feat: add dt.to_string to DaskExpr

[lucianosrp]

What type of PR is this? (check all applicable)

• 💾 Refactor
• ✨ Feature
• 🐛 Bug Fix
• 🔧 Optimization
• 📝 Documentation
• ✅ Test
• 🐳 Other

Related issues

• Related issue feat: add methods to Dask backend #637
• Closes #

Checklist

• Code follows style guide (ruff)
• Tests added
• Documented the changes

If you have comments or can explain your changes, please do so below.

Dask seems to be failing for %.f, any suggesstions on how to handle it ? Solved

[FBruzzesi]

That's definitly more effort than expected! Thanks for putting the work on this 🙌🏼
I left only a nitpick comment on how to test/compare series values.

Edit: Just noticed the test failing, maybe a deeper look is needed 🥲

[lucianosrp]

@FBruzzesi thanks as always, could you check now?
Edit: ah nevermind - still checking

[lucianosrp]

I am not able to replicate this failure in my local, any suggestions ?

[FBruzzesi]

I am not able to replicate this failure in my local, any suggestions ?

Sharing what I am discovering along the way: replacing with_columns with select does not raise an error. I am a bit lost as well 🙈

What I mean exactly is:

result = nw.from_native(df).select(
    nw.col("a").dt.to_string("%Y-%m-%d")
)

[lucianosrp]

replacing with_columns with select does not raise an error

Do you think there might be an issue with Narwhals' internals ? Or just Modin?
Shall we just change the test and monitor if we see this moving forward? 😓

[FBruzzesi]

I would not assume is something going wrong with narwhals since for pandas there are no issues and the logic should be the same. Apart from that, I never worked with modin directly 🥲

@MarcoGorelli would you mind weighting in for this whenever you have the time?

[MarcoGorelli]

thanks for the ping! will take a look at this and other open prs later today 😇

[MarcoGorelli]

looks like a bug in Modin - I've reported it to them here modin-project/modin#7371

shall we just xfail this test for modin for now?

[MarcoGorelli]

thanks @lucianosrp and @FBruzzesi for reviewing

I think you also need to xfail test_dt_to_string_iso_local_datetime_expr - then it should be good to go, feel free to merge once that's addressed (just check that all tests, apart from pyarrow-nightly, pass)

[lucianosrp]

Thanks @FBruzzesi and @MarcoGorelli as always 🫡