Support for query parameterization

[dylanscott]

I was wondering if there were plans to expose the ability to pass bound parameters (either named or qmark style) when constructing constructing a bigframes DataFrame e.g. via Session.read_gbq_query? My understanding is they would need to be thread through to the queryParameters field on the QueryJobConfig that gets constructed.

I realize it's probably a bit of a niche use-case since this is a code-based API but for context we were looking to integrate support for bigframes in a product that allows users to build Python notebook-based interactive apps which may be used by external/untrusted users (not the ones writing the app/code) and so our SQL functionality relies on using parameterization whenever these kinds of inputs are involved.

[tswast]

In pandas-gbq this is exposed via a generic "configuration" parameter, where the JobConfig REST object can be passed directly as a dictionary.

I suggest that read_gbq and read_gbq_query add the same configuration argument in BigQuery DataFrames.

[tswast]

Draft PR: #401

[tswast]

Expected usage will be the same as pandas-gbq. See: https://stackoverflow.com/a/62171130/101923

sql = """
    SELECT name
    FROM `bigquery-public-data.usa_names.usa_1910_current`
    WHERE state = @state
"""
query_config = {
    'query': {
        'parameterMode': 'NAMED',
        'queryParameters': [
            {
                'name': 'state',
                'parameterType': {'type': 'STRING'},
                'parameterValue': {'value': 'TX'}
            },
        ]
    }
}
df = pandas_gbq.read_gbq(sql, configuration=query_config)

[tswast]

This has been released in version 1.0.0