[Python] Performance difference during fetching between pandas-gbq and bigquery_storage api in...

I can fetch data from gbq using two methods:

1. 
   df = pd.io.gbq.read_gbq(
   query,
   project_id=project_id
   use_bqstorage_api=True,
   credentials=credentials,
   configuration=dict(
   query=dict(
   maximumBytesBilled=maximum_bytes_billed
   ),
   labels=labels
   )
   )
2. 
   client = bigquery.Client(
   project=project_id,
   credentials=credentials
   )
   bq_storage_client = bigquery_storage.BigQueryReadClient(credentials=credentials)
   job_config = bigquery.QueryJobConfig(
   use_legacy_sql=False,
   maximum_bytes_billed=maximum_bytes_billed
   )
   query_job = client.query(
   query,
   job_config=job_config
   )
   df = query_job.to_dataframe(
   bqstorage_client=bq_storage_client
   )

Is there any advantage of using pandas-gbq over plain bigquery_storage API? For now, I've noticed, that pandas-gbq is slower and provides less flexibility. However, maybe there is something that I'm missing here.

When should I use method 1, and when method 2 is better?

Continue reading...