Fix as_dataarray to apply coords parameter for DataArray input

[FBumann]

Fix as_dataarray to apply coords parameter for DataArray input

Summary

Fixes an API inconsistency in as_dataarray() where the coords parameter was silently ignored when the input was already a DataArray.

API Change

Before (broken)

da = xr.DataArray([20, 10], dims=['x'], coords={'x': ['b', 'a']})
result = as_dataarray(da, coords={'x': ['a', 'b']})
# result.coords['x'] = ['b', 'a']  # coords parameter IGNORED

After (fixed)

da = xr.DataArray([20, 10], dims=['x'], coords={'x': ['b', 'a']})
result = as_dataarray(da, coords={'x': ['a', 'b']})
# result.coords['x'] = ['a', 'b']  # coords parameter APPLIED
# result.values = [10, 20]         # values correctly reindexed

Behavior by input type

Input Type	coords param (before)	coords param (after)
np.ndarray	✅ Applied	✅ Applied
pd.Series	✅ Applied	✅ Applied
DataArray	❌ Ignored	✅ Applied

What the fix does for DataArray input

When coords is a dict:

1. Reindex existing dimensions to match the provided coordinate order
2. Expand to new dimensions from coords (via expand_dims)

Extra dimensions in the DataArray (not in coords) are preserved.

Breaking Change

This is technically a breaking change for code that:

• Passed a DataArray to as_dataarray() with a coords parameter
• Relied on the coords parameter being silently ignored

Migration: If you need the old behavior (pass-through), don't pass the coords parameter.

In practice, this is a bug fix - if you passed coords, you likely expected them to be applied.

Edge Cases

Scenario	Behavior
DataArray coords differ in order	Reindex to match coords
DataArray has fewer dims than coords	Expand to include new dims
DataArray has more dims than coords	Preserve extra dims
coords is a list (not dict)	No change (lists don't specify dim names)

Checklist

• Code changes are sufficiently documented
• Unit tests pass (1186 passed)
• A note for the release notes
• I consent to the release of this PR's code under the MIT license