access to the dataset (dataset-level permissions, not just project-level)
  • Slow queries — Queries run in your project using your compute quota. Consider using partitioned tables and clustering for large datasets
Last updated on