specify geographical location of table in pandas read_gbq

[rmporsch]

I am having trouble setting the geographical location in pandas.read_gbq.

In google.cloud.bigquery I can set the location to asia-northeast1 the following:

from google.cloud import bigquery
bigquery_client = bigquery.Client(project_name)    
query_job = bigquery_client.query(query, location='asia-northeast1')

However, I fail to do this in read_gbq. The following seems not to work:

from pandas_gbq import gbq    
dat = gbq.read_gbq(query, project_id=project_name, jobReference={'location':'asia-northeast1'}, dialect='standard')

[max-sixty]

Thanks for the report!

I'm not familiar with this - what's the error you get when setting this? And when it's not set?

[tswast]

@rmporsch You're correct that this requires modifications to pandas-gbq.

Note: the way I would expect this to work (but it doesn't) with the current library is by using the configuration argument to read_gbq.

import pandas_gbq
pandas_gbq.read_gbq(
    "SELECT name FROM tokyo_dataset.us_states WHERE post_abbr LIKE 'W%'",
    configuration={'jobReference':{'location':'asia-northeast1'}},
    project_id='swast-scratch')

but this fails with GenericGBQException: Reason: 404 Not found: Table swast-scratch:tokyo_dataset.us_states. Please verify that the table exists and the correct location was used for the job.. I believe because the jobReference key gets overwritten by google-cloud-bigquery when no job_id is passed in.

[tswast]

I think the proper fix here is to add a location parameter to read_gbq and to_gbq so that we can pass it on to the google-cloud-bigquery library when we run the query or load job.

Aside: I'd like to make this less verbose once there is a way to set a default location in the BigQuery client object. googleapis/google-cloud-python#5148

[tswast]

Once #185 is in, we'll need to

• Release pandas-gbq 0.5.0.
• Update pandas to add the location parameter to DataFrame.to_gbq and pandas.read_gbq.

[tswast]

Version 0.5.0 is now released to PyPI with this fix. Next steps are to add the location parameters to Pandas.