test_backends.py

from __future__ import annotations

import contextlib
import gzip
import itertools
import math
import os.path
import pickle
import platform
import re
import shutil
import sys
import tempfile
import uuid
import warnings
from collections.abc import Generator, Iterator, Mapping
from contextlib import ExitStack
from io import BytesIO
from os import listdir
from pathlib import Path
from typing import TYPE_CHECKING, Any, Final, Literal, cast
from unittest.mock import patch

import numpy as np
import pandas as pd
import pytest
from packaging.version import Version
from pandas.errors import OutOfBoundsDatetime

import xarray as xr
from xarray import (
    DataArray,
    Dataset,
    backends,
    load_dataarray,
    load_dataset,
    open_dataarray,
    open_dataset,
    open_mfdataset,
    save_mfdataset,
)
from xarray.backends.common import robust_getitem
from xarray.backends.h5netcdf_ import H5netcdfBackendEntrypoint
from xarray.backends.netcdf3 import _nc3_dtype_coercions
from xarray.backends.netCDF4_ import (
    NetCDF4BackendEntrypoint,
    _extract_nc4_variable_encoding,
)
from xarray.backends.pydap_ import PydapDataStore
from xarray.backends.scipy_ import ScipyBackendEntrypoint
from xarray.coding.cftime_offsets import cftime_range
from xarray.coding.strings import check_vlen_dtype, create_vlen_dtype
from xarray.coding.variables import SerializationWarning
from xarray.conventions import encode_dataset_coordinates
from xarray.core import indexing
from xarray.core.options import set_options
from xarray.namedarray.pycompat import array_type
from xarray.tests import (
    assert_allclose,
    assert_array_equal,
    assert_equal,
    assert_identical,
    assert_no_warnings,
    has_dask,
    has_netCDF4,
    has_scipy,
    mock,
    network,
    requires_cftime,
    requires_dask,
    requires_fsspec,
    requires_h5netcdf,
    requires_h5netcdf_ros3,
    requires_iris,
    requires_netCDF4,
    requires_netCDF4_1_6_2_or_above,
    requires_pydap,
    requires_pynio,
    requires_scipy,
    requires_scipy_or_netCDF4,
    requires_zarr,
)
from xarray.tests.test_coding_times import (
    _ALL_CALENDARS,
    _NON_STANDARD_CALENDARS,
    _STANDARD_CALENDARS,
)
from xarray.tests.test_dataset import (
    create_append_string_length_mismatch_test_data,
    create_append_test_data,
    create_test_data,
)

try:
    import netCDF4 as nc4
except ImportError:
    pass

try:
    import dask
    import dask.array as da
except ImportError:
    pass

have_zarr_kvstore = False
try:
    from zarr.storage import KVStore

    have_zarr_kvstore = True
except ImportError:
    KVStore = None

have_zarr_v3 = False
try:
    # as of Zarr v2.13 these imports require environment variable
    # ZARR_V3_EXPERIMENTAL_API=1
    from zarr import DirectoryStoreV3, KVStoreV3

    have_zarr_v3 = True
except ImportError:
    KVStoreV3 = None

ON_WINDOWS = sys.platform == "win32"
default_value = object()
dask_array_type = array_type("dask")

if TYPE_CHECKING:
    from xarray.backends.api import T_NetcdfEngine, T_NetcdfTypes


def open_example_dataset(name, *args, **kwargs) -> Dataset:
    return open_dataset(
        os.path.join(os.path.dirname(__file__), "data", name), *args, **kwargs
    )


def open_example_mfdataset(names, *args, **kwargs) -> Dataset:
    return open_mfdataset(
        [os.path.join(os.path.dirname(__file__), "data", name) for name in names],
        *args,
        **kwargs,
    )


def create_masked_and_scaled_data(dtype: np.dtype) -> Dataset:
    x = np.array([np.nan, np.nan, 10, 10.1, 10.2], dtype=dtype)
    encoding = {
        "_FillValue": -1,
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
        "dtype": "i2",
    }
    return Dataset({"x": ("t", x, {}, encoding)})


def create_encoded_masked_and_scaled_data(dtype: np.dtype) -> Dataset:
    attributes = {
        "_FillValue": -1,
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    return Dataset(
        {"x": ("t", np.array([-1, -1, 0, 1, 2], dtype=np.int16), attributes)}
    )


def create_unsigned_masked_scaled_data(dtype: np.dtype) -> Dataset:
    encoding = {
        "_FillValue": 255,
        "_Unsigned": "true",
        "dtype": "i1",
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    x = np.array([10.0, 10.1, 22.7, 22.8, np.nan], dtype=dtype)
    return Dataset({"x": ("t", x, {}, encoding)})


def create_encoded_unsigned_masked_scaled_data(dtype: np.dtype) -> Dataset:
    # These are values as written to the file: the _FillValue will
    # be represented in the signed form.
    attributes = {
        "_FillValue": -1,
        "_Unsigned": "true",
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    # Create unsigned data corresponding to [0, 1, 127, 128, 255] unsigned
    sb = np.asarray([0, 1, 127, -128, -1], dtype="i1")
    return Dataset({"x": ("t", sb, attributes)})


def create_bad_unsigned_masked_scaled_data(dtype: np.dtype) -> Dataset:
    encoding = {
        "_FillValue": 255,
        "_Unsigned": True,
        "dtype": "i1",
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    x = np.array([10.0, 10.1, 22.7, 22.8, np.nan], dtype=dtype)
    return Dataset({"x": ("t", x, {}, encoding)})


def create_bad_encoded_unsigned_masked_scaled_data(dtype: np.dtype) -> Dataset:
    # These are values as written to the file: the _FillValue will
    # be represented in the signed form.
    attributes = {
        "_FillValue": -1,
        "_Unsigned": True,
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    # Create signed data corresponding to [0, 1, 127, 128, 255] unsigned
    sb = np.asarray([0, 1, 127, -128, -1], dtype="i1")
    return Dataset({"x": ("t", sb, attributes)})


def create_signed_masked_scaled_data(dtype: np.dtype) -> Dataset:
    encoding = {
        "_FillValue": -127,
        "_Unsigned": "false",
        "dtype": "i1",
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    x = np.array([-1.0, 10.1, 22.7, np.nan], dtype=dtype)
    return Dataset({"x": ("t", x, {}, encoding)})


def create_encoded_signed_masked_scaled_data(dtype: np.dtype) -> Dataset:
    # These are values as written to the file: the _FillValue will
    # be represented in the signed form.
    attributes = {
        "_FillValue": -127,
        "_Unsigned": "false",
        "add_offset": dtype.type(10),
        "scale_factor": dtype.type(0.1),
    }
    # Create signed data corresponding to [0, 1, 127, 128, 255] unsigned
    sb = np.asarray([-110, 1, 127, -127], dtype="i1")
    return Dataset({"x": ("t", sb, attributes)})


def create_boolean_data() -> Dataset:
    attributes = {"units": "-"}
    return Dataset({"x": ("t", [True, False, False, True], attributes)})


class TestCommon:
    def test_robust_getitem(self) -> None:
        class UnreliableArrayFailure(Exception):
            pass

        class UnreliableArray:
            def __init__(self, array, failures=1):
                self.array = array
                self.failures = failures

            def __getitem__(self, key):
                if self.failures > 0:
                    self.failures -= 1
                    raise UnreliableArrayFailure
                return self.array[key]

        array = UnreliableArray([0])
        with pytest.raises(UnreliableArrayFailure):
            array[0]
        assert array[0] == 0

        actual = robust_getitem(array, 0, catch=UnreliableArrayFailure, initial_delay=0)
        assert actual == 0


class NetCDF3Only:
    netcdf3_formats: tuple[T_NetcdfTypes, ...] = ("NETCDF3_CLASSIC", "NETCDF3_64BIT")

    @requires_scipy
    def test_dtype_coercion_error(self) -> None:
        """Failing dtype coercion should lead to an error"""
        for dtype, format in itertools.product(
            _nc3_dtype_coercions, self.netcdf3_formats
        ):
            if dtype == "bool":
                # coerced upcast (bool to int8) ==> can never fail
                continue

            # Using the largest representable value, create some data that will
            # no longer compare equal after the coerced downcast
            maxval = np.iinfo(dtype).max
            x = np.array([0, 1, 2, maxval], dtype=dtype)
            ds = Dataset({"x": ("t", x, {})})

            with create_tmp_file(allow_cleanup_failure=False) as path:
                with pytest.raises(ValueError, match="could not safely cast"):
                    ds.to_netcdf(path, format=format)


class DatasetIOBase:
    engine: T_NetcdfEngine | None = None
    file_format: T_NetcdfTypes | None = None

    def create_store(self):
        raise NotImplementedError()

    @contextlib.contextmanager
    def roundtrip(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        if save_kwargs is None:
            save_kwargs = {}
        if open_kwargs is None:
            open_kwargs = {}
        with create_tmp_file(allow_cleanup_failure=allow_cleanup_failure) as path:
            self.save(data, path, **save_kwargs)
            with self.open(path, **open_kwargs) as ds:
                yield ds

