def _backfill_staging_table()

in backfill/2024-03-23-clients_last_seen/telemetry_derived_clients_last_seen_v2_20240322/backfill_clients_last_seen_v2.py [0:0]

• 39 lines of code
• 4 McCabe index (conditional complexity)

def _backfill_staging_table(client, job_config, project_id, dataset, destination_table, bigquery_schema, submission_date, sample_id):
    """Backfill for a submission_date, sample_id combination."""
    full_table_id = f"{project_id}.{dataset}.{destination_table}_{sample_id}"
    try:
        table = client.get_table(full_table_id)
    except NotFound:
        table = bigquery.Table(full_table_id)
        table.schema = bigquery_schema
        table.time_partitioning = bigquery.TimePartitioning(
            type_=bigquery.TimePartitioningType.DAY,
            field='submission_date'
        )
        table.clustering_fields =['normalized_channel', 'sample_id']

    if not table.created:
        client.create_table(table)
        click.echo(f"Destination table {full_table_id} created.")
    else:
        client.update_table(table, ["schema"])

    if (
            partition := get_backfill_partition(
                submission_date,
                "submission_date",
                0,
                PartitionType.DAY
            )
        ) is not None:
        dest_table = f"{destination_table}_{sample_id}${partition}"

    print(f"Running a backfill for sample_id={sample_id} to {dest_table}")

    arguments = (
            ['query', '--use_legacy_sql=false', '--replace', '--project_id=moz-fx-data-shared-prod',
             '--format=none']
            + [f'--dataset_id={dataset}']
            + [f'--destination_table={dest_table}']
    )

    with tempfile.NamedTemporaryFile(mode="w+") as query_stream:
        query_stream.write(
            PARTITION_QUERY.format(submission_date=submission_date, sample_id=sample_id, full_table_id=full_table_id)
        )
        query_stream.seek(0)
        subprocess.check_call(["bq"] + arguments, stdin=query_stream)