What happened?

Dataset object could not be fetched with Dataset[variables] when variables is a frozenlist. There is some source code that can replicate this behaviour.

What did you expect to happen?

Dataset[variables] returning a new fetched dataset. It returns a key error.

Minimal Complete Verifiable Example

```Python

%%

from dagster import Output, asset, Field, build_op_context import numpy as np import pandas as pd import xarray as xr

def retrieve_dataset(date, fct_step, lats, lons, runtime, source, variables):

dt = str(date)[0:10].replace("-", "")
year = dt[0:4]
prev_n = "{}".format(fct_step).zfill(3)

lat_toplot = np.arange(
    min(lats), max(lats), 0.25
)  # last number is exclusive
lon_toplot = (
    np.arange(min(lons), max(lons), 0.25) % 360
)  # last number is exclusive

gfs_model = "ds084.1"
URL = f"https://rda.ucar.edu/thredds/dodsC/files/g/{gfs_model}/{year}/{dt}/gfs.0p25.{dt}{runtime}.f{prev_n}.grib2"

dataset = xr.open_dataset(URL, cache=True, engine="netcdf4")[variables]

# Select required data via xarray
dataset = dataset.sel(lon=lon_toplot, lat=lat_toplot)[variables]

return dataset

@asset( description="GFS dataset fetched for NOAA Forecast data from NOAA NOMADS server.", config_schema=dict( variables=Field(list), ), output_required=True, ) def ncar_dataset(context):

date = "2022-02-01"
fct_step = 0
lats = [-4.0, -2.75]
lons = [-41.5, -40.25]
runtime = "00"
# variables = list(context.op_config["variables"])  <--- This way it runs
variables = context.op_config["variables"]
source = "model-gfs-003"

dataset = retrieve_dataset(
    date, fct_step, lats, lons, runtime, source, variables
)

return Output(dataset)

test_variable = "u-component_of_wind_height_above_ground"

with build_op_context( op_config={ "variables": [test_variable], } ) as context:

dataset_ast = ncar_dataset(context)

dataset = dataset_ast.value ```

MVCE confirmation

• [X] 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.
• [X] Verifiable example — the example copy & pastes into an IPython prompt or Binder notebook, returning the result.
• [X] New issue — a search of GitHub Issues suggests this is not a duplicate.

Relevant log output

Python I raises KeyError

Anything else we need to know?

No response

Environment