    @contextlib.contextmanager
    def roundtrip_append(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        if save_kwargs is None:
            save_kwargs = {}
        if open_kwargs is None:
            open_kwargs = {}
        with create_tmp_file(allow_cleanup_failure=allow_cleanup_failure) as path:
            for i, key in enumerate(data.variables):
                mode = "a" if i > 0 else "w"
                self.save(data[[key]], path, mode=mode, **save_kwargs)
            with self.open(path, **open_kwargs) as ds:
                yield ds

    # The save/open methods may be overwritten below
    def save(self, dataset, path, **kwargs):
        return dataset.to_netcdf(
            path, engine=self.engine, format=self.file_format, **kwargs
        )

    @contextlib.contextmanager
    def open(self, path, **kwargs):
        with open_dataset(path, engine=self.engine, **kwargs) as ds:
            yield ds

    def test_zero_dimensional_variable(self) -> None:
        expected = create_test_data()
        expected["float_var"] = ([], 1.0e9, {"units": "units of awesome"})
        expected["bytes_var"] = ([], b"foobar")
        expected["string_var"] = ([], "foobar")
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)

    def test_write_store(self) -> None:
        expected = create_test_data()
        with self.create_store() as store:
            expected.dump_to_store(store)
            # we need to cf decode the store because it has time and
            # non-dimension coordinates
            with xr.decode_cf(store) as actual:
                assert_allclose(expected, actual)

    def check_dtypes_roundtripped(self, expected, actual):
        for k in expected.variables:
            expected_dtype = expected.variables[k].dtype

            # For NetCDF3, the backend should perform dtype coercion
            if (
                isinstance(self, NetCDF3Only)
                and str(expected_dtype) in _nc3_dtype_coercions
            ):
                expected_dtype = np.dtype(_nc3_dtype_coercions[str(expected_dtype)])

            actual_dtype = actual.variables[k].dtype
            # TODO: check expected behavior for string dtypes more carefully
            string_kinds = {"O", "S", "U"}
            assert expected_dtype == actual_dtype or (
                expected_dtype.kind in string_kinds
                and actual_dtype.kind in string_kinds
            )

    def test_roundtrip_test_data(self) -> None:
        expected = create_test_data()
        with self.roundtrip(expected) as actual:
            self.check_dtypes_roundtripped(expected, actual)
            assert_identical(expected, actual)

    def test_load(self) -> None:
        expected = create_test_data()

        @contextlib.contextmanager
        def assert_loads(vars=None):
            if vars is None:
                vars = expected
            with self.roundtrip(expected) as actual:
                for k, v in actual.variables.items():
                    # IndexVariables are eagerly loaded into memory
                    assert v._in_memory == (k in actual.dims)
                yield actual
                for k, v in actual.variables.items():
                    if k in vars:
                        assert v._in_memory
                assert_identical(expected, actual)

        with pytest.raises(AssertionError):
            # make sure the contextmanager works!
            with assert_loads() as ds:
                pass

        with assert_loads() as ds:
            ds.load()

        with assert_loads(["var1", "dim1", "dim2"]) as ds:
            ds["var1"].load()

        # verify we can read data even after closing the file
        with self.roundtrip(expected) as ds:
            actual = ds.load()
        assert_identical(expected, actual)

    def test_dataset_compute(self) -> None:
        expected = create_test_data()

        with self.roundtrip(expected) as actual:
            # Test Dataset.compute()
            for k, v in actual.variables.items():
                # IndexVariables are eagerly cached
                assert v._in_memory == (k in actual.dims)

            computed = actual.compute()

            for k, v in actual.variables.items():
                assert v._in_memory == (k in actual.dims)
            for v in computed.variables.values():
                assert v._in_memory

            assert_identical(expected, actual)
            assert_identical(expected, computed)

    def test_pickle(self) -> None:
        expected = Dataset({"foo": ("x", [42])})
        with self.roundtrip(expected, allow_cleanup_failure=ON_WINDOWS) as roundtripped:
            with roundtripped:
                # Windows doesn't like reopening an already open file
                raw_pickle = pickle.dumps(roundtripped)
            with pickle.loads(raw_pickle) as unpickled_ds:
                assert_identical(expected, unpickled_ds)

    @pytest.mark.filterwarnings("ignore:deallocating CachingFileManager")
    def test_pickle_dataarray(self) -> None:
        expected = Dataset({"foo": ("x", [42])})
        with self.roundtrip(expected, allow_cleanup_failure=ON_WINDOWS) as roundtripped:
            with roundtripped:
                raw_pickle = pickle.dumps(roundtripped["foo"])
            # TODO: figure out how to explicitly close the file for the
            # unpickled DataArray?
            unpickled = pickle.loads(raw_pickle)
            assert_identical(expected["foo"], unpickled)

    def test_dataset_caching(self) -> None:
        expected = Dataset({"foo": ("x", [5, 6, 7])})
        with self.roundtrip(expected) as actual:
            assert isinstance(actual.foo.variable._data, indexing.MemoryCachedArray)
            assert not actual.foo.variable._in_memory
            actual.foo.values  # cache
            assert actual.foo.variable._in_memory

        with self.roundtrip(expected, open_kwargs={"cache": False}) as actual:
            assert isinstance(actual.foo.variable._data, indexing.CopyOnWriteArray)
            assert not actual.foo.variable._in_memory
            actual.foo.values  # no caching
            assert not actual.foo.variable._in_memory

    @pytest.mark.filterwarnings("ignore:deallocating CachingFileManager")
    def test_roundtrip_None_variable(self) -> None:
        expected = Dataset({None: (("x", "y"), [[0, 1], [2, 3]])})
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_object_dtype(self) -> None:
        floats = np.array([0.0, 0.0, 1.0, 2.0, 3.0], dtype=object)
        floats_nans = np.array([np.nan, np.nan, 1.0, 2.0, 3.0], dtype=object)
        bytes_ = np.array([b"ab", b"cdef", b"g"], dtype=object)
        bytes_nans = np.array([b"ab", b"cdef", np.nan], dtype=object)
        strings = np.array(["ab", "cdef", "g"], dtype=object)
        strings_nans = np.array(["ab", "cdef", np.nan], dtype=object)
        all_nans = np.array([np.nan, np.nan], dtype=object)
        original = Dataset(
            {
                "floats": ("a", floats),
                "floats_nans": ("a", floats_nans),
                "bytes": ("b", bytes_),
                "bytes_nans": ("b", bytes_nans),
                "strings": ("b", strings),
                "strings_nans": ("b", strings_nans),
                "all_nans": ("c", all_nans),
                "nan": ([], np.nan),
            }
        )
        expected = original.copy(deep=True)
        with self.roundtrip(original) as actual:
            try:
                assert_identical(expected, actual)
            except AssertionError:
                # Most stores use '' for nans in strings, but some don't.
                # First try the ideal case (where the store returns exactly)
                # the original Dataset), then try a more realistic case.
                # This currently includes all netCDF files when encoding is not
                # explicitly set.
                # https://github.com/pydata/xarray/issues/1647
                expected["bytes_nans"][-1] = b""
                expected["strings_nans"][-1] = ""
                assert_identical(expected, actual)

    def test_roundtrip_string_data(self) -> None:
        expected = Dataset({"x": ("t", ["ab", "cdef"])})
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_string_encoded_characters(self) -> None:
        expected = Dataset({"x": ("t", ["ab", "cdef"])})
        expected["x"].encoding["dtype"] = "S1"
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)
            assert actual["x"].encoding["_Encoding"] == "utf-8"

        expected["x"].encoding["_Encoding"] = "ascii"
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)
            assert actual["x"].encoding["_Encoding"] == "ascii"

    def test_roundtrip_numpy_datetime_data(self) -> None:
        times = pd.to_datetime(["2000-01-01", "2000-01-02", "NaT"])
        expected = Dataset({"t": ("t", times), "t0": times[0]})
        kwargs = {"encoding": {"t0": {"units": "days since 1950-01-01"}}}
        with self.roundtrip(expected, save_kwargs=kwargs) as actual:
            assert_identical(expected, actual)
            assert actual.t0.encoding["units"] == "days since 1950-01-01"

