If someone has a df: nw.DataFrame and they do

nw.dependencies.is_pandas_dataframe(df)

then they almost certainly meant to call

nw.dependencies.is_pandas_dataframe(df.to_native())

Source: I made this mistake 😳

To save others from making this mistake (as well as to save myself from re-making it in the future), we could:

• raise an error
• always raise a warning

Maybe just raising an error would be fine, I can't imagine a situation where you have an object which is either a narwhals dataframe or a native object and in both cases you pass it as-is to is_pandas_dataframe