DataTree attribute-like access bug

What happened?

I have a datatree with a dataset in a group. I tried to filter this dataset and store in the datatree. The dataset showed the filtered dataset and the unfiltered dataset, depending on the method to get the group. dt.variable1 or dt['variable1'].

What did you expect to happen?

I expect that both datasets are equal.

Minimal Complete Verifiable Example

Note the difference when `dt.variable1` and `dt['variable1']` are printed. When the datatree is exported the value of `dt['variable1']` is used. 


import numpy as np
import xarray as xr

ds1 = xr.Dataset(
    dict(
        variable1=xr.DataArray(
            np.random.randn(10),
            dims="x",
            coords=dict(x=np.linspace(1, 10, num=10)),
        )
    )
)
dt = xr.DataTree(ds1)
print("Datatree with 1 dataset (variable1)")
print(dt)

# filter on x
dt.variable1 = dt.variable1.sel(x=slice(1, 5))
print("\nPrint sliced dataset variable 1 using 'dt.variable1'")
print(dt.variable1)
print("\nPrint sliced dataset variable 1 using 'dt['variable1']'")
print(dt["variable1"])
print(
    "\nThe whole datatree how it is stored (see variable1, which is not sliced)"
)
print(dt)

Datatree with 1 dataset (variable1)
<xarray.DataTree>
Group: /
    Dimensions:    (x: 10)
    Coordinates:
      * x          (x) float64 80B 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
    Data variables:
        variable1  (x) float64 80B -0.03606 -0.5738 -1.331 ... -0.3654 1.669 0.9354

Print sliced dataset variable 1 using 'dt.variable1'
<xarray.DataArray 'variable1' (x: 5)> Size: 40B
array([-0.03605657, -0.57384656, -1.3312692 ,  1.0832746 , -0.07725834])
Coordinates:
  * x        (x) float64 40B 1.0 2.0 3.0 4.0 5.0

Print sliced dataset variable 1 using 'dt['variable1']'
<xarray.DataArray 'variable1' (x: 10)> Size: 80B
array([-0.03605657, -0.57384656, -1.3312692 ,  1.0832746 , -0.07725834,
       -0.45822326, -1.32865998, -0.36541352,  1.66870598,  0.93536491])
Coordinates:
  * x        (x) float64 80B 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0

The whole datatree how it is stored (see variable1, which is not sliced)
<xarray.DataTree>
Group: /
    Dimensions:    (x: 10)
    Coordinates:
      * x          (x) float64 80B 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
    Data variables:
        variable1  (x) float64 80B -0.03606 -0.5738 -1.331 ... -0.3654 1.669 0.9354


>>> import xarray
>>> xarray.__version__
'2025.1.0'

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

No response

Anything else we need to know?

No response

Environment

Start by running the minimal reproducible example from the issue and compare dt.variable1 with dt["variable1"] after assignment. Trace the DataTree attribute-access and item-access entry points, then add a regression test showing both accesses return the sliced dataset and confirm the displayed or exported tree uses that value.