    @requires_cftime
    def test_roundtrip_cftime_datetime_data(self) -> None:
        from xarray.tests.test_coding_times import _all_cftime_date_types

        date_types = _all_cftime_date_types()
        for date_type in date_types.values():
            times = [date_type(1, 1, 1), date_type(1, 1, 2)]
            expected = Dataset({"t": ("t", times), "t0": times[0]})
            kwargs = {"encoding": {"t0": {"units": "days since 0001-01-01"}}}
            expected_decoded_t = np.array(times)
            expected_decoded_t0 = np.array([date_type(1, 1, 1)])
            expected_calendar = times[0].calendar

            with warnings.catch_warnings():
                if expected_calendar in {"proleptic_gregorian", "standard"}:
                    warnings.filterwarnings("ignore", "Unable to decode time axis")

                with self.roundtrip(expected, save_kwargs=kwargs) as actual:
                    abs_diff = abs(actual.t.values - expected_decoded_t)
                    assert (abs_diff <= np.timedelta64(1, "s")).all()
                    assert (
                        actual.t.encoding["units"]
                        == "days since 0001-01-01 00:00:00.000000"
                    )
                    assert actual.t.encoding["calendar"] == expected_calendar

                    abs_diff = abs(actual.t0.values - expected_decoded_t0)
                    assert (abs_diff <= np.timedelta64(1, "s")).all()
                    assert actual.t0.encoding["units"] == "days since 0001-01-01"
                    assert actual.t.encoding["calendar"] == expected_calendar

    def test_roundtrip_timedelta_data(self) -> None:
        time_deltas = pd.to_timedelta(["1h", "2h", "NaT"])
        expected = Dataset({"td": ("td", time_deltas), "td0": time_deltas[0]})
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_float64_data(self) -> None:
        expected = Dataset({"x": ("y", np.array([1.0, 2.0, np.pi], dtype="float64"))})
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_example_1_netcdf(self) -> None:
        with open_example_dataset("example_1.nc") as expected:
            with self.roundtrip(expected) as actual:
                # we allow the attributes to differ since that
                # will depend on the encoding used.  For example,
                # without CF encoding 'actual' will end up with
                # a dtype attribute.
                assert_equal(expected, actual)

    def test_roundtrip_coordinates(self) -> None:
        original = Dataset(
            {"foo": ("x", [0, 1])}, {"x": [2, 3], "y": ("a", [42]), "z": ("x", [4, 5])}
        )

        with self.roundtrip(original) as actual:
            assert_identical(original, actual)

        original["foo"].encoding["coordinates"] = "y"
        with self.roundtrip(original, open_kwargs={"decode_coords": False}) as expected:
            # check roundtripping when decode_coords=False
            with self.roundtrip(
                expected, open_kwargs={"decode_coords": False}
            ) as actual:
                assert_identical(expected, actual)

    def test_roundtrip_global_coordinates(self) -> None:
        original = Dataset(
            {"foo": ("x", [0, 1])}, {"x": [2, 3], "y": ("a", [42]), "z": ("x", [4, 5])}
        )
        with self.roundtrip(original) as actual:
            assert_identical(original, actual)

        # test that global "coordinates" is as expected
        _, attrs = encode_dataset_coordinates(original)
        assert attrs["coordinates"] == "y"

        # test warning when global "coordinates" is already set
        original.attrs["coordinates"] = "foo"
        with pytest.warns(SerializationWarning):
            _, attrs = encode_dataset_coordinates(original)
            assert attrs["coordinates"] == "foo"

    def test_roundtrip_coordinates_with_space(self) -> None:
        original = Dataset(coords={"x": 0, "y z": 1})
        expected = Dataset({"y z": 1}, {"x": 0})
        with pytest.warns(SerializationWarning):
            with self.roundtrip(original) as actual:
                assert_identical(expected, actual)

    def test_roundtrip_boolean_dtype(self) -> None:
        original = create_boolean_data()
        assert original["x"].dtype == "bool"
        with self.roundtrip(original) as actual:
            assert_identical(original, actual)
            assert actual["x"].dtype == "bool"
            # this checks for preserving dtype during second roundtrip
            # see https://github.com/pydata/xarray/issues/7652#issuecomment-1476956975
            with self.roundtrip(actual) as actual2:
                assert_identical(original, actual2)
                assert actual2["x"].dtype == "bool"

    def test_orthogonal_indexing(self) -> None:
        in_memory = create_test_data()
        with self.roundtrip(in_memory) as on_disk:
            indexers = {"dim1": [1, 2, 0], "dim2": [3, 2, 0, 3], "dim3": np.arange(5)}
            expected = in_memory.isel(indexers)
            actual = on_disk.isel(**indexers)
            # make sure the array is not yet loaded into memory
            assert not actual["var1"].variable._in_memory
            assert_identical(expected, actual)
            # do it twice, to make sure we're switched from orthogonal -> numpy
            # when we cached the values
            actual = on_disk.isel(**indexers)
            assert_identical(expected, actual)

    def test_vectorized_indexing(self) -> None:
        in_memory = create_test_data()
        with self.roundtrip(in_memory) as on_disk:
            indexers = {
                "dim1": DataArray([0, 2, 0], dims="a"),
                "dim2": DataArray([0, 2, 3], dims="a"),
            }
            expected = in_memory.isel(indexers)
            actual = on_disk.isel(**indexers)
            # make sure the array is not yet loaded into memory
            assert not actual["var1"].variable._in_memory
            assert_identical(expected, actual.load())
            # do it twice, to make sure we're switched from
            # vectorized -> numpy when we cached the values
            actual = on_disk.isel(**indexers)
            assert_identical(expected, actual)

        def multiple_indexing(indexers):
            # make sure a sequence of lazy indexings certainly works.
            with self.roundtrip(in_memory) as on_disk:
                actual = on_disk["var3"]
                expected = in_memory["var3"]
                for ind in indexers:
                    actual = actual.isel(ind)
                    expected = expected.isel(ind)
                    # make sure the array is not yet loaded into memory
                    assert not actual.variable._in_memory
                assert_identical(expected, actual.load())

        # two-staged vectorized-indexing
        indexers2 = [
            {
                "dim1": DataArray([[0, 7], [2, 6], [3, 5]], dims=["a", "b"]),
                "dim3": DataArray([[0, 4], [1, 3], [2, 2]], dims=["a", "b"]),
            },
            {"a": DataArray([0, 1], dims=["c"]), "b": DataArray([0, 1], dims=["c"])},
        ]
        multiple_indexing(indexers2)

        # vectorized-slice mixed
        indexers3 = [
            {
                "dim1": DataArray([[0, 7], [2, 6], [3, 5]], dims=["a", "b"]),
                "dim3": slice(None, 10),
            }
        ]
        multiple_indexing(indexers3)

        # vectorized-integer mixed
        indexers4 = [
            {"dim3": 0},
            {"dim1": DataArray([[0, 7], [2, 6], [3, 5]], dims=["a", "b"])},
            {"a": slice(None, None, 2)},
        ]
        multiple_indexing(indexers4)

        # vectorized-integer mixed
        indexers5 = [
            {"dim3": 0},
            {"dim1": DataArray([[0, 7], [2, 6], [3, 5]], dims=["a", "b"])},
            {"a": 1, "b": 0},
        ]
        multiple_indexing(indexers5)

    def test_vectorized_indexing_negative_step(self) -> None:
        # use dask explicitly when present
        open_kwargs: dict[str, Any] | None
        if has_dask:
            open_kwargs = {"chunks": {}}
        else:
            open_kwargs = None
        in_memory = create_test_data()

        def multiple_indexing(indexers):
            # make sure a sequence of lazy indexings certainly works.
            with self.roundtrip(in_memory, open_kwargs=open_kwargs) as on_disk:
                actual = on_disk["var3"]
                expected = in_memory["var3"]
                for ind in indexers:
                    actual = actual.isel(ind)
                    expected = expected.isel(ind)
                    # make sure the array is not yet loaded into memory
                    assert not actual.variable._in_memory
                assert_identical(expected, actual.load())

        # with negative step slice.
        indexers = [
            {
                "dim1": DataArray([[0, 7], [2, 6], [3, 5]], dims=["a", "b"]),
                "dim3": slice(-1, 1, -1),
            }
        ]
        multiple_indexing(indexers)

