Code Sample, a copy-pastable example if possible

```python import xarray as xr import dask.array as da from dask.diagnostics import ProgressBar a = xr.DataArray(da.zeros((3, 5, 5), chunks=(1, 2, 2)), dims=('bands', 'y', 'x'))

with ProgressBar(): b = a.transpose('y', 'x', 'bands')

dask does not show a progress bar due to no computation

with ProgressBar(): b = a.transpose('y', 'x', 'bands') b.compute()

dask computes the array (since we told it to) and we see a progress bar

```

Question

The documentation for transpose says that it is not lazy. Is this only in certain situations? By not lazy does it mean that when the data is computed that the transpose task will require all data to be loaded at once (one large chunk) or does it mean that the transpose operation will immediately compute the transposed array? My test above does not seem to compute the data when transpose is called.

Or is the documentation just outdated?