A data scientist wants to run SQL in BigQuery over Parquet files that remain in a Cloud Storage bucket, with no copy made into BigQuery storage. Which table type supports this?
Show answer and explanation
Answer: D
A BigLake table points BigQuery at files in Cloud Storage, such as Parquet, ORC, Avro, or CSV. The data stays in the bucket and is queried in place, while BigLake adds unified access controls. Views and materialized views operate on data already in BigQuery, and federated queries target external database services rather than files in a bucket.
Why the other options are wrong
- A. An authorized view restricts which columns or rows a user sees from a table that already lives in BigQuery. It does not expose files sitting in a bucket.
- B. Federated queries reach operational databases such as Cloud SQL or Spanner. They are not the mechanism for reading files stored in Cloud Storage.
- C. A materialized view stores precomputed results from a BigQuery table, so it still depends on data already loaded into BigQuery storage.