        # with negative step slice.
        indexers = [
            {
                "dim1": DataArray([[0, 7], [2, 6], [3, 5]], dims=["a", "b"]),
                "dim3": slice(-1, 1, -2),
            }
        ]
        multiple_indexing(indexers)

    def test_outer_indexing_reversed(self) -> None:
        # regression test for GH6560
        ds = xr.Dataset(
            {"z": (("t", "p", "y", "x"), np.ones((1, 1, 31, 40)))},
        )

        with self.roundtrip(ds) as on_disk:
            subset = on_disk.isel(t=[0], p=0).z[:, ::10, ::10][:, ::-1, :]
            assert subset.sizes == subset.load().sizes

    def test_isel_dataarray(self) -> None:
        # Make sure isel works lazily. GH:issue:1688
        in_memory = create_test_data()
        with self.roundtrip(in_memory) as on_disk:
            expected = in_memory.isel(dim2=in_memory["dim2"] < 3)
            actual = on_disk.isel(dim2=on_disk["dim2"] < 3)
            assert_identical(expected, actual)

    def validate_array_type(self, ds):
        # Make sure that only NumpyIndexingAdapter stores a bare np.ndarray.
        def find_and_validate_array(obj):
            # recursively called function. obj: array or array wrapper.
            if hasattr(obj, "array"):
                if isinstance(obj.array, indexing.ExplicitlyIndexed):
                    find_and_validate_array(obj.array)
                else:
                    if isinstance(obj.array, np.ndarray):
                        assert isinstance(obj, indexing.NumpyIndexingAdapter)
                    elif isinstance(obj.array, dask_array_type):
                        assert isinstance(obj, indexing.DaskIndexingAdapter)
                    elif isinstance(obj.array, pd.Index):
                        assert isinstance(obj, indexing.PandasIndexingAdapter)
                    else:
                        raise TypeError(f"{type(obj.array)} is wrapped by {type(obj)}")

        for k, v in ds.variables.items():
            find_and_validate_array(v._data)

    def test_array_type_after_indexing(self) -> None:
        in_memory = create_test_data()
        with self.roundtrip(in_memory) as on_disk:
            self.validate_array_type(on_disk)
            indexers = {"dim1": [1, 2, 0], "dim2": [3, 2, 0, 3], "dim3": np.arange(5)}
            expected = in_memory.isel(indexers)
            actual = on_disk.isel(**indexers)
            assert_identical(expected, actual)
            self.validate_array_type(actual)
            # do it twice, to make sure we're switched from orthogonal -> numpy
            # when we cached the values
            actual = on_disk.isel(**indexers)
            assert_identical(expected, actual)
            self.validate_array_type(actual)

    def test_dropna(self) -> None:
        # regression test for GH:issue:1694
        a = np.random.randn(4, 3)
        a[1, 1] = np.nan
        in_memory = xr.Dataset(
            {"a": (("y", "x"), a)}, coords={"y": np.arange(4), "x": np.arange(3)}
        )

        assert_identical(
            in_memory.dropna(dim="x"), in_memory.isel(x=slice(None, None, 2))
        )

        with self.roundtrip(in_memory) as on_disk:
            self.validate_array_type(on_disk)
            expected = in_memory.dropna(dim="x")
            actual = on_disk.dropna(dim="x")
            assert_identical(expected, actual)

    def test_ondisk_after_print(self) -> None:
        """Make sure print does not load file into memory"""
        in_memory = create_test_data()
        with self.roundtrip(in_memory) as on_disk:
            repr(on_disk)
            assert not on_disk["var1"]._in_memory


class CFEncodedBase(DatasetIOBase):
    def test_roundtrip_bytes_with_fill_value(self) -> None:
        values = np.array([b"ab", b"cdef", np.nan], dtype=object)
        encoding = {"_FillValue": b"X", "dtype": "S1"}
        original = Dataset({"x": ("t", values, {}, encoding)})
        expected = original.copy(deep=True)
        with self.roundtrip(original) as actual:
            assert_identical(expected, actual)

        original = Dataset({"x": ("t", values, {}, {"_FillValue": b""})})
        with self.roundtrip(original) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_string_with_fill_value_nchar(self) -> None:
        values = np.array(["ab", "cdef", np.nan], dtype=object)
        expected = Dataset({"x": ("t", values)})

        encoding = {"dtype": "S1", "_FillValue": b"X"}
        original = Dataset({"x": ("t", values, {}, encoding)})
        # Not supported yet.
        with pytest.raises(NotImplementedError):
            with self.roundtrip(original) as actual:
                assert_identical(expected, actual)

    def test_roundtrip_empty_vlen_string_array(self) -> None:
        # checks preserving vlen dtype for empty arrays GH7862
        dtype = create_vlen_dtype(str)
        original = Dataset({"a": np.array([], dtype=dtype)})
        assert check_vlen_dtype(original["a"].dtype) == str
        with self.roundtrip(original) as actual:
            assert_identical(original, actual)
            if np.issubdtype(actual["a"].dtype, object):
                # only check metadata for capable backends
                # eg. NETCDF3 based backends do not roundtrip metadata
                if actual["a"].dtype.metadata is not None:
                    assert check_vlen_dtype(actual["a"].dtype) == str
            else:
                assert actual["a"].dtype == np.dtype("<U1")

    @pytest.mark.parametrize(
        "decoded_fn, encoded_fn",
        [
            (
                create_unsigned_masked_scaled_data,
                create_encoded_unsigned_masked_scaled_data,
            ),
            pytest.param(
                create_bad_unsigned_masked_scaled_data,
                create_bad_encoded_unsigned_masked_scaled_data,
                marks=pytest.mark.xfail(reason="Bad _Unsigned attribute."),
            ),
            (
                create_signed_masked_scaled_data,
                create_encoded_signed_masked_scaled_data,
            ),
            (create_masked_and_scaled_data, create_encoded_masked_and_scaled_data),
        ],
    )
    @pytest.mark.parametrize("dtype", [np.dtype("float64"), np.dtype("float32")])
    def test_roundtrip_mask_and_scale(self, decoded_fn, encoded_fn, dtype) -> None:
        if hasattr(self, "zarr_version") and dtype == np.float32:
            pytest.skip("float32 will be treated as float64 in zarr")
        decoded = decoded_fn(dtype)
        encoded = encoded_fn(dtype)
        with self.roundtrip(decoded) as actual:
            for k in decoded.variables:
                assert decoded.variables[k].dtype == actual.variables[k].dtype
            assert_allclose(decoded, actual, decode_bytes=False)

        with self.roundtrip(decoded, open_kwargs=dict(decode_cf=False)) as actual:
            # TODO: this assumes that all roundtrips will first
            # encode.  Is that something we want to test for?
            for k in encoded.variables:
                assert encoded.variables[k].dtype == actual.variables[k].dtype
            assert_allclose(encoded, actual, decode_bytes=False)

        with self.roundtrip(encoded, open_kwargs=dict(decode_cf=False)) as actual:
            for k in encoded.variables:
                assert encoded.variables[k].dtype == actual.variables[k].dtype
            assert_allclose(encoded, actual, decode_bytes=False)

        # make sure roundtrip encoding didn't change the
        # original dataset.
        assert_allclose(encoded, encoded_fn(dtype), decode_bytes=False)

        with self.roundtrip(encoded) as actual:
            for k in decoded.variables:
                assert decoded.variables[k].dtype == actual.variables[k].dtype
            assert_allclose(decoded, actual, decode_bytes=False)

    @staticmethod
    def _create_cf_dataset():
        original = Dataset(
            dict(
                variable=(
                    ("ln_p", "latitude", "longitude"),
                    np.arange(8, dtype="f4").reshape(2, 2, 2),
                    {"ancillary_variables": "std_devs det_lim"},
                ),
                std_devs=(
                    ("ln_p", "latitude", "longitude"),
                    np.arange(0.1, 0.9, 0.1).reshape(2, 2, 2),
                    {"standard_name": "standard_error"},
                ),
                det_lim=(
                    (),
                    0.1,
                    {"standard_name": "detection_minimum"},
                ),
            ),
            dict(
                latitude=("latitude", [0, 1], {"units": "degrees_north"}),
                longitude=("longitude", [0, 1], {"units": "degrees_east"}),
                latlon=((), -1, {"grid_mapping_name": "latitude_longitude"}),
                latitude_bnds=(("latitude", "bnds2"), [[0, 1], [1, 2]]),
                longitude_bnds=(("longitude", "bnds2"), [[0, 1], [1, 2]]),
                areas=(
                    ("latitude", "longitude"),
                    [[1, 1], [1, 1]],
                    {"units": "degree^2"},
                ),
                ln_p=(
                    "ln_p",
                    [1.0, 0.5],
                    {
                        "standard_name": "atmosphere_ln_pressure_coordinate",
                        "computed_standard_name": "air_pressure",
                    },
                ),
                P0=((), 1013.25, {"units": "hPa"}),
            ),
        )
        original["variable"].encoding.update(
            {"cell_measures": "area: areas", "grid_mapping": "latlon"},
        )
        original.coords["latitude"].encoding.update(
            dict(grid_mapping="latlon", bounds="latitude_bnds")
        )
        original.coords["longitude"].encoding.update(
            dict(grid_mapping="latlon", bounds="longitude_bnds")
        )
        original.coords["ln_p"].encoding.update({"formula_terms": "p0: P0 lev : ln_p"})
        return original

    def test_grid_mapping_and_bounds_are_not_coordinates_in_file(self) -> None:
        original = self._create_cf_dataset()
        with create_tmp_file() as tmp_file:
            original.to_netcdf(tmp_file)
            with open_dataset(tmp_file, decode_coords=False) as ds:
                assert ds.coords["latitude"].attrs["bounds"] == "latitude_bnds"
                assert ds.coords["longitude"].attrs["bounds"] == "longitude_bnds"
                assert "coordinates" not in ds["variable"].attrs
                assert "coordinates" not in ds.attrs

    def test_coordinate_variables_after_dataset_roundtrip(self) -> None:
        original = self._create_cf_dataset()
        with self.roundtrip(original, open_kwargs={"decode_coords": "all"}) as actual:
            assert_identical(actual, original)

        with self.roundtrip(original) as actual:
            expected = original.reset_coords(
                ["latitude_bnds", "longitude_bnds", "areas", "P0", "latlon"]
            )
            # equal checks that coords and data_vars are equal which
            # should be enough
            # identical would require resetting a number of attributes
            # skip that.
            assert_equal(actual, expected)

    def test_grid_mapping_and_bounds_are_coordinates_after_dataarray_roundtrip(
        self,
    ) -> None:
        original = self._create_cf_dataset()
        # The DataArray roundtrip should have the same warnings as the
        # Dataset, but we already tested for those, so just go for the
        # new warnings.  It would appear that there is no way to tell
        # pytest "This warning and also this warning should both be
        # present".
        # xarray/tests/test_conventions.py::TestCFEncodedDataStore
        # needs the to_dataset. The other backends should be fine
        # without it.
        with pytest.warns(
            UserWarning,
            match=(
                r"Variable\(s\) referenced in bounds not in variables: "
                r"\['l(at|ong)itude_bnds'\]"
            ),
        ):
            with self.roundtrip(
                original["variable"].to_dataset(), open_kwargs={"decode_coords": "all"}
            ) as actual:
                assert_identical(actual, original["variable"].to_dataset())

