How to install gcp in Python?

Question

Lots of the BigQuery examples begin with:

import gcp.bigquery as bq

But I get ImportError: No module named gcp.bigquery whenever I try to run this.

How do I install this library?

I'm working in a virtualenv with python 2.7. I've tried pip install gcp, pip install gcloud, and pip install google-api-python-client.

None of them help and I can't find any documentation. Help!

UPDATE: the reason I want to use gcp is that I want to get data from BigQuery, preferably in CSV form, from within a Python script. If there's a better way to do this, I'm all ears...

Answer 1

You can build the library from datalab teams' content on github:

1. Clone the repository https://github.com/GoogleCloudPlatform/datalab/
2. Run tools/initenv.sh
3. Run sources/build.sh
4. There you have it under sources/build/

Hope this helps. Executing the docker image locally does not work for me at least.

Answer 2

For anyone with this problem, it looks like the datalabs library was updated and now you should import things differently.

import datalab.bigquery as bq

Answer 3

Use pandas and google-api-python-client. The function you are looking for is pd.read_gbq http://pandas.pydata.org/pandas-docs/stable/generated/pandas.io.gbq.read_gbq.html

Answer 4

If you're accessing BigQuery in python, you can do that using the gcloud library.

First, install the gcloud library:

$ pip install --upgrade gcloud

Then, after setting up your auth and project info, you can make api calls in python like this (adapted from the gcloud-python docs):

from gcloud import bigquery

client = bigquery.Client()
datasets, next_page_token = client.list_datasets()
print([dataset.name for dataset in datasets])

(As someone mentioned previously, you can also do it using the google-api-python-client.)

License: Apache 2

Answer 5

You should try a simple:

$ pip install --upgrade google-api-python-client

as discussed in the documentation.

Furthermore, gcp.bigquery is part of Google Cloud DataLab, so you should try from that angle if you are still interested.

Answer 6

gcp.bigquery is a library specific to Cloud Datalab (as would be any samples you saw such an import in).