Output formats

Pick a format on the Delivery step of the generator. Every table in the job is written in that format. Large tables are split into parts rather than one huge file.

CSV

Comma-separated text with a header row. Nested values (arrays, structs, maps) are encoded as JSON in the cell. Binary values are lowercase hex; decimals are exact decimal strings.

JSONL

JSON Lines: one JSON object per row. Nested values stay structured. Binary values are lowercase hex; decimals are exact decimal strings, not JSON numbers.

Parquet

Columnar files that keep native types (including nested columns). This is usually the best choice if you will query the dataset in a lake or warehouse.

Where the files go — a zip download or your own bucket — is separate from the format. See cloud destinations.