    @requires_iris
    def test_coordinate_variables_after_iris_roundtrip(self) -> None:
        original = self._create_cf_dataset()
        iris_cube = original["variable"].to_iris()
        actual = DataArray.from_iris(iris_cube)
        # Bounds will be missing (xfail)
        del original.coords["latitude_bnds"], original.coords["longitude_bnds"]
        # Ancillary vars will be missing
        # Those are data_vars, and will be dropped when grabbing the variable
        assert_identical(actual, original["variable"])

    def test_coordinates_encoding(self) -> None:
        def equals_latlon(obj):
            return obj == "lat lon" or obj == "lon lat"

        original = Dataset(
            {"temp": ("x", [0, 1]), "precip": ("x", [0, -1])},
            {"lat": ("x", [2, 3]), "lon": ("x", [4, 5])},
        )
        with self.roundtrip(original) as actual:
            assert_identical(actual, original)
        with create_tmp_file() as tmp_file:
            original.to_netcdf(tmp_file)
            with open_dataset(tmp_file, decode_coords=False) as ds:
                assert equals_latlon(ds["temp"].attrs["coordinates"])
                assert equals_latlon(ds["precip"].attrs["coordinates"])
                assert "coordinates" not in ds.attrs
                assert "coordinates" not in ds["lat"].attrs
                assert "coordinates" not in ds["lon"].attrs

        modified = original.drop_vars(["temp", "precip"])
        with self.roundtrip(modified) as actual:
            assert_identical(actual, modified)
        with create_tmp_file() as tmp_file:
            modified.to_netcdf(tmp_file)
            with open_dataset(tmp_file, decode_coords=False) as ds:
                assert equals_latlon(ds.attrs["coordinates"])
                assert "coordinates" not in ds["lat"].attrs
                assert "coordinates" not in ds["lon"].attrs

        original["temp"].encoding["coordinates"] = "lat"
        with self.roundtrip(original) as actual:
            assert_identical(actual, original)
        original["precip"].encoding["coordinates"] = "lat"
        with create_tmp_file() as tmp_file:
            original.to_netcdf(tmp_file)
            with open_dataset(tmp_file, decode_coords=True) as ds:
                assert "lon" not in ds["temp"].encoding["coordinates"]
                assert "lon" not in ds["precip"].encoding["coordinates"]
                assert "coordinates" not in ds["lat"].encoding
                assert "coordinates" not in ds["lon"].encoding

    def test_roundtrip_endian(self) -> None:
        ds = Dataset(
            {
                "x": np.arange(3, 10, dtype=">i2"),
                "y": np.arange(3, 20, dtype="<i4"),
                "z": np.arange(3, 30, dtype="=i8"),
                "w": ("x", np.arange(3, 10, dtype=float)),
            }
        )

        with self.roundtrip(ds) as actual:
            # technically these datasets are slightly different,
            # one hold mixed endian data (ds) the other should be
            # all big endian (actual).  assertDatasetIdentical
            # should still pass though.
            assert_identical(ds, actual)

        if self.engine == "netcdf4":
            ds["z"].encoding["endian"] = "big"
            with pytest.raises(NotImplementedError):
                with self.roundtrip(ds) as actual:
                    pass

    def test_invalid_dataarray_names_raise(self) -> None:
        te = (TypeError, "string or None")
        ve = (ValueError, "string must be length 1 or")
        data = np.random.random((2, 2))
        da = xr.DataArray(data)
        for name, (error, msg) in zip([0, (4, 5), True, ""], [te, te, te, ve]):
            ds = Dataset({name: da})
            with pytest.raises(error) as excinfo:
                with self.roundtrip(ds):
                    pass
            excinfo.match(msg)
            excinfo.match(repr(name))

    def test_encoding_kwarg(self) -> None:
        ds = Dataset({"x": ("y", np.arange(10.0))})

        kwargs: dict[str, Any] = dict(encoding={"x": {"dtype": "f4"}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            encoded_dtype = actual.x.encoding["dtype"]
            # On OS X, dtype sometimes switches endianness for unclear reasons
            assert encoded_dtype.kind == "f" and encoded_dtype.itemsize == 4
        assert ds.x.encoding == {}

        kwargs = dict(encoding={"x": {"foo": "bar"}})
        with pytest.raises(ValueError, match=r"unexpected encoding"):
            with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                pass

        kwargs = dict(encoding={"x": "foo"})
        with pytest.raises(ValueError, match=r"must be castable"):
            with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                pass

        kwargs = dict(encoding={"invalid": {}})
        with pytest.raises(KeyError):
            with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                pass

    def test_encoding_kwarg_dates(self) -> None:
        ds = Dataset({"t": pd.date_range("2000-01-01", periods=3)})
        units = "days since 1900-01-01"
        kwargs = dict(encoding={"t": {"units": units}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert actual.t.encoding["units"] == units
            assert_identical(actual, ds)

    def test_encoding_kwarg_fixed_width_string(self) -> None:
        # regression test for GH2149
        for strings in [[b"foo", b"bar", b"baz"], ["foo", "bar", "baz"]]:
            ds = Dataset({"x": strings})
            kwargs = dict(encoding={"x": {"dtype": "S1"}})
            with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                assert actual["x"].encoding["dtype"] == "S1"
                assert_identical(actual, ds)

    def test_default_fill_value(self) -> None:
        # Test default encoding for float:
        ds = Dataset({"x": ("y", np.arange(10.0))})
        kwargs = dict(encoding={"x": {"dtype": "f4"}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert math.isnan(actual.x.encoding["_FillValue"])
        assert ds.x.encoding == {}

        # Test default encoding for int:
        ds = Dataset({"x": ("y", np.arange(10.0))})
        kwargs = dict(encoding={"x": {"dtype": "int16"}})
        with warnings.catch_warnings():
            warnings.filterwarnings("ignore", ".*floating point data as an integer")
            with self.roundtrip(ds, save_kwargs=kwargs) as actual:
                assert "_FillValue" not in actual.x.encoding
        assert ds.x.encoding == {}

        # Test default encoding for implicit int:
        ds = Dataset({"x": ("y", np.arange(10, dtype="int16"))})
        with self.roundtrip(ds) as actual:
            assert "_FillValue" not in actual.x.encoding
        assert ds.x.encoding == {}

    def test_explicitly_omit_fill_value(self) -> None:
        ds = Dataset({"x": ("y", [np.pi, -np.pi])})
        ds.x.encoding["_FillValue"] = None
        with self.roundtrip(ds) as actual:
            assert "_FillValue" not in actual.x.encoding

    def test_explicitly_omit_fill_value_via_encoding_kwarg(self) -> None:
        ds = Dataset({"x": ("y", [np.pi, -np.pi])})
        kwargs = dict(encoding={"x": {"_FillValue": None}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert "_FillValue" not in actual.x.encoding
        assert ds.y.encoding == {}

    def test_explicitly_omit_fill_value_in_coord(self) -> None:
        ds = Dataset({"x": ("y", [np.pi, -np.pi])}, coords={"y": [0.0, 1.0]})
        ds.y.encoding["_FillValue"] = None
        with self.roundtrip(ds) as actual:
            assert "_FillValue" not in actual.y.encoding

    def test_explicitly_omit_fill_value_in_coord_via_encoding_kwarg(self) -> None:
        ds = Dataset({"x": ("y", [np.pi, -np.pi])}, coords={"y": [0.0, 1.0]})
        kwargs = dict(encoding={"y": {"_FillValue": None}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert "_FillValue" not in actual.y.encoding
        assert ds.y.encoding == {}

