id,node_id,number,title,user,state,locked,assignee,milestone,comments,created_at,updated_at,closed_at,author_association,active_lock_reason,draft,pull_request,body,reactions,performed_via_github_app,state_reason,repo,type
244016361,MDU6SXNzdWUyNDQwMTYzNjE=,1483,Loss of coordinate information from groupby.apply() on a stacked object,17701232,open,0,,,5,2017-07-19T11:59:48Z,2020-10-04T16:09:22Z,,NONE,,,,"I use this stack, groupby, unstack quite frequently.  e.g. [here](https://gist.github.com/rabernat/bc4c6990eb20942246ce967e6c9c3dbe)

An issue I have is that after  `groupby('allpoints').apply()`,  the coordinate names do not get carried through. i.e. the coordinate names are now: `allpoints_level_0` and `allpoints_level_1`. Then after `unstacking` I rename them back to lat/lon etc. Do you ever encounter this?

Is there a way to carry them through and is  this  an issue for others? 

```
import xarray as xr  
import numpy as np
ds = xr.DataArray(np.ndarray((180,360,2000)), coords={'lat':np.arange(90,-90,-1), 'lon':np.arange(-180,180), 'time':range(2000)})
ds

<xarray.DataArray (lat: 180, lon: 360, time: 2000)>
array([[[ 0.623891, -0.044304, ...,  1.015785,  0.009088],
        [-0.7375  ,  0.380369, ...,  0.788351, -0.69295 ],
        ..., 
        [ 0.171894,  0.517164, ..., -0.946908, -0.597802],
        [ 0.353743,  0.005539, ..., -1.436965, -0.190099]],
....
Coordinates:
  * lat      (lat) int32 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 ...
  * lon      (lon) int32 -180 -179 -178 -177 -176 -175 -174 -173 -172 -171 ...
  * time     (time) int32 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ..
```
Now we stack the data by allpoints. **Note that the info about original coordinates (lat / lon) is still there...**
`dst = ds.stack(allpoints=['lat','lon'])`
```
<xarray.DataArray (time: 2000, allpoints: 64800)>
array([[ 0.623891, -0.7375  ,  0.053525, ...,  0.379701,  0.130618,  0.11094 ],
       [-0.044304,  0.380369, -0.410632, ..., -0.739881,  0.203219, -0.506303],
       [-1.762024, -1.019424,  2.580218, ...,  1.491677,  1.189149, -0.072223],
       ..., 
       [-0.896298,  0.333163, -1.751641, ...,  1.90315 ,  2.642813, -0.913787],
       [ 1.015785,  0.788351,  0.379997, ...,  0.864934,  0.889001, -1.363458],
       [ 0.009088, -0.69295 , -1.276184, ...,  1.220656,  0.895599,  0.848757]])
Coordinates:
  * time       (time) int32 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 ...
  * allpoints  (allpoints) MultiIndex
  - lat        (allpoints) int64 90 90 90 90 90 90 90 90 90 90 90 90 90 90 ...
  - lon        (allpoints) int64 -180 -179 -178 -177 -176 -175 -174 -173 ...
```
Now apply `groupby().apply()`
`dsg=dst.groupby('allpoints').apply(my_custom_function)
`
```
<xarray.DataArray (allpoints: 64800)>
array([ 0.013697,  0.006272,  0.009744, ..., -0.016265, -0.002108, -0.014733])
Coordinates:
  * allpoints          (allpoints) MultiIndex
  - allpoints_level_0  (allpoints) int64 -89 -89 -89 -89 -89 -89 -89 -89 -89 ...
  - allpoints_level_1  (allpoints) int64 -180 -179 -178 -177 -176 -175 -174 ...

```
So now we have lost the `'lat','lon'`. However **if we skip the groupby part** and go straight to `unstack`, this would be carried through.
`dst.unstack('allpoints')`

```
<xarray.DataArray (time: 2000, lat: 180, lon: 360)>
array([[[ 0.623891, -0.7375  , ...,  0.171894,  0.353743],
        [ 1.780691, -0.747431, ...,  0.038754,  0.615228],
        ..., 
Coordinates:
  * time     (time) int32 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 ...
  * lat      (lat) int64 90 89 88 87 86 85 84 83 82 81 80 79 78 77 76 75 74 ...
  * lon      (lon) int64 -180 -179 -178 -177 -176 -175 -174 -173 -172 -171 ... 
```

","{""url"": ""https://api.github.com/repos/pydata/xarray/issues/1483/reactions"", ""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,,13221727,issue