Use the jupyter_client API to assign SQL output to variables

[tmthyjames]

Currently I'm assigning variables using the __builtins__ module.

I think the better way to accomplish this is using the BlockingKernelClient class from the jupyter_client API.

Maybe the following:

from jupyter_client.blocking import BlockingKernelClient

kc = BlockingKernelClient()
kc.load_connection_info(<connect_info>)
kc.start_channels()

Then to create variables do kc.execute('foo = "bar"')

[tmthyjames]

To pass a dataframe we could do:

code = """
import pandas as pd
{0} = pd.DataFrame({1})
{0}.columns = {1}[0].keys()
""".format(var_name, data)

kc.execute(code)

Something like this should work.