MIGRATE: pandas-gbq

[jreback]

I created the pandas-gbq package to migrate the guts of pandas/io/gbq.py. Once this is release and on PyPi and conda-forge. Then we can switch the internal interface to use this code. I think this will allow this package to grow.

[jreback]

I have add all pandas to the repo as well as @parthea . If you would like to be added, please ping.

[max-sixty]

CC @clifton

[jorisvandenbossche]

Big +1 to have this as a separate package.

Then we can switch the internal interface to use this code

@jreback What is your idea on how to approach this?

Let read_gqb/to_gbq call the external package and pass everything on?

Roughly like

def read_gbq(*args, **kwargs):
    from pandas_gbq import read_gbq
    return read_gbq(*args, **kwargs)

class DataFrame():
    ...
    def to_gbq(self, *args, **kwargs):
        from pandas_gbq import to_gbq
        to_gbq(self, *args, **kwargs)

(with of course some checking of the import etc)

[jreback]

exactly the internal pandas implementation becomes trivial

[jreback]

ok moved all of the docs, so will change that as well: http://pandas-gbq.readthedocs.io/en/latest/index.html#