issue_comments

Finished first complete part, ready for merging: https://github.com/pydata/xarray/pull/963

Note: there is a failed unit test, TestDataset.test_broadcast_nocopy, which shows broadcast on dataset doing a data copy whereas it shouldn't. Could you look into it?

I'm now moving to auto-calling broadcast inside concat()...

Started working on it. Still a lot of cleanup to be done. https://github.com/crusaderky/xarray/commit/8830437865c43e472101ca91a12f714ee43546cc

Observations: - I can't put any sense in the skip_single_target hack or in the whole special treatment inside align() for when there is only one arg. What's the benefit of the whole thing? What would happen if we simply removed the special logic? - DataArray.reindex(copy=False) still performs a copy, even if there's nothing to do. I'm a bit afraid to go and fix it right now as I don't want to trigger domino effects - I'm experiencing a lot of grief because assertDatasetIdentical expects both the coords and the data vars to have the same order, which in some situations it's simply impossible to control without touching vast areas of your code. As a more general and fundamental point, I cannot understand what's the benefit of using OrderedDict instead of a plain dict for coords, attrs, and Dataset.data_vars?

I can work on it. Should we go for a default outer join + broadcast inside concat? If the input already arrive aligned, or if the user wants a different type of join, this will slow things down with useless code though. What's your policy on this?

I'm now facing the equivalent problem with broadcast(). I need to concat 3 arrays:

python a = xarray.DataArray([[1,2],[3,4]], dims=['time', 'scenario'], coords={'time': ['t1', 't2']}) b = xarray.DataArray([5,6], dims=['time'], coords={'time': ['t3', 't4']}) c = xarray.DataArray(7, coords={'time': 't5'})

I want to broadcast dimension 'scenario' and concatenate on dimension 'time'. However, I can't invoke broadcast() because 1. a and b have misaligned time 2. c has a time coord, but not a time dim - if you broadcast a and c, you will get incorrect results

Again the solution would be to have an optional parameter exclude=['time']

Just found xarray.core.alignment.partial_align(), which does exactly that. Not sure why the functionality is not part of the public API? Looks like the solution is to simply merge partial_align() into align()?