Warehouses

Connect BigQuery

Connect BigQuery with your Google account, choose read-only access, and pick the project that runs and pays for each query.

On this page

BigQuery connects with a Google account, in Google's own sign-in window. Ahamo's AI can then list your projects, datasets and tables, and run SQL on them.

Availability. The BigQuery connector must be on for your workspace. See Connect a data warehouse.

Before you connect

The Google account needs two roles from your project's admin:

  • BigQuery Job User in the project that runs the queries;
  • BigQuery Data Viewer on the datasets to read.

The project that runs a query also pays for it, so it needs billing, or must stay within BigQuery's free tier.

Connect

  1. Open Connectors, find BigQuery and choose Connect.
  2. Under the permissions, View your data in BigQuery is required. Leave View and manage your data in BigQuery off unless Ahamo should also change data.
  3. Name the connection, choose who may use it, and continue.
  4. Sign in to Google in the window that opens and allow the access it asks for.
  5. Wait for Connected to BigQuery. It works.

With only View your data in BigQuery, the connection can only read, and Ahamo's AI runs queries without asking. With View and manage your data in BigQuery, you approve each query in a card.

Ask a question

For example, ask this in a chat.

How many orders did each city have this year? Use the shop dataset in BigQuery.

When you have more than one project, Ahamo asks which one runs and pays for the query.

Tables are named project.dataset.table. A dataset's location, such as EU or US, is fixed, and a query joins only tables in one location.

Limits

A query may bill at most 10 GiB. A query above that is refused and costs nothing, so ask for a narrower one, such as fewer columns or a shorter period. BigQuery bills by the columns a query reads.

Every query carries your Ahamo id as the label ahamo_person, so it shows in BigQuery's own job history. Deleting datasets or tables, and changing who can see a dataset, are always refused.

Troubleshooting

Google shows a warning that the app isn't verified yet. Choose Advanced, then the link to continue. An admin can remove the warning for everyone with your organisation's own Google sign-in app. See Use your organisation's own sign-in app.

Your Google Workspace admin blocks apps like this one. Ask your Google Workspace admin to trust Ahamo under Security › Access and data control › API controls › Manage App Access.

This Google account may not run queries in that project. Ask the project's admin for BigQuery Job User, or run the query in a project where the account has it.

BigQuery can't find that project, dataset or table for this account. Check the names and the dataset's location.

Google ended this connection, for example after a password change. Open the connection and choose Sign in again.

This connection may only read BigQuery data. Something asked for a change. If Ahamo should change data, open the connection, choose Change permissions, tick View and manage your data in BigQuery and choose Reconnect with these.

On this page