    def test_encoding_same_dtype(self) -> None:
        ds = Dataset({"x": ("y", np.arange(10.0, dtype="f4"))})
        kwargs = dict(encoding={"x": {"dtype": "f4"}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            encoded_dtype = actual.x.encoding["dtype"]
            # On OS X, dtype sometimes switches endianness for unclear reasons
            assert encoded_dtype.kind == "f" and encoded_dtype.itemsize == 4
        assert ds.x.encoding == {}

    def test_append_write(self) -> None:
        # regression for GH1215
        data = create_test_data()
        with self.roundtrip_append(data) as actual:
            assert_identical(data, actual)

    def test_append_overwrite_values(self) -> None:
        # regression for GH1215
        data = create_test_data()
        with create_tmp_file(allow_cleanup_failure=False) as tmp_file:
            self.save(data, tmp_file, mode="w")
            data["var2"][:] = -999
            data["var9"] = data["var2"] * 3
            self.save(data[["var2", "var9"]], tmp_file, mode="a")
            with self.open(tmp_file) as actual:
                assert_identical(data, actual)

    def test_append_with_invalid_dim_raises(self) -> None:
        data = create_test_data()
        with create_tmp_file(allow_cleanup_failure=False) as tmp_file:
            self.save(data, tmp_file, mode="w")
            data["var9"] = data["var2"] * 3
            data = data.isel(dim1=slice(2, 6))  # modify one dimension
            with pytest.raises(
                ValueError, match=r"Unable to update size for existing dimension"
            ):
                self.save(data, tmp_file, mode="a")

    def test_multiindex_not_implemented(self) -> None:
        ds = Dataset(coords={"y": ("x", [1, 2]), "z": ("x", ["a", "b"])}).set_index(
            x=["y", "z"]
        )
        with pytest.raises(NotImplementedError, match=r"MultiIndex"):
            with self.roundtrip(ds):
                pass

        # regression GH8628 (can serialize reset multi-index level coordinates)
        ds_reset = ds.reset_index("x")
        with self.roundtrip(ds_reset) as actual:
            assert_identical(actual, ds_reset)


class NetCDFBase(CFEncodedBase):
    """Tests for all netCDF3 and netCDF4 backends."""

    @pytest.mark.skipif(
        ON_WINDOWS, reason="Windows does not allow modifying open files"
    )
    def test_refresh_from_disk(self) -> None:
        # regression test for https://github.com/pydata/xarray/issues/4862

        with create_tmp_file() as example_1_path:
            with create_tmp_file() as example_1_modified_path:
                with open_example_dataset("example_1.nc") as example_1:
                    self.save(example_1, example_1_path)

                    example_1.rh.values += 100
                    self.save(example_1, example_1_modified_path)

                a = open_dataset(example_1_path, engine=self.engine).load()

                # Simulate external process modifying example_1.nc while this script is running
                shutil.copy(example_1_modified_path, example_1_path)

                # Reopen example_1.nc (modified) as `b`; note that `a` has NOT been closed
                b = open_dataset(example_1_path, engine=self.engine).load()

                try:
                    assert not np.array_equal(a.rh.values, b.rh.values)
                finally:
                    a.close()
                    b.close()


_counter = itertools.count()


@contextlib.contextmanager
def create_tmp_file(
    suffix: str = ".nc", allow_cleanup_failure: bool = False
) -> Iterator[str]:
    temp_dir = tempfile.mkdtemp()
    path = os.path.join(temp_dir, f"temp-{next(_counter)}{suffix}")
    try:
        yield path
    finally:
        try:
            shutil.rmtree(temp_dir)
        except OSError:
            if not allow_cleanup_failure:
                raise


@contextlib.contextmanager
def create_tmp_files(
    nfiles: int, suffix: str = ".nc", allow_cleanup_failure: bool = False
) -> Iterator[list[str]]:
    with ExitStack() as stack:
        files = [
            stack.enter_context(create_tmp_file(suffix, allow_cleanup_failure))
            for _ in range(nfiles)
        ]
        yield files


class NetCDF4Base(NetCDFBase):
    """Tests for both netCDF4-python and h5netcdf."""

    engine: T_NetcdfEngine = "netcdf4"

    def test_open_group(self) -> None:
        # Create a netCDF file with a dataset stored within a group
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, "w") as rootgrp:
                foogrp = rootgrp.createGroup("foo")
                ds = foogrp
                ds.createDimension("time", size=10)
                x = np.arange(10)
                ds.createVariable("x", np.int32, dimensions=("time",))
                ds.variables["x"][:] = x

            expected = Dataset()
            expected["x"] = ("time", x)

            # check equivalent ways to specify group
            for group in "foo", "/foo", "foo/", "/foo/":
                with self.open(tmp_file, group=group) as actual:
                    assert_equal(actual["x"], expected["x"])

            # check that missing group raises appropriate exception
            with pytest.raises(OSError):
                open_dataset(tmp_file, group="bar")
            with pytest.raises(ValueError, match=r"must be a string"):
                open_dataset(tmp_file, group=(1, 2, 3))

    def test_open_subgroup(self) -> None:
        # Create a netCDF file with a dataset stored within a group within a
        # group
        with create_tmp_file() as tmp_file:
            rootgrp = nc4.Dataset(tmp_file, "w")
            foogrp = rootgrp.createGroup("foo")
            bargrp = foogrp.createGroup("bar")
            ds = bargrp
            ds.createDimension("time", size=10)
            x = np.arange(10)
            ds.createVariable("x", np.int32, dimensions=("time",))
            ds.variables["x"][:] = x
            rootgrp.close()

            expected = Dataset()
            expected["x"] = ("time", x)

            # check equivalent ways to specify group
            for group in "foo/bar", "/foo/bar", "foo/bar/", "/foo/bar/":
                with self.open(tmp_file, group=group) as actual:
                    assert_equal(actual["x"], expected["x"])

    def test_write_groups(self) -> None:
        data1 = create_test_data()
        data2 = data1 * 2
        with create_tmp_file() as tmp_file:
            self.save(data1, tmp_file, group="data/1")
            self.save(data2, tmp_file, group="data/2", mode="a")
            with self.open(tmp_file, group="data/1") as actual1:
                assert_identical(data1, actual1)
            with self.open(tmp_file, group="data/2") as actual2:
                assert_identical(data2, actual2)

    @pytest.mark.parametrize(
        "input_strings, is_bytes",
        [
            ([b"foo", b"bar", b"baz"], True),
            (["foo", "bar", "baz"], False),
            (["foó", "bár", "baź"], False),
        ],
    )
    def test_encoding_kwarg_vlen_string(
        self, input_strings: list[str], is_bytes: bool
    ) -> None:
        original = Dataset({"x": input_strings})

        expected_string = ["foo", "bar", "baz"] if is_bytes else input_strings
        expected = Dataset({"x": expected_string})
        kwargs = dict(encoding={"x": {"dtype": str}})
        with self.roundtrip(original, save_kwargs=kwargs) as actual:
            assert actual["x"].encoding["dtype"] == "<U3"
            assert actual["x"].dtype == "<U3"
            assert_identical(actual, expected)

    @pytest.mark.parametrize("fill_value", ["XXX", "", "bár"])
    def test_roundtrip_string_with_fill_value_vlen(self, fill_value: str) -> None:
        values = np.array(["ab", "cdef", np.nan], dtype=object)
        expected = Dataset({"x": ("t", values)})

        original = Dataset({"x": ("t", values, {}, {"_FillValue": fill_value})})
        with self.roundtrip(original) as actual:
            assert_identical(expected, actual)

        original = Dataset({"x": ("t", values, {}, {"_FillValue": ""})})
        with self.roundtrip(original) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_character_array(self) -> None:
        with create_tmp_file() as tmp_file:
            values = np.array([["a", "b", "c"], ["d", "e", "f"]], dtype="S")

