Support for xarray-datatrees?

[ahuang11]

Hi! Amazing documentation and really excited for this. Was wondering if it'd be valuable to have a from_datatree method:

import numpy as np
import xarray as xr
from xarray_sql import XarrayContext

data = xr.DataArray(np.random.randn(2, 3), dims=("x", "y"), coords={"x": [10, 20]})
ds = xr.Dataset({"foo": data, "bar": ("x", [1, 2]), "baz": np.pi})
ds

ds2 = ds.interp(coords={"x": [10, 12, 14, 16, 18, 20]})
ds2

ds3 = xr.Dataset(
    {"people": ["alice", "bob"], "heights": ("people", [1.57, 1.82])},
    coords={"species": "human"},
)
ds3

dt = xr.DataTree.from_dict({"simulation/coarse": ds, "simulation/fine": ds2, "/": ds3})
dt

XarrayContext.from_datatree(dt)

Which would simply loop through each dataset and register them?

[alxmrs]

I really like this idea, and this this is generally good approach for registering DataTrees to be read as tables. However, I think making a CatalogProvider for the entire tree/ds, then allowing users to open specific tables, may be a better affordance for this looping logic. What do you think?

https://datafusion.apache.org/library-user-guide/catalogs.html

[ahuang11]

Yes that sounds good to me. I was not aware of datafusion catalogs.