feature request: to_gbq with a bigframes.pandas.DataFrame should call .to_gbq() from BigQuery DataFrames not download and re-upload

[tswast]

Is your feature request related to a problem? Please describe.

Mixing pandas-gbq and BigFrames can be a recipe for confusion and increased costs. Sending a bigframes.pandas.DataFrame to pandas_gbq.to_gbq results in many queries as pandas-gbq tries to determine the appropriate schema to use for upload.

Describe the solution you'd like

Ideally, if bigframes is installed, pandas-gbq can detect if someone is trying to use a BigQuery DataFrames object with it and call the .to_gbq() method on the object, which would avoid unnecessary queries, a download, and a re-upload of the data.

Describe alternatives you've considered

• Make BigFrames.pandas.DataFrame less compatible with the pandas interface to avoid confusion by pandas-gbq.
• Still download and re-upload but skip the schema detection, which should be unnecessary with a BigFrames DataFrame.

Additional context

See internal issue 374235961

[tswast]

Since I'm trying to have a light day today due to OOO 1/2 day, I can pick this up, since I think I can get it done relatively quickly.