Google BigQuery + Go Fig
Integrate BigQuery data into Go Fig for powerful financial analytics at any scale.
BigQuery handles petabyte-scale analytics and is the default warehouse for many Google-centric data stacks. Go Fig connects natively, querying partitioned tables in place or replicating them into the Financial Intelligence Graph alongside your accounting and CRM data, so Celeste can reach across BigQuery and the rest of your finance stack without anyone hand-stitching extracts.
Key facts
- Auth
- Service-account JSON or Workload Identity Federation
- Access mode
- Query-in-place or replicated
- Read API
- Storage Read API used for high-throughput pulls
- Project scope
- Multi-project per connection
- Pricing model
- On-demand or reservation slots
What you can do with Google BigQuery data in Go Fig
Large-Scale Analysis
Analyze partitioned, multi-billion-row tables alongside accounting and CRM data without copying datasets out of BigQuery.
Marketing Attribution
Join BigQuery marketing event data (GA4, Ads exports) to revenue and cost data for true blended-channel ROI.
Cross-Cloud Joins
Combine BigQuery data with Snowflake, S3, or Azure-resident sources in one query layer without replicating between clouds.
Data available from Google BigQuery
Go Fig extracts and normalizes the following data from your Google BigQuery account:
How to connect Google BigQuery
Create a Go Fig service account
In Google Cloud IAM, create a service account and assign BigQuery Data Viewer on the datasets you want exposed plus BigQuery Job User on the project that will run queries. Generate and download the JSON key. Skip this step if you prefer Workload Identity Federation, which avoids handling key files entirely.
Choose datasets, projects, and partition strategy
Specify the project, datasets, and tables Go Fig should access. Partitioned tables are queried with partition pruning so you do not scan the full table on every sync. Clustered tables are also respected. Multi-project setups are supported in a single connection if your service account has access in each project.
Decide between on-demand and reservation pricing
On-demand pricing charges per TiB scanned and works well for moderate workloads. Reservation slots are flat-rate and predictable for heavy users. Go Fig's connector dashboard shows bytes scanned per query so finance can attribute BigQuery cost to specific reports and decide whether to move syncs into a reservation.
Schedule sync via the Storage Read API
For high-throughput pulls, Go Fig uses the BigQuery Storage Read API, which is optimized for streaming large reads at lower cost than standard query pricing. Incremental sync uses partition or last-modified columns to avoid full scans. Default cadence is hourly; near-real-time is available for streaming-inserted tables.
Authentication: Service-account JSON key with the BigQuery Data Viewer and BigQuery Job User roles scoped to the datasets you want exposed. Workload Identity Federation is supported for organizations that prohibit long-lived service-account keys, allowing Go Fig to assume a federated identity instead.
Common Questions About Google BigQuery Integration
Service account or Workload Identity Federation, which should I use?
Service-account JSON keys are the fastest path and are fine for most mid-market deployments. Workload Identity Federation is recommended if your security policy prohibits long-lived service-account keys, which is increasingly common in regulated environments. Federation lets Go Fig assume a short-lived credential via your IdP without ever holding a static key. Both paths use the same BigQuery IAM roles.
How does Go Fig avoid scanning entire tables on every sync?
Partition pruning. If your tables are partitioned by ingestion time, date, or a custom column, Go Fig issues queries that only touch the partitions changed since the last sync. Clustered tables further reduce bytes scanned. For tables without partitioning, an incremental sync column (such as updated_at) is used. Bytes scanned per query is visible in the connector dashboard so cost is auditable.
Does Go Fig use the BigQuery Storage Read API?
Yes, for high-throughput pulls. The Storage Read API is faster and lower cost than running standard SELECT queries for moving large volumes of data. It is used automatically for backfills and for replicated tables above a size threshold. Standard query API is used for small tables, views, and ad-hoc query-in-place lookups where the Storage Read API is not appropriate.
Can Go Fig query BigQuery views and materialized views?
Yes. Standard views, authorized views, and materialized views are all queryable. Authorized views are useful when you want to expose a curated slice of a sensitive table to Go Fig without granting access to the underlying table. Materialized views save query cost on aggregations you compute repeatedly. External tables backed by GCS, Sheets, or Bigtable are also accessible.
How is BigQuery cost controlled and attributed?
Every Go Fig query is tagged with a connector label that surfaces in BigQuery's INFORMATION_SCHEMA.JOBS view, so your data team can attribute cost down to the report or flow. The connector dashboard shows bytes scanned and estimated dollar cost per query. For predictable spend, point Go Fig's queries at a flat-rate reservation, which converts on-demand cost into a fixed slot allocation.
Ready to connect Google BigQuery?
See how your Google BigQuery data looks in Go Fig with a personalized demo.
Book a Demo