I got the exact same error, also randomly working for one table and erroring on the other, but I've managed to resolve it by converting every field in my dataframe to string:

final_output_df.astype(str)

It's an unsatisfactory answer, but it seems to do the trick.