I am using pandas with BigQuery tables. Because pandas doesn't support boolean missing values, BigQuery BOOLEAN columns become dataframe object columns. I have to use to_gbq() table_schema to convert force those columns back to BOOLEAN.

My dataframe has 62 columns. table_schema has to include every column. It would be nice if I could only pass in the boolean columns in table_schema; for the other columns, to_gbq should default to dataframe column type.