The short answer
BigQuery is Google's serverless cloud data warehouse. You load or query data with SQL; Google manages storage and compute. You typically pay for storage plus bytes scanned (on-demand) or for slot reservations.
GCS / Pub/Sub / Datastream
|
v
BigQuery datasets.tables
|
+--> Looker / BI / dbt / notebooksWhy it feels different from classic warehouses
- No cluster nodes to resize for basic use
- + separation of storage and compute
- Partitioning (often by date) and clustering cut scan cost
- External / federated queries can read GCS or other sources
Tiny cost habit
-- Prefer partitioned filters so you do not full-scan
SELECT COUNT(*)
FROM sales.fct_orders
WHERE order_date BETWEEN '2026-09-01' AND '2026-09-05';DE workflow
- 1Land raw files in GCS
- 2Load or query into BigQuery staging tables
- 3Transform with SQL/dbt into marts
- 4Govern with datasets and authorized views
Interview tip: "BigQuery = serverless GCP warehouse; cost follows scanned bytes unless you reserve slots." Mention partitioning/clustering as the fresher cost lever.