> ## Documentation Index
> Fetch the complete documentation index at: https://docs.dqlabs.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# BigQuery Overview

> What the BigQuery connector covers: metadata, tags, lineage, usage, and how it authenticates.

Prizm connects to Google BigQuery to catalog your projects, datasets, tables, and columns, bring in tags and policy tags, build lineage, and report usage and performance. Once connected, that metadata appears in the Prizm catalog alongside your other sources.

## Why Connect BigQuery to Prizm?

**Catalog & Context** Prizm discovers projects, datasets, tables, and columns, including nested fields, and adds them to the Prizm catalog.

**Tags & Classification** Labels, policy tags (with their column mappings), resource tags, and Data Catalog tags and aspects are collected so governance context follows the data.

**Lineage** Prizm builds lineage from either the Google Data Lineage API (Dataplex) or BigQuery job logs, depending on the lineage mode you choose when you connect.

**Usage & Performance** Job history gives you who is querying what, how often, and how those queries perform.

## What Prizm Extracts

| Category | What Prizm collects |
| :- | :- |
| Metadata | Projects, datasets, tables, columns, and nested fields |
| Tags | Labels, policy tags and their column mappings, resource tags (keys and values), and Data Catalog tags and aspects |
| Lineage | Table-level lineage from the Data Lineage API or from job logs, depending on the lineage mode. Column-level lineage isn't covered yet |
| Usage and performance | Job history and timelines at the project or organization level |

See [What We Collect](/sources/bigquery/what-we-collect) for the source of each item.

## Authentication

Prizm supports two ways to authenticate to Google Cloud:

* **Service account key.** A JSON key file for a service account that has the roles Prizm needs. You upload it on the connection form, or retrieve it from a vault.
* **Workload Identity.** Keyless authentication using Google's identity federation. It appears on the connection form, but service account key is the method to use today. Workload Identity hasn't been verified end to end.

See [Setup](/sources/bigquery/setup) for the APIs to enable and the roles to grant.

## Lineage Sources

When you connect, you turn **Extract Lineage** on or off and choose a **Lineage Mode**:

* **Via DataPlex API** (default). Uses the Google Data Lineage API. It covers query, load, and copy jobs that ran after the API was enabled.
* **Job Logs.** Derives lineage from BigQuery job history. Use it if you can't enable the Data Lineage API, or you need history from before it was enabled. Source and destination tables are read from each job, so lineage from DML targets and from views is less complete.

> [!WARNING]
>
> The Data Lineage API records lineage only from the moment it is enabled in a project. It does not backfill earlier jobs. Enable it early. See [Setup](/sources/bigquery/setup#enable-the-required-apis).

## Next Steps

> **[Setup](/sources/bigquery/setup)**
>
> Enable APIs, grant roles, and connect BigQuery to Prizm.

> **[What We Collect](/sources/bigquery/what-we-collect)**
>
> The metadata, tags, lineage, and usage data Prizm extracts.

> **[FAQ](/sources/bigquery/faq)**
>
> Common questions about the BigQuery connector.

> **[Troubleshooting](/help/troubleshooting)**
>
> Diagnose connection, permission, and performance errors.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.