            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("x", 2)
                nc.createDimension("string3", 3)
                v = nc.createVariable("x", np.dtype("S1"), ("x", "string3"))
                v[:] = values

            values = np.array(["abc", "def"], dtype="S")
            expected = Dataset({"x": ("x", values)})
            with open_dataset(tmp_file) as actual:
                assert_identical(expected, actual)
                # regression test for #157
                with self.roundtrip(actual) as roundtripped:
                    assert_identical(expected, roundtripped)

    def test_default_to_char_arrays(self) -> None:
        data = Dataset({"x": np.array(["foo", "zzzz"], dtype="S")})
        with self.roundtrip(data) as actual:
            assert_identical(data, actual)
            assert actual["x"].dtype == np.dtype("S4")

    def test_open_encodings(self) -> None:
        # Create a netCDF file with explicit time units
        # and make sure it makes it into the encodings
        # and survives a round trip
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, "w") as ds:
                ds.createDimension("time", size=10)
                ds.createVariable("time", np.int32, dimensions=("time",))
                units = "days since 1999-01-01"
                ds.variables["time"].setncattr("units", units)
                ds.variables["time"][:] = np.arange(10) + 4

            expected = Dataset()

            time = pd.date_range("1999-01-05", periods=10)
            encoding = {"units": units, "dtype": np.dtype("int32")}
            expected["time"] = ("time", time, {}, encoding)

            with open_dataset(tmp_file) as actual:
                assert_equal(actual["time"], expected["time"])
                actual_encoding = {
                    k: v
                    for k, v in actual["time"].encoding.items()
                    if k in expected["time"].encoding
                }
                assert actual_encoding == expected["time"].encoding

    def test_dump_encodings(self) -> None:
        # regression test for #709
        ds = Dataset({"x": ("y", np.arange(10.0))})
        kwargs = dict(encoding={"x": {"zlib": True}})
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert actual.x.encoding["zlib"]

    def test_dump_and_open_encodings(self) -> None:
        # Create a netCDF file with explicit time units
        # and make sure it makes it into the encodings
        # and survives a round trip
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, "w") as ds:
                ds.createDimension("time", size=10)
                ds.createVariable("time", np.int32, dimensions=("time",))
                units = "days since 1999-01-01"
                ds.variables["time"].setncattr("units", units)
                ds.variables["time"][:] = np.arange(10) + 4

            with open_dataset(tmp_file) as xarray_dataset:
                with create_tmp_file() as tmp_file2:
                    xarray_dataset.to_netcdf(tmp_file2)
                    with nc4.Dataset(tmp_file2, "r") as ds:
                        assert ds.variables["time"].getncattr("units") == units
                        assert_array_equal(ds.variables["time"], np.arange(10) + 4)

    def test_compression_encoding_legacy(self) -> None:
        data = create_test_data()
        data["var2"].encoding.update(
            {
                "zlib": True,
                "chunksizes": (5, 5),
                "fletcher32": True,
                "shuffle": True,
                "original_shape": data.var2.shape,
            }
        )
        with self.roundtrip(data) as actual:
            for k, v in data["var2"].encoding.items():
                assert v == actual["var2"].encoding[k]

        # regression test for #156
        expected = data.isel(dim1=0)
        with self.roundtrip(expected) as actual:
            assert_equal(expected, actual)

    def test_encoding_kwarg_compression(self) -> None:
        ds = Dataset({"x": np.arange(10.0)})
        encoding = dict(
            dtype="f4",
            zlib=True,
            complevel=9,
            fletcher32=True,
            chunksizes=(5,),
            shuffle=True,
        )
        kwargs = dict(encoding=dict(x=encoding))

        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert_equal(actual, ds)
            assert actual.x.encoding["dtype"] == "f4"
            assert actual.x.encoding["zlib"]
            assert actual.x.encoding["complevel"] == 9
            assert actual.x.encoding["fletcher32"]
            assert actual.x.encoding["chunksizes"] == (5,)
            assert actual.x.encoding["shuffle"]

        assert ds.x.encoding == {}

    def test_keep_chunksizes_if_no_original_shape(self) -> None:
        ds = Dataset({"x": [1, 2, 3]})
        chunksizes = (2,)
        ds.variables["x"].encoding = {"chunksizes": chunksizes}

        with self.roundtrip(ds) as actual:
            assert_identical(ds, actual)
            assert_array_equal(
                ds["x"].encoding["chunksizes"], actual["x"].encoding["chunksizes"]
            )

    def test_preferred_chunks_is_present(self) -> None:
        ds = Dataset({"x": [1, 2, 3]})
        chunksizes = (2,)
        ds.variables["x"].encoding = {"chunksizes": chunksizes}

        with self.roundtrip(ds) as actual:
            assert actual["x"].encoding["preferred_chunks"] == {"x": 2}

    @requires_dask
    def test_auto_chunking_is_based_on_disk_chunk_sizes(self) -> None:
        x_size = y_size = 1000
        y_chunksize = y_size
        x_chunksize = 10

        with dask.config.set({"array.chunk-size": "100KiB"}):
            with self.chunked_roundtrip(
                (1, y_size, x_size),
                (1, y_chunksize, x_chunksize),
                open_kwargs={"chunks": "auto"},
            ) as ds:
                t_chunks, y_chunks, x_chunks = ds["image"].data.chunks
                assert all(np.asanyarray(y_chunks) == y_chunksize)
                # Check that the chunk size is a multiple of the file chunk size
                assert all(np.asanyarray(x_chunks) % x_chunksize == 0)

    @requires_dask
    def test_base_chunking_uses_disk_chunk_sizes(self) -> None:
        x_size = y_size = 1000
        y_chunksize = y_size
        x_chunksize = 10

        with self.chunked_roundtrip(
            (1, y_size, x_size),
            (1, y_chunksize, x_chunksize),
            open_kwargs={"chunks": {}},
        ) as ds:
            for chunksizes, expected in zip(
                ds["image"].data.chunks, (1, y_chunksize, x_chunksize)
            ):
                assert all(np.asanyarray(chunksizes) == expected)

    @contextlib.contextmanager
    def chunked_roundtrip(
        self,
        array_shape: tuple[int, int, int],
        chunk_sizes: tuple[int, int, int],
        open_kwargs: dict[str, Any] | None = None,
    ) -> Generator[Dataset, None, None]:
        t_size, y_size, x_size = array_shape
        t_chunksize, y_chunksize, x_chunksize = chunk_sizes

        image = xr.DataArray(
            np.arange(t_size * x_size * y_size, dtype=np.int16).reshape(
                (t_size, y_size, x_size)
            ),
            dims=["t", "y", "x"],
        )
        image.encoding = {"chunksizes": (t_chunksize, y_chunksize, x_chunksize)}
        dataset = xr.Dataset(dict(image=image))

        with self.roundtrip(dataset, open_kwargs=open_kwargs) as ds:
            yield ds

    def test_preferred_chunks_are_disk_chunk_sizes(self) -> None:
        x_size = y_size = 1000
        y_chunksize = y_size
        x_chunksize = 10

        with self.chunked_roundtrip(
            (1, y_size, x_size), (1, y_chunksize, x_chunksize)
        ) as ds:
            assert ds["image"].encoding["preferred_chunks"] == {
                "t": 1,
                "y": y_chunksize,
                "x": x_chunksize,
            }

    def test_encoding_chunksizes_unlimited(self) -> None:
        # regression test for GH1225
        ds = Dataset({"x": [1, 2, 3], "y": ("x", [2, 3, 4])})
        ds.variables["x"].encoding = {
            "zlib": False,
            "shuffle": False,
            "complevel": 0,
            "fletcher32": False,
            "contiguous": False,
            "chunksizes": (2**20,),
            "original_shape": (3,),
        }
        with self.roundtrip(ds) as actual:
            assert_equal(ds, actual)

    def test_mask_and_scale(self) -> None:
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("t", 5)
                nc.createVariable("x", "int16", ("t",), fill_value=-1)
                v = nc.variables["x"]
                v.set_auto_maskandscale(False)
                v.add_offset = 10
                v.scale_factor = 0.1
                v[:] = np.array([-1, -1, 0, 1, 2])
                dtype = type(v.scale_factor)

            # first make sure netCDF4 reads the masked and scaled data
            # correctly
            with nc4.Dataset(tmp_file, mode="r") as nc:
                expected = np.ma.array(
                    [-1, -1, 10, 10.1, 10.2], mask=[True, True, False, False, False]
                )
                actual = nc.variables["x"][:]
                assert_array_equal(expected, actual)

            # now check xarray
            with open_dataset(tmp_file) as ds:
                expected = create_masked_and_scaled_data(np.dtype(dtype))
                assert_identical(expected, ds)

    def test_0dimensional_variable(self) -> None:
        # This fix verifies our work-around to this netCDF4-python bug:
        # https://github.com/Unidata/netcdf4-python/pull/220
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                v = nc.createVariable("x", "int16")
                v[...] = 123

            with open_dataset(tmp_file) as ds:
                expected = Dataset({"x": ((), 123)})
                assert_identical(expected, ds)

    def test_read_variable_len_strings(self) -> None:
        with create_tmp_file() as tmp_file:
            values = np.array(["foo", "bar", "baz"], dtype=object)

            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("x", 3)
                v = nc.createVariable("x", str, ("x",))
                v[:] = values

            expected = Dataset({"x": ("x", values)})
            for kwargs in [{}, {"decode_cf": True}]:
                with open_dataset(tmp_file, **cast(dict, kwargs)) as actual:
                    assert_identical(expected, actual)

    def test_encoding_unlimited_dims(self) -> None:
        ds = Dataset({"x": ("y", np.arange(10.0))})
        with self.roundtrip(ds, save_kwargs=dict(unlimited_dims=["y"])) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)
        ds.encoding = {"unlimited_dims": ["y"]}
        with self.roundtrip(ds) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)

    def test_raise_on_forward_slashes_in_names(self) -> None:
        # test for forward slash in variable names and dimensions
        # see GH 7943
        data_vars: list[dict[str, Any]] = [
            {"PASS/FAIL": (["PASSFAIL"], np.array([0]))},
            {"PASS/FAIL": np.array([0])},
            {"PASSFAIL": (["PASS/FAIL"], np.array([0]))},
        ]
        for dv in data_vars:
            ds = Dataset(data_vars=dv)
            with pytest.raises(ValueError, match="Forward slashes '/' are not allowed"):
                with self.roundtrip(ds):
                    pass

