Francesco Bruzzesi

@MarcoGorelli I have a couple of questions: - Could you expand a little bit more, here or somewhere else on the following: > we should not add anything which modifies...

(Asking for a friend 👀) how much cheating are we allowed to? Specifically, pandas-like `translate_dtype` should apply one-to-one for dask.

Just to give a sense of how far we have gone and what's still missing. `DaskExpr` methods (those marked with * at the end means that modify the index and...

As there is not much left: for anyone interested, we could use the DaskNamespace `concat` implementation 😇

Closing this for now as the only methods which are left modify the index and #743 has just been closed

> probably fine to ignore them for now? or even, ignore them completely. we could be a bit stricter than Polars here, and guide people towards https://docs.pola.rs/api/python/stable/reference/expressions/api/polars.concat_str.html to concatenate strings...

> then maybe we want to support `concat_str` first Well it turns out that I am facing one of the very painful pandas issue: ```py import pandas as pd df...

This would official be a regression. I [searched on github](https://github.com/search?q=%22nw.sum_horizontal%22&type=code&p=1) if anyone is using the feature on a public repo and ended up empty handed, which is a good sign....

Hey sorry I am from mobile, so this could be a sloppy comment. I will try to do my best. > Think you're right around the behaviour of "outer" by...

> * keyword argument of "inner" now passed into `dd.concat` - I think this is the desired behaviour? Thanks! I am still a bit unsure about the difference. I can...