Paul Ockenfuß

I would also be very interested in this! There is another Python-Markdown extension, which does this: [Markdown_Captions](https://github.com/Evidlo/markdown_captions) So far, it seems to work in combination with pelican and elegant.

Thank you for the quick reply! The ideal solution for an end user like me would be if `np.inf` is supported, but if there are technical challenges, I would opt...

I just found another, quite weird, example where I get different values. Is it just on my machine? Can you reproduce this? ```python xr.set_options(use_numbagg=False) n1=9.9e+36 n2=7e+36 arr=[1,1,1,n1,1,n2,n1,1, 1, 1, 1,...

Ok, I see, makes sense regarding the continuously accumulating implementation in numbagg. Thanks for the clarification!

This behaviour is intended. In xarray, many operations broadcast the arrays first, i.e. xarray tries to bring them to the same shape before applying an operation. If one dimension is...

Two comments if the pandas arguments are introduced: * If the pandas arguments are introduced, one might raise the question, whether the `max_gap` argument of xarray is still necessary, since...

Do you have any thoughts on this? I think with the following signature, the function will be backward compatible: ``` interpolate_na(dim, method='linear', use_coordinate=True, limit=None, limit_direction='forward', limit_area=None, limit_use_coordinate=False, max_gap=None) ```

Thanks for the feedback @scottyhq! I used the time over christmas and tried to come up with an implementation, see #8577! As you suggested, I pointed out specific details to...

Some more results from the discussion: - This affects all filling methods: interpolate_na, fillna, ffill, bfill, ... - We want two functionalities: limit (or cutoff) to limit filling distance, max_gap...