Ritchie Vink
Ritchie Vink
I was, but I bounced because it is a bit more complicated than I had anticipated. For instance `pl.count()` does not have a column origin, so it does not know...
> Isn't `pl.count("col_a")` just a shorthening of `pl.col("col_a").count()` so the `include_nulls` would work here too? No, `pl.count()` does not refer to any column. > Do you mean that only for...
Yes, agree that we should have a `glob` flag. Also interesting recursion going on in the error message? :thinking:
Not yet, we are working on the architecture to do more of the expression API streaming. But this requires some patience. There is a lot of pre-work to be done.
@stinodego was their something that should still be done on OPs side? I think we still had to get to this one. Evwn though we cannot get to it soon...
Ok, then I understand. :)
Whatever the conclusions, this peeks my interest. ^^ Let's make clear what we can do on the Polars side. - Zero copy from a 2D numpy matrix to a Polars...
> in between sklearn.Pipeline steps pandas might be faster, due to lack of memory copies. Note that lack of memory copies isn't entirely true. If users enable pandas COW (which...
> Currently, we use copy=False when wrapping the ndarray to avoid the copy: Right. That does not copy memory. And I think it is ok internally for scikit-learn IFF you...