`to_pandas` function - update docs

[bwalchen]

Good afternoon,
First, thank you for all the work you've put into creating this Python package.

Problem Statement
to_pandas() call is extremely slow if the returned data has a lot of columns.

Example: a simple query to the user table will return 7 rows of data, but will also return 100-200 columns. The to_pandas() call in this case will take several minutes to complete even though there are only 7 rows.

Request

1. Print a warning when running that function if number of columns is larger than 20.
2. Update the function docstring and recommend specifying the desired fields with gr.fields