PygQuery

🐷 PygQuery

Multi-treaded wrapper to read and write Pandas dataframes with Google BigQuery without the hassle of the heavy BigQuery API.

PygQuery is multi-treaded by design, meaning that any SQL request is a thread of its own. The advantage of it is that you can send multiple requests in parallel, and have awaiting inbound data ready for later.

Install

On CLI, just type:

pip install pygquery

Read Data

Let's import the module first

from pygquery.bigquery import BigQueryReader

The module takes 3 arguments as an input:

1. request : A string of your query. E.g. """SELECT * FROM myproject.dataset.table"""
2. project : The string of the project you are currently gathering data from
3. api_key_path : a path of the G Service Account key, you can create one in the IAM tab of your GCP interface

Let's instantiate our data reader:

reader_dict = {
  'request' : """SELECT * FROM myproject.dataset.table""",
  'project' : 'myproject',
  'api_key_path' : 'folder/key.json'
}

# If there any error in your query at the instantiation stage, BigQuery will let you know
my_request = BigQueryReader(**reader_dict) 

Now you have an object ready to be launched. If the line of code above executes, you know that:

1. There is no error in the SQL
2. There is no credentials failure

Let's fire up this object:

my_request.start() # Launch the thread for downloading data

"# ... Do other things while data is downloading, like launching another request ... #"

my_request.join() # Tell to Python to wait for your download to complete

my_data = myRequest.data # Get your data

Et voilà! You have your data in Pandas DataFrame format ready to be crunched.

my_data.info()
my_data.head()