TST: skipping a gbq test for now

[jreback]

I am xfailing most of these tests for now till we have the testing sorted: d9e75c7

[parthea]

At first glance, it appears that the test started failing 3 days ago. The test was written based on the assumption that BigQuery's streaming system caches the table schema for up to 2 minutes. This is mentioned in the following stackoverflow post.
http://stackoverflow.com/questions/25279116/cannot-insert-new-value-to-bigquery-table-after-updating-with-new-column-using-s

This assumption is documented in gbq.py.

pandas/pandas/io/gbq.py

Line 569 in 7d40f18

# <https://code.google.com/p/google-bigquery/issues/detail?id=191>`__

It appears that this assumption is no longer valid. Another user mentioned a similar issue in Google BigQuery Issue 191:

Our team has noticed that updates are taking over 2 minutes now (more like ~10 minutes) as of the past 12 hours.

pandas.io.gbq.to_gbq() will also fail for end users under the following condition:

• Users set the if_exists parameter of pandas.io.gbq.to_gbq() to 'replace'
• The schema differs from the remote BigQuery schema
• The BigQuery system cache is greater than 2 minutes (i.e. change in schema is not reflected in BigQuery).

    if_exists : {'fail', 'replace', 'append'}, default 'fail'
        'fail': If table exists, do nothing.
        'replace': If table exists, drop it, recreate it, and insert data.
        'append': If table exists, insert data. Create if does not exist.

[jreback]

ok we can mark this disabled so it's not run by default (but u can still run locally)
maybe also add s comment about the timeout as you describe above

[parthea]

I'd like to keep this issue open because this exact issue is currently open on the BigQuery bug tracking site.
https://code.google.com/p/google-bigquery/issues/detail?id=771

[jreback]

closed #15381

let's migrate this (and others) to pandas-gbq