Thomas Bierhance

I would say the same way (AND?!) as it works with column names. Isn’t the type just another way of referring to columns instead of the name?

This is now consistently failing with column names as well and works when wrapping with `pl.all`: ```python select_dtypes_ok = df.select(pl.col(DTYPES_FLOAT)) #filter_colnames_fails = df.filter(pl.col(float_columns).is_not_nan()) filter_colnames_all_ok = df.filter(pl.all(pl.col(float_columns).is_not_nan())) #filter_dtypes_fails = df.filter(pl.col(DTYPES_FLOAT).is_not_nan()) filter_dtypes_all_ok...

I was not able to run your code and see what the problem is. I think you can get quiet close (see code below). Wouldn't this rather be a feature...

> @thomasbierhance I currently have the same code above in an application. However, doing it this way means that the legend shows the same value (the mean) for both "upr80"...

> > However, I fear that this can not be done using the current version of the R interface to dygraphs. > > Which is (of course) my use case!...

@amanda-hand Sure, I did a small [blog post for towardsdatascience](https://towardsdatascience.com/how-to-create-better-interactive-forecast-plots-using-r-and-dygraph-29bdd7146066) about it with step-by-step explanations. In case you hit against the paywall you can find the code without explanations in...

I don’t think that this can be done easily. The R wrapper does not seem to support the legendFormatter. If I’m not mistaken the dygraphs version that ships with the...

While `pl.List` gets first exploded and is reshaped according to the spec... ``` print(pl.DataFrame(pl.Series(name="a", values=[[1, 2, 3, 4]], dtype=pl.List(pl.Int64))).select(pl.col("a").reshape((2,2)))) print(pl.DataFrame(pl.Series(name="a", values=[[1], [2], [3], [4]], dtype=pl.List(pl.Int64))).select(pl.col("a").reshape((2,2)))) ┌───────────┐ │ a │ │...

Created a PR for this that hardens the implementation. I think `0` should only be available for empty series and only in the combination of `(0, 0)`, `(0, -1)`, `(-1,...

I somehow managed to close the preceding PR https://github.com/pola-rs/polars/pull/16223 where @stinodego suggested to create a utility function instead of putting this into the DataType classes.