check_dim_order=False not working well when passing xr.Dataset to xarray.testing.assert_equal #10704
Description
What happened?
xarray.testing.assert_equal with check_dim_order=False
fails with two datasets that only differ in the dim_order of individual data vars or coords when the relative order of the dims is not consistent across all variables in the dataset.
I believe this is because this line
b.transpose(*a.dims)
is getting applied when a is a dataset, but a.dims order is not meaningful when a is a dataset.
I believe potential solutions are:
- transpose the dims on a var by var basis.
- Do something like
dim_order = list(a.dims)
b = b.transpose(*dim_order)
a = a.transpose(*dim_order) # Note this is necesary as not all vars in `a` may actually follow a.dims order.
What did you expect to happen?
xarray.testing.assert_equal with check_dim_order=False
should not fail with two datasets only differ in the dim_order of individual data vars or coords.
Minimal Complete Verifiable Example
import xarray
import numpy as np
dataset_1 = xarray.Dataset(
{"foo": xarray.DataArray(np.zeros([4, 5]), dims=("a", "b")),
"bar": xarray.DataArray(np.zeros([5, 4]), dims=("b", "a"))})
dataset_2 = xarray.Dataset(
{"foo": xarray.DataArray(np.zeros([5, 4]), dims=("b", "a")),
"bar": xarray.DataArray(np.zeros([4, 5]), dims=("a", "b"))})
xarray.testing.assert_equal(dataset_1["foo"], dataset_2["foo"], check_dim_order=False) # passes
xarray.testing.assert_equal(dataset_1["bar"], dataset_2["bar"], check_dim_order=False) # passes
xarray.testing.assert_equal(dataset_1, dataset_2, check_dim_order=False) # Fails with log below.MVCE confirmation
- Minimal example — the example is as focused as reasonably possible to demonstrate the underlying issue in xarray.
- Complete example — the example is self-contained, including all data and the text of any traceback.
- Verifiable example — the example copy & pastes into an IPython prompt or Binder notebook, returning the result.
- New issue — a search of GitHub Issues suggests this is not a duplicate.
- Recent environment — the issue occurs with the latest version of xarray and its dependencies.
Relevant log output
> AssertionError Traceback (most recent call last)
/tmp/ipython-input-1957692165.py in <cell line: 0>()
2 xarray.testing.assert_equal(dataset_1["bar"], dataset_2["bar"], check_dim_order=False)
3
----> 4 xarray.testing.assert_equal(dataset_1, dataset_2, check_dim_order=False)
[... skipping hidden 1 frame]
/usr/local/lib/python3.12/dist-packages/xarray/testing/assertions.py in assert_equal(a, b, check_dim_order)
144 assert a.equals(b), formatting.diff_array_repr(a, b, "equals")
145 elif isinstance(a, Dataset):
--> 146 assert a.equals(b), formatting.diff_dataset_repr(a, b, "equals")
147 elif isinstance(a, Coordinates):
148 assert a.equals(b), formatting.diff_coords_repr(a, b, "equals")
AssertionError: Left and right Dataset objects are not equal
Differing data variables:
L bar (b, a) float64 160B 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0
R bar (a, b) float64 160B 0.0 0.0 0.0 0.0 0.0 0.0 ... 0.0 0.0 0.0 0.0 0.0Anything else we need to know?
No response