Francesco Bruzzesi

Hellooo 👋🏼 Apologies for the late feedback on this - My days at work are a bit hectic. I have been thinking a bit on how to extend `Implementation` but...

Thanks @dangotbanned (https://github.com/narwhals-dev/narwhals/pull/3335#issuecomment-3591831415) - I certainly missed that. I assumed that if you had a solution already, this would have been pushed to main rather than in a branch 😂...

Thanks @MarcoGorelli > * I think the function just needs to not make any guarantees about order, I don't think we need to intentionally add randomness, just taking `first` for...

@MarcoGorelli just to make sure we align before going on a tangent. We briefly touched the possibility of _explicitly_ marking this feature as unstable as it's outside of polars API....

Hey @dangotbanned 👋🏼 I am missing some context for the https://github.com/narwhals-dev/narwhals/pull/3220#issuecomment-3570084015 What would you like to see achieved? If it's dataframe with one or more datetime with timezones colujmns, then...

> [!WARNING] > This proposal might end up be a bit overwhelming in one go The best way to test this implementation would be to start rolling it out as...

This is not supported in polars either 🤔 ```py import polars as pl data = {"a": [5, 4, 3, 2, 1]} pl.DataFrame(data).with_columns(a_max=pl.col("a").max().over()) ``` > TypeError: Expr.over() missing 1 required positional...

@MarcoGorelli this is currently possible with order dependent ops, but not with general aggregation. I am not sure which is the sweet spot you want to reach 👀

Regarding CI failures: the only one related to this PR seem to be tubular 👀

Wouldn't `Expr.sample(n=1)` be what we are after without having to diverge from the polars API?