This seems to be a strictly better version of the dataframe equality check than this: automatic conversion to Pandas, prints the entire DataFrames if it fails.

We should just use it everywhere?