`flatten` as convinience method for stack all?

[mathause]

There is a bunch of operations that need to be conducted on a 1D array. It could be handy to have a convinience method that mimicks numpys flatten, i.e. works along the lines:

def flatten(dataarray, name='stacked'):
    return dataarray.stack(**{name: dataarray.dims})

What do you think?

[shoyer]

Yes, seems like a great idea to me!

[feldman4]

This is a good idea, any reason not to add linear indexing with array.flat[:] as well?

[stale]

In order to maintain a list of currently relevant issues, we mark issues as stale after a period of inactivity
If this issue remains relevant, please comment here; otherwise it will be marked as closed automatically

[shoyer]

We might call this stack_all() instead -- that's closer to .stack() with alphabetical sorting.

[dcherian]

Solved by .stack=[...]?

[mathause]

Yes indeed, good catch.

Closed by #3826