Implement `Array.__iter__`

This came up recently in https://github.com/pandas-dev/pandas/issues/38645 and I took a greedy stab at implementing it in: https://github.com/dask/dask/pull/7888

I abandoned that PR because a bunch of tests failed and I feel like it’s at least a little surprising to have list(array) return a list of real values. It seems too easy to do by mistake.

For context, currently on main if you call list(array) you get an array of getitem calls with the same length as the 0th dimension of the array. It’s possible that is already the best solution. In which case we can probably close this.

[davidhassell]

Hello, I’ve been following this with interest, and also agree that the “array of getitem calls with the same length as the 0th” is what I would expect to occur (i.e. not returning computed numpy arrays).

Thanks, David

[jsignell]

On main you get

ValueError: Arrays chunk sizes are unknown: (nan,)

A possible solution: https://docs.dask.org/en/latest/array-chunks.html#unknown-chunks
Summary: to compute chunks sizes, use

   x.compute_chunk_sizes()  # for Dask Array `x`
   ddf.to_dask_array(lengths=True)  # for Dask DataFrame `ddf`