[Enh]: Add support for DataFrame.rows

[Priyansh121096]

We would like to learn about your use case. For example, if this feature is needed to adopt Narwhals in an open source project, could you please enter the link to it below?

No response

Please describe the purpose of the new feature or describe the problem to solve.

I have a library which accepts a pandas dataframe and serializes it into a list of lists (df.values.tolist()) before processing it further. I would like to make the library work with polars dataframes as well and would like to use narwhals for the same.

Suggest a solution if possible.

For polars, nw.DataFrame.rows() would be the same as pl.DataFrame.rows().
For pandas, nw.DataFrame.rows() would be pd.DataFrame.values.tolist().

If you have tried alternatives, please describe them below.

No response

Would you want to open a pull request?

yes

Additional information that may help us understand your needs.

No response

[MarcoGorelli]

thanks @Priyansh121096 for the request! sure, open to this!

I think for pandas we might want to use https://pandas.pydata.org/docs/reference/api/pandas.DataFrame.itertuples.html#pandas.DataFrame.itertuples though

In [12]: pl.DataFrame({'a': [1,2], 'b': [3., 4.]}).rows()
Out[12]: [(1, 3.0), (2, 4.0)]

In [13]: list(pd.DataFrame({'a': [1,2], 'b': [3., 4.]}).itertuples(index=False, name=None))
Out[13]: [(1, 3.0), (2, 4.0)]

else .values would upcast

[Priyansh121096]

Thanks @MarcoGorelli . itertuples also seems to return a list of tuples which is consistent with the polars output as opposed to values.tolist() which returns a list of lists so I agree.

I'm interested in opening a PR for this. Let me know if that's fine.

[MarcoGorelli]

yes, of course, thanks! check contributing.md if you need help

[MarcoGorelli]

hey @Priyansh121096 - just wanted to ask how you were finding this, and if you appreciate any help

[Priyansh121096]

Hello @MarcoGorelli, thanks for reaching out. I have started the work on this today and should be raising a PR soon. I'll reach out if I have any questions.

[Priyansh121096]

@MarcoGorelli I've raised #317 for this

[MarcoGorelli]

I spoke to Polars devs about this today, and they said that if you need a list, it's actually more efficient to use .rows directly

So, sorry for my bad suggestion - I'll reopen this then, let's add .rows as well 💪

[Priyansh121096]

Sure @MarcoGorelli, I'll raise a PR for .rows then.

[Priyansh121096]

@MarcoGorelli raised #351 for .rows.