Finishing up this line of though: without the assumption that the relative order of dimensions is maintained across arrays in a set, this feature is impossible to implement as a neat function call. You would have to specify exactly how to expand each of the coordinates which can get pretty long.

I wrote some code, that I think should have worked if relative ordering was a valid assumption: Here it is for reference https://github.com/hmaarrfk/xarray/pull/1

To obtain the desired effect, you have to expand the dimensions of the coordinates individually: ```python import xarray as xr import numpy as np

Setup an array with coordinates

n = np.arange(1, 13).reshape(3, 2, 2) coords={'y': np.arange(1, 4), 'x': np.arange(1, 3), 'xi': np.arange(2)}

%%

z = xr.DataArray(n[..., 0]2, dims=['y', 'x']) a = xr.DataArray(n, dims=['y', 'x', 'xi'], coords={*coords, 'z': z})

sliced = a[0] print("The original xarray") print(a.z) print("The sliced xarray") print(sliced.z)

%%

expanded = sliced.expand_dims('y', 0) expanded['z'] = expanded.z.expand_dims('y', 0) print(expanded) ```

@shoyer Thank you for your first detailed response. It definitely took me a while to understand all the intricacies of your words.

In fact, if you do use a Dataset, as opposed to a DataArray, you are able to expand dims for multiple simultaneous arrays together.

For my particular case, a Dataset is probably more appropriate, but I liked the fact that the DataArray had one of the sets as the main set and that I could do math on that one.

Here is an example showing how if you use a Dataset, the behaviour I expect is obtained:

```python

%%

import xarray as xa import numpy as np

n = np.zeros((3, 2))

data = xa.DataArray(n, dims=['y', 'x'], coords={'y':range(3), 'x':range(2)})

z=xa.DataArray(np.arange(6).reshape((3, 2)), dims=['y', 'x'])

my_set = xa.Dataset({'mydata':data, 'z':z})

print('Original Data') print('=============') print(my_set)

%%

my_slice = my_set[{'x': 0, 'y':1}] print("Sliced data") print("===========") print("z coordinate remembers it's own x value") print(f'x = {my_slice.z.x}')

%%

expanded_slice = my_slice.expand_dims('x') print("expanded slice") print("==============") print("forgot that 'z' had 'x' coordinates") print("but remembered it had a 'y' coordinate") print(f"z = {expanded_slice.z}") print(expanded_slice.z.x) ``` Output:

python Original Data ============= <xarray.Dataset> Dimensions: (x: 2, y: 3, z2: 0) Coordinates: * y (y) int64 0 1 2 * x (x) int64 0 1 * z2 (z2) float64 Data variables: mydata (y, x) float64 0.0 0.0 0.0 0.0 0.0 0.0 z (y, x) int32 0 1 2 3 4 5 Sliced data =========== z coordinate remembers it's own x value x = <xarray.DataArray 'x' ()> array(0, dtype=int64) Coordinates: y int64 1 x int64 0 expanded slice ============== z = <xarray.DataArray 'z' (x: 1)> array([2]) Coordinates: y int64 1 * x (x) int64 0 <xarray.DataArray 'x' (x: 1)> array([0], dtype=int64) Coordinates: y int64 1 * x (x) int64 0

I guess I probably erroneously assumed a "scalar" was a 1x1x1x1....x1 array. It really isn't. It is just a scalar. When the number of dimensions hits 0, an IndexVariable, it automatically gets demoted to a Variable somewhere here https://github.com/pydata/xarray/blob/56381ef444c5e699443e8b4e08611060ad5c9507/xarray/core/variable.py#L465

I guess maybe: 1. Maybe what I'm looking for is a Dataset 2. Maybe sticky dimensions would be nice to have? We could maybe flag those dimensions as being sticky, and even if you select a single element, it would be maintained as a dimension. Is this already possible?

@shoyer thank you for the quick response.