Fetching interval data to arrow (and thus pandas) doesn't quite work using the REST API

[jimfulton]

Given:

create table i (v interval)

insert into i (v) values (make_interval(1, 2, 3, 4, 5, 6))

Trying to fetch data as arrow:

>>> client.query("select * from i").to_arrow(create_bqstorage_client=False)
/home/jim/p/g/python-bigquery/google/cloud/bigquery/_pandas_helpers.py:244: UserWarning: Unable to determine type for field 'v'.
  warnings.warn("Unable to determine type for field '{}'.".format(bq_field.name))
pyarrow.Table
v: string

This is because BQ_TO_ARROW_SCALARS doesn't have an entry for INTERVAL.

I think we should stamp the array with an interval extension type, and convert timedelta when we covert to pandas.

Also #949

[tswast]

Duplicate of #836