In the Google Cloud console open BigQuery, go to Explorer, expand the project, open Datasets, select the dataset, open Overview then Tables, select the table, then click Export / sync then Cloud Storage, pick the destination, format and compression and click Save. On the command line the same job is bq extract --destination_format FORMAT --compression TYPE project:dataset.table gs://bucket/filename. Formats are CSV and JSON with GZIP, Avro with DEFLATE or SNAPPY, and Parquet with SNAPPY, GZIP or ZSTD. Reading the data needs roles/bigquery.dataViewer or higher, running the extract job needs roles/bigquery.user or roles/bigquery.jobUser, and writing to the bucket needs roles/storage.objectAdmin. BigQuery writes at most 1 GB of logical data to a single file, so anything larger needs a wildcard in the filename to shard across files, and it cannot export to a local file, Google Sheets or Drive. Nested and repeated data cannot go to CSV.
What it holds
Dataset tables and their rows, including partitioned and clustered tables.
Find the size
Query INFORMATION_SCHEMA.TABLE_STORAGE, a snapshot of current storage for the project, one row per table: TOTAL_LOGICAL_BYTES for the logical size and TOTAL_ROWS for the row count. TABLE_STORAGE_BY_FOLDER and TABLE_STORAGE_BY_ORGANIZATION give the same figures at wider scope.
What to enter on the form
Google BigQuery in GB, and the year your company started using it. Seller form