html_url,issue_url,id,node_id,user,created_at,updated_at,author_association,body,reactions,performed_via_github_app,issue
https://github.com/pydata/xarray/pull/1496#issuecomment-319471941,https://api.github.com/repos/pydata/xarray/issues/1496,319471941,MDEyOklzc3VlQ29tbWVudDMxOTQ3MTk0MQ==,5635139,2017-08-01T19:28:15Z,2017-08-01T19:28:15Z,MEMBER,"Not a clear comment. I had meant for this to work, which it does:

```python

In [1]: import xarray as xr

In [2]: import numpy as np

In [3]: a = xr.DataArray(np.random.rand(3,4), dims=['x','y'])

In [4]: b = xr.DataArray(np.random.rand(3,1), dims=['x','y'])

In [5]: a+b.squeeze()
Out[5]:
<xarray.DataArray (x: 3, y: 4)>
array([[ 1.394345,  0.916842,  0.806284,  1.604707],
       [ 1.011313,  0.736347,  0.677679,  0.970856],
       [ 0.477433,  0.825672,  1.014959,  0.495829]])
Dimensions without coordinates: x, y

In [6]: a.where(a>0.5, b.squeeze())
Out[6]:
<xarray.DataArray (x: 3, y: 4)>
array([[ 0.676462,  0.717883,  0.717883,  0.886823],
       [ 0.568829,  0.442484,  0.442484,  0.528372],
       [ 0.100403,  0.725269,  0.914556,  0.100403]])
Dimensions without coordinates: x, y
```

I had meant: `'exact'` includes `other` having a _subset_ of dimensions, where the missing dimensions are easy to broadcast out","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,246502828
https://github.com/pydata/xarray/pull/1496#issuecomment-319423470,https://api.github.com/repos/pydata/xarray/issues/1496,319423470,MDEyOklzc3VlQ29tbWVudDMxOTQyMzQ3MA==,5635139,2017-08-01T16:27:41Z,2017-08-01T16:27:41Z,MEMBER,"This is excellent! Thank you @shoyer . 

I agree with forcing exact alignment for the 3 arg version. Does `exact` allow a scalar on a dimension? That's then simple and unambiguous to broadcast.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,246502828