What’s supported
Supported actions
This connector supports:- Read Actions, including full historic backfill and incremental read.
- Proxy Actions, using the base URL
https://bigquery.googleapis.com.
Supported objects
The BigQuery connector can read from any table in the specified dataset. Tables are referenced by name in theobjectName field of your manifest (e.g., my_table).
For each table, the connector supports:
- All columns, subject to the field limit per read operation.
- Standard BigQuery types: STRING, INT64, FLOAT64, BOOL, TIMESTAMP, DATE, BYTES, NUMERIC, RECORD, and more.
Limits
- Maximum fields per read: 7. BigQuery is a columnar store, so reading additional columns is significantly more expensive.
- Default page size: 50,000 rows per page.
- Timestamp column required: Every table you read must have a
TIMESTAMPorDATETIMEcolumn for incremental reads and backfill windowing.
Example integration
To define an integration for BigQuery, create a manifest file that looks like this:Using the connector
This connector uses a service account key for authentication, which means that you do not need to set up a Provider App before getting started. (Provider apps are only required for providers that use OAuth2 Authorization Code grant type.) To start integrating with BigQuery:- Create a manifest file like the example above.
- Deploy it using the amp CLI.
- If you are using Read Actions, create a destination.
- Embed the InstallIntegration UI component. The UI component will prompt the customer for their service account key, GCP project ID, dataset name, and timestamp column.
- Start using the connector!
- If your integration has Read Actions, you’ll start getting webhook messages.
- If your integration has Proxy Actions, you can start making Proxy API calls. Please note that this connector’s base URL is
https://bigquery.googleapis.com.

