5-minute tutorial
Migrate Google BigQuery to Databricks in 60 Seconds
Learn how to copy your Google BigQuery data to Databricks with a single command using ingestr - no code required.
What you'll learn
Prerequisites
- Python 3.8 or higher installed
- Google Cloud project with BigQuery API enabled
- Service account with BigQuery Data Editor and Job User roles
- Downloaded service account JSON key file
- Dataset created in BigQuery (or permissions to create one)
- Databricks workspace with SQL endpoint
- Personal access token generated
- SQL endpoint running (not terminated)
- Appropriate permissions on catalog/schema
Step 1: Install ingestr
Install ingestr in seconds using pip. Choose the method that works best for you:
Recommended: Using uv (fastest)
# Install uv first if you haven't already
pip install uv
# Run ingestr using uvx
uvx ingestr
Alternative: Global installation
# Install globally using uv
uv pip install --system ingestr
# Or using standard pip
pip install ingestr
Verify installation: Run ingestr --version
to confirm it's installed correctly.
Step 2: Your First Migration
Let's copy a table from Google BigQuery to Databricks. This example shows a complete, working command you can adapt to your needs.
Set up your connections
Google BigQuery connection format:
bigquery://project-id?credentials_path=/path/to/service-account.json
Parameters:
- • project-id: Your Google Cloud project ID
- • credentials_path: Path to service account JSON key file
- • location: Optional dataset location (e.g., US, EU)
Databricks connection format:
databricks://token@host:port/http_path
Parameters:
- • token: Personal access token (use as username)
- • host: Workspace URL
- • port: Port number (usually 443)
- • http_path: SQL endpoint HTTP path
BigQuery Setup Required
Before running the command:
- Create a service account in Google Cloud Console
- Grant it BigQuery Data Editor and Job User roles
- Download the JSON key file
- Use the path to this file in your connection string
Run your first copy
Copy the entire users table from Google BigQuery to Databricks:
ingestr ingest \
--source-uri 'bigquery://my-project?credentials_path=/path/to/credentials.json' \
--source-table 'events' \
--dest-uri 'databricks://[email protected]:443/sql/1.0/endpoints/abc123' \
--dest-table 'raw.events'
What this does:
- • Connects to your Google BigQuery database
- • Reads all data from the specified table
- • Creates the table in Databricks if needed
- • Copies all rows to the destination
Command breakdown:
--source-uri
Your source database--source-table
Table to copy from--dest-uri
Your destination--dest-table
Where to write data
Step 3: Verify your data
After the migration completes, verify your data was copied correctly:
Check row count in Databricks:
-- Run this in Databricks
SELECT COUNT(*) as row_count
FROM raw.events;
-- Check a sample of the data
SELECT *
FROM raw.events
LIMIT 10;
Advanced Patterns
Once you've mastered the basics, use these patterns for production workloads.
Only copy new or updated records since the last sync. Perfect for daily updates.
ingestr ingest \
--source-uri 'bigquery://my-project?credentials_path=/path/to/credentials.json' \
--source-table 'public.orders' \
--dest-uri 'databricks://[email protected]:443/sql/1.0/endpoints/abc123' \
--dest-table 'raw.orders' \
--incremental-strategy merge \
--incremental-key updated_at \
--primary-key order_id
How it works: The merge strategy updates existing rows and inserts new ones based on the primary key. Only rows where updated_at
has changed will be processed.
Common Use Cases
Ready-to-use commands for typical Google BigQuery to Databricks scenarios.
Daily Customer Data Sync
Keep your analytics warehouse updated with the latest customer information every night.
# Add this to your cron job or scheduler
ingestr ingest \
--source-uri 'bigquery://my-project?credentials_path=/path/to/credentials.json' \
--source-table 'public.customers' \
--dest-uri 'databricks://[email protected]:443/sql/1.0/endpoints/abc123' \
--dest-table 'analytics.customers' \
--incremental-strategy merge \
--incremental-key updated_at \
--primary-key customer_id
Historical Data Migration
One-time migration of all historical records to your data warehouse.
# One-time full table copy
ingestr ingest \
--source-uri 'bigquery://my-project?credentials_path=/path/to/credentials.json' \
--source-table 'public.transactions' \
--dest-uri 'databricks://[email protected]:443/sql/1.0/endpoints/abc123' \
--dest-table 'warehouse.transactions_historical'
Development Environment Sync
Copy production data to your development Databricks instance (with sensitive data excluded).
# Copy sample data to development
ingestr ingest \
--source-uri 'bigquery://my-project?credentials_path=/path/to/credentials.json' \
--source-table 'public.products' \
--dest-uri 'databricks://[email protected]:443/sql/1.0/endpoints/abc123' \
--dest-table 'dev.products' \
--limit 1000 # Only copy 1000 rows for testing
Troubleshooting Guide
Solutions to common issues when migrating from Google BigQuery to Databricks.
Connection refused or timeout errors
Check your connection details:
- Verify the project ID matches your Google Cloud project
- Ensure the service account has necessary permissions
- Check if BigQuery API is enabled in your project
- Confirm the credentials file path is absolute, not relative
- Ensure SQL endpoint is running
- Verify personal access token is valid
- Check workspace URL is correct
- Confirm HTTP path matches your endpoint
Authentication failures
Common authentication issues:
- Verify the project ID matches your Google Cloud project
- Ensure the service account has necessary permissions
- Check if BigQuery API is enabled in your project
- Confirm the credentials file path is absolute, not relative
- Ensure SQL endpoint is running
- Verify personal access token is valid
- Check workspace URL is correct
- Confirm HTTP path matches your endpoint
Schema or data type mismatches
Handling data type differences:
- ingestr automatically handles most type conversions
- Google BigQuery: JSON/JSONB types become STRING (use JSON functions to query)
- Google BigQuery: Arrays are converted to JSON arrays
- Google BigQuery: TIMESTAMP types are preserved with timezone information
- Databricks: Delta tables support schema evolution
- Databricks: Complex types (arrays, maps, structs) supported
- Databricks: Photon acceleration for certain operations
- Databricks: Partitioning affects query performance
Performance issues with large tables
Optimize large data transfers:
- Use incremental loading to process data in chunks
- Run migrations during off-peak hours
- Split very large tables by date ranges using interval parameters
Ready to scale your data pipeline?
You've learned how to migrate data from Google BigQuery to Databricks with ingestr. For production workloads with monitoring, scheduling, and data quality checks, explore Bruin Cloud.