Loading data

Status: shipped.

Load jobs accept CSV, JSON, Parquet, Avro, and ORC, with gzip, zstd, and snappy compression. Source URIs may be local file paths or gs://… URIs (resolved under BQEMU_GCS_LOCAL_ROOT). writeDisposition supports WRITE_APPEND, WRITE_TRUNCATE, WRITE_EMPTY.

job_config = bigquery.LoadJobConfig(
    source_format=bigquery.SourceFormat.PARQUET,
    write_disposition=bigquery.WriteDisposition.WRITE_TRUNCATE,
)
load_job = client.load_table_from_uri(
    "file:///tmp/orders.parquet",
    "my-project.sales.orders",
    job_config=job_config,
)
load_job.result()

Tracking: issue.