Rafael Schouten

Maybe we can achieve this in a more general way by ~using `Iterators.Stateful` to iterate over the chunks,~ caching the chunks along the column as we load them and swapping...

@meggart it seems we need to add that `DOCUMENTER_KEY` There is also another failure there I don't totally understand https://github.com/meggart/DiskArrays.jl/runs/8109369706?check_suite_focus=true#step:5:18

Perfect. We will also need to handle the non-lazy case of concatenating a disk array with other kinds of `AbstractArray`. Maybe by broadcasting each array to views of a preallocated...

Thanks for starting this. I haven't done anything yet, I have a lot of packages mid-way to release so my capacity is pretty low for a while, but I would...

I agree broadcast will be the hardest part, that's a good idea to leave it until last. I was imagining chunking would be integral to a lot of this too,...

This looks good. Seems pretty comprehensive for indexing it will be great to have everything standardised. My vote is also to implement it in ArchGDAL, then GeoData.jl can use it...

I was also wondering if it should be in JuliaArrays... HDF5 is a lot more general than geospatial data, and there are probably other non geo applications.

Yeah something like that could work. And we just use the largest possible chunks of the second? So some duplication of chunk loads but maybe not terrible

Changing the name now means changing the dependencies of quite a few packages. @xiaodaigh you would want to commit to making all the PRs, or this is a headache for...

In Julia generally only vectors can grow. Multidimensional arrays can't. And with a vector, `setindex!` cant do this either. I would argue that `setindex!` shouldn't change the array size like...