Open
Description
What is your issue?
I wanted to highlight a typing issue with xarray.core.combine.combine_by_coords where the first argument is data_objects: Iterable[Dataset | DataArray] = []. Specifically the type issue exists when a Generator which is an Iterable is passed as the data_objects argument. See the example below...
from typing import Iterable
import xarray as xr
def dataset_maker(key:str) -> xr.Dataset:
return xr.Dataset(
data_vars={
key: xr.DataArray(data=[1, 2, 3], dims=["x"])
},
coords=dict(
x=xr.DataArray(data=[1, 2, 3], dims=["x"]),
),
)
data_list = [dataset_maker('a'), dataset_maker('b')]
assert isinstance(data_list, Iterable)
print("... { data_list } ...")
print(xr.combine_by_coords(data_list))
data_tuple = dataset_maker('a'), dataset_maker('b')
assert isinstance(data_tuple, Iterable)
print("... { data_tuple } ...")
print(xr.combine_by_coords(data_tuple))
data_iterator = (dataset_maker(k) for k in ['a', 'b'])
assert isinstance(data_iterator, Iterable)
print("... { data_iterator } ...")
print(xr.combine_by_coords(data_iterator))The results.
... { data_list } ...
<xarray.Dataset> Size: 72B
Dimensions: (x: 3)
Coordinates:
* x (x) int64 24B 1 2 3
Data variables:
a (x) int64 24B 1 2 3
b (x) int64 24B 1 2 3
... { data_tuple } ...
<xarray.Dataset> Size: 72B
Dimensions: (x: 3)
Coordinates:
* x (x) int64 24B 1 2 3
Data variables:
a (x) int64 24B 1 2 3
b (x) int64 24B 1 2 3
... { data_iterator } ...
<xarray.Dataset> Size: 0B
Dimensions: ()
Data variables:
*empty*
The issue arises at the head of the function where the generator is exhausted.
objs_are_unnamed_dataarrays = [
isinstance(data_object, DataArray) and data_object.name is None
for data_object in data_objects
]A simple solution would be to change the type annotations to either that of a Sequence or memoize the data_objects variable as...
data_objects = list(data_objects)