    @requires_netCDF4
    def test_encoding_enum__no_fill_value(self):
        with create_tmp_file() as tmp_file:
            cloud_type_dict = {"clear": 0, "cloudy": 1}
            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("time", size=2)
                cloud_type = nc.createEnumType("u1", "cloud_type", cloud_type_dict)
                v = nc.createVariable(
                    "clouds",
                    cloud_type,
                    "time",
                    fill_value=None,
                )
                v[:] = 1
            with open_dataset(tmp_file) as original:
                save_kwargs = {}
                if self.engine == "h5netcdf":
                    save_kwargs["invalid_netcdf"] = True
                with self.roundtrip(original, save_kwargs=save_kwargs) as actual:
                    assert_equal(original, actual)
                    assert (
                        actual.clouds.encoding["dtype"].metadata["enum"]
                        == cloud_type_dict
                    )
                    if self.engine != "h5netcdf":
                        # not implemented in h5netcdf yet
                        assert (
                            actual.clouds.encoding["dtype"].metadata["enum_name"]
                            == "cloud_type"
                        )

    @requires_netCDF4
    def test_encoding_enum__multiple_variable_with_enum(self):
        with create_tmp_file() as tmp_file:
            cloud_type_dict = {"clear": 0, "cloudy": 1, "missing": 255}
            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("time", size=2)
                cloud_type = nc.createEnumType("u1", "cloud_type", cloud_type_dict)
                nc.createVariable(
                    "clouds",
                    cloud_type,
                    "time",
                    fill_value=255,
                )
                nc.createVariable(
                    "tifa",
                    cloud_type,
                    "time",
                    fill_value=255,
                )
            with open_dataset(tmp_file) as original:
                save_kwargs = {}
                if self.engine == "h5netcdf":
                    save_kwargs["invalid_netcdf"] = True
                with self.roundtrip(original, save_kwargs=save_kwargs) as actual:
                    assert_equal(original, actual)
                    assert (
                        actual.clouds.encoding["dtype"] == actual.tifa.encoding["dtype"]
                    )
                    assert (
                        actual.clouds.encoding["dtype"].metadata
                        == actual.tifa.encoding["dtype"].metadata
                    )
                    assert (
                        actual.clouds.encoding["dtype"].metadata["enum"]
                        == cloud_type_dict
                    )
                    if self.engine != "h5netcdf":
                        # not implemented in h5netcdf yet
                        assert (
                            actual.clouds.encoding["dtype"].metadata["enum_name"]
                            == "cloud_type"
                        )

    @requires_netCDF4
    def test_encoding_enum__error_multiple_variable_with_changing_enum(self):
        """
        Given 2 variables, if they share the same enum type,
        the 2 enum definition should be identical.
        """
        with create_tmp_file() as tmp_file:
            cloud_type_dict = {"clear": 0, "cloudy": 1, "missing": 255}
            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("time", size=2)
                cloud_type = nc.createEnumType("u1", "cloud_type", cloud_type_dict)
                nc.createVariable(
                    "clouds",
                    cloud_type,
                    "time",
                    fill_value=255,
                )
                nc.createVariable(
                    "tifa",
                    cloud_type,
                    "time",
                    fill_value=255,
                )
            with open_dataset(tmp_file) as original:
                assert (
                    original.clouds.encoding["dtype"].metadata
                    == original.tifa.encoding["dtype"].metadata
                )
                modified_enum = original.clouds.encoding["dtype"].metadata["enum"]
                modified_enum.update({"neblig": 2})
                original.clouds.encoding["dtype"] = np.dtype(
                    "u1",
                    metadata={"enum": modified_enum, "enum_name": "cloud_type"},
                )
                if self.engine != "h5netcdf":
                    # not implemented yet in h5netcdf
                    with pytest.raises(
                        ValueError,
                        match=(
                            "Cannot save variable .*"
                            " because an enum `cloud_type` already exists in the Dataset .*"
                        ),
                    ):
                        with self.roundtrip(original):
                            pass


@requires_netCDF4
class TestNetCDF4Data(NetCDF4Base):
    @contextlib.contextmanager
    def create_store(self):
        with create_tmp_file() as tmp_file:
            with backends.NetCDF4DataStore.open(tmp_file, mode="w") as store:
                yield store

    def test_variable_order(self) -> None:
        # doesn't work with scipy or h5py :(
        ds = Dataset()
        ds["a"] = 1
        ds["z"] = 2
        ds["b"] = 3
        ds.coords["c"] = 4

        with self.roundtrip(ds) as actual:
            assert list(ds.variables) == list(actual.variables)

    def test_unsorted_index_raises(self) -> None:
        # should be fixed in netcdf4 v1.2.1
        random_data = np.random.random(size=(4, 6))
        dim0 = [0, 1, 2, 3]
        dim1 = [0, 2, 1, 3, 5, 4]  # We will sort this in a later step
        da = xr.DataArray(
            data=random_data,
            dims=("dim0", "dim1"),
            coords={"dim0": dim0, "dim1": dim1},
            name="randovar",
        )
        ds = da.to_dataset()

        with self.roundtrip(ds) as ondisk:
            inds = np.argsort(dim1)
            ds2 = ondisk.isel(dim1=inds)
            # Older versions of NetCDF4 raise an exception here, and if so we
            # want to ensure we improve (that is, replace) the error message
            try:
                ds2.randovar.values
            except IndexError as err:
                assert "first by calling .load" in str(err)

    def test_setncattr_string(self) -> None:
        list_of_strings = ["list", "of", "strings"]
        one_element_list_of_strings = ["one element"]
        one_string = "one string"
        attrs = {
            "foo": list_of_strings,
            "bar": one_element_list_of_strings,
            "baz": one_string,
        }
        ds = Dataset({"x": ("y", [1, 2, 3], attrs)}, attrs=attrs)

        with self.roundtrip(ds) as actual:
            for totest in [actual, actual["x"]]:
                assert_array_equal(list_of_strings, totest.attrs["foo"])
                assert_array_equal(one_element_list_of_strings, totest.attrs["bar"])
                assert one_string == totest.attrs["baz"]

    @pytest.mark.parametrize(
        "compression",
        [
            None,
            "zlib",
            "szip",
            "zstd",
            "blosc_lz",
            "blosc_lz4",
            "blosc_lz4hc",
            "blosc_zlib",
            "blosc_zstd",
        ],
    )
    @requires_netCDF4_1_6_2_or_above
    @pytest.mark.xfail(ON_WINDOWS, reason="new compression not yet implemented")
    def test_compression_encoding(self, compression: str | None) -> None:
        data = create_test_data(dim_sizes=(20, 80, 10))
        encoding_params: dict[str, Any] = dict(compression=compression, blosc_shuffle=1)
        data["var2"].encoding.update(encoding_params)
        data["var2"].encoding.update(
            {
                "chunksizes": (20, 40),
                "original_shape": data.var2.shape,
                "blosc_shuffle": 1,
                "fletcher32": False,
            }
        )
        with self.roundtrip(data) as actual:
            expected_encoding = data["var2"].encoding.copy()
            # compression does not appear in the retrieved encoding, that differs
            # from the input encoding. shuffle also chantges. Here we modify the
            # expected encoding to account for this
            compression = expected_encoding.pop("compression")
            blosc_shuffle = expected_encoding.pop("blosc_shuffle")
            if compression is not None:
                if "blosc" in compression and blosc_shuffle:
                    expected_encoding["blosc"] = {
                        "compressor": compression,
                        "shuffle": blosc_shuffle,
                    }
                    expected_encoding["shuffle"] = False
                elif compression == "szip":
                    expected_encoding["szip"] = {
                        "coding": "nn",
                        "pixels_per_block": 8,
                    }
                    expected_encoding["shuffle"] = False
                else:
                    # This will set a key like zlib=true which is what appears in
                    # the encoding when we read it.
                    expected_encoding[compression] = True
                    if compression == "zstd":
                        expected_encoding["shuffle"] = False
            else:
                expected_encoding["shuffle"] = False

            actual_encoding = actual["var2"].encoding
            assert expected_encoding.items() <= actual_encoding.items()
        if (
            encoding_params["compression"] is not None
            and "blosc" not in encoding_params["compression"]
        ):
            # regression test for #156
            expected = data.isel(dim1=0)
            with self.roundtrip(expected) as actual:
                assert_equal(expected, actual)

    @pytest.mark.skip(reason="https://github.com/Unidata/netcdf4-python/issues/1195")
    def test_refresh_from_disk(self) -> None:
        super().test_refresh_from_disk()


@requires_netCDF4
class TestNetCDF4AlreadyOpen:
    def test_base_case(self) -> None:
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                v = nc.createVariable("x", "int")
                v[...] = 42

            nc = