CODEWITHSUNDEEP
numpypython-xarray
Canol Gökel
Canol Gökel

Reputation: 1275

How to check if two xarray objects have the same dimensions and coordinates (but not necessarily values)

I have 2 xarray objects. I would like to check if these 2 arrays have the same dimensions and coordinates. But the two objects don't need to have the same values or meta data.

For example:

import xarray as xr

arr1_data = [[1, 2], [3, 4]]
arr2_data = [[5, 6], [7, 8]]

arr1 = xr.DataArray(arr1_data, dims=["y", "x"], coords={"y": ["coord1", "coord2"]})
arr2 = xr.DataArray(arr2_data, dims=["y", "x"], coords={"y": ["coord1", "coord2"]})

compare_xarray_objects(arr1, arr2)  # Should return True

The .dims attribute returns a tuple, so it is easy to tell if dimensions are the same. But .coords attribute returns a DataArrayCoordinates object and I don't know how to compare a DataArrayCoordinates object with another one.

Upvotes: 4

Views: 2091

Answers (1)

Michael Delgado
Michael Delgado

Reputation: 15442

For cases like the one in your example where you're trying to compare indexing coordinates, simply use xr.align with the join='exact' argument.

xr.align(arr1, arr2, join='exact')  # will raise a ValueError if not aligned

If you're trying to also compare non-indexing coordinates, you could also use xr.testing.assert_equal. This will check all coordinates, including those not in the dimension of the array (e.g. if you had another coordinate z indexed by x and y). However, it's a bit pickier than xr.align, as it requires the dimension ordering to be the same. If this were an issue, you could loop through the non-indexing coordinates and use this testing function to check equality for each coordinate.

Upvotes: 3

Related Questions