Jérôme Dockès
Jérôme Dockès
I agree with you the `raw` probably is worse than `skb`, especially since it would not be the actual value but still a lazy object, so it would be super...
@Vincent-Maladiere you are right of course, and sorry for pushing back too quickly and too hard on the `skrub_` solution @MarieSacksick . To me the deeper issue is that we...
Given the above, one thing to consider when choosing between `skb_` and `skb.` is which one more clearly sets apart dataframe functions from skrub functions, I think a case could...
regarding user feedback indeed it is important to gather it but note it will remain very scarce as long as it is only possible to try the feature by installing...
why close as not planned? we should at least have a better error message if we don't want to support return_indices
ah ok, thanks! gh interface can be a bit confusing 😅
Maybe allowing to specify the thousands separator & the decimal separator would be a good starting point indeed. It can be handled efficiently because the thousands separator can just be...
awesome @gabrielapgomezji ! I think a first step could be looking for example datasets with such columns. I suspect in many cases it won't be quite as clean as "40...
> But do we want to address these cases for now? I don't think we ever want to address these cases (they are too varied to come up with a...
Also if the simple case is very easily handled with pandas or polars directly, for example with something like ``` >>> df weight 0 4 kg 1 2.3 kg >>>...