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/issues/1482#issuecomment-316328685,https://api.github.com/repos/pydata/xarray/issues/1482,316328685,MDEyOklzc3VlQ29tbWVudDMxNjMyODY4NQ==,6815844,2017-07-19T09:33:41Z,2017-07-19T09:42:23Z,MEMBER,"I have a similar use case and I often use MultiIndex,
which (partly) enables to handle hierarchical data structure.

For example, 
```python
In [1]: import xarray as xr
   ...: import numpy as np
   ...:
   ...: # image 0, size [3, 4]
   ...: data0 = xr.DataArray(np.arange(12).reshape(3, 4), dims=['x', 'y'],
   ...:                      coords={'x': np.linspace(0, 1, 3), 
   ...:                              'y': np.linspace(0, 1, 4),
   ...:                              'image_index': 0})
   ...: # image 1, size [4, 5]
   ...: data1 = xr.DataArray(np.arange(20).reshape(4, 5), dims=['x', 'y'],
   ...:                      coords={'x': np.linspace(0, 1, 4), 
   ...:                              'y': np.linspace(0, 1, 5),
   ...:                              'image_index': 1})
   ...: 
   ...: data = xr.concat([data0.expand_dims('image_index').stack(xy=['x', 'y', 'image_index']),
   ...:                   data1.expand_dims('image_index').stack(xy=['x', 'y', 'image_index'])],
   ...:                  dim='xy')

In [2]: data
Out[2]: 
<xarray.DataArray (xy: 32)>
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11,  0,  1,  2,  3,  4,  5,
        6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19])
Coordinates:
  * xy           (xy) MultiIndex
  - x            (xy) float64 0.0 0.0 0.0 0.0 0.5 0.5 0.5 0.5 1.0 1.0 1.0 ...
  - y            (xy) float64 0.0 0.3333 0.6667 1.0 0.0 0.3333 0.6667 1.0 ...
  - image_index  (xy) int64 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 ...

In [3]:  data.sel(image_index=0)  # gives data0
Out[3]: 
<xarray.DataArray (xy: 12)>
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11])
Coordinates:
  * xy       (xy) MultiIndex
  - x        (xy) float64 0.0 0.0 0.0 0.0 0.5 0.5 0.5 0.5 1.0 1.0 1.0 1.0
  - y        (xy) float64 0.0 0.3333 0.6667 1.0 0.0 0.3333 0.6667 1.0 0.0 ...

In [4]: data.sel(x=0.0)  # x==0.0 for both images
Out[4]: 
<xarray.DataArray (xy: 9)>
array([0, 1, 2, 3, 0, 1, 2, 3, 4])
Coordinates:
  * xy           (xy) MultiIndex
  - y            (xy) float64 0.0 0.3333 0.6667 1.0 0.0 0.25 0.5 0.75 1.0
  - image_index  (xy) int64 0 0 0 0 1 1 1 1 1
```

<s>I think the above solution is essentially equivalent with
> all images of the same size into its own dimension</s>

EDIT:
I didn't understand the comment correctly.
The above corresponds to that all the images are flattened out and combined along one large dimension.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",,243964948