Hans Dembinski

Thank you for confirming, so this is not a bug, but this somewhat surprising behavior probably should be documented.

Thanks, please do!

Hi @dantrim I think what you have is a good solution, thank you for sharing. It is one of the possibilities that I would have suggested to you, but you...

Thank you for the offer, I will come back to you about that.

Some notes: 1) Due to the structure of Boost.Histogram, the storage is always equal to the product of axis sizes. The storage does not know anything about the axis structure...

> On more word on sparse storage: we should benchmark with this one https://github.com/greg7mdp/sparsepp. It is supposed to be a drop-in replacement for unordered_map, and so it should work out...

The author of sparsepp suggests this: https://github.com/greg7mdp/parallel-hashmap

Adding this in Boost.Histogram is not so easy, because it requires a large change to `reduce`. Currently, `reduce` can only select ranges. It is not possible to pick the 1,...

@paulgessinger I think I can easily add slicing of a category axis, you specify a range of indices and the resulting histogram has a category axis with that sub-range. This...

I added it to my issues. I am currently busy with some other stuff, but I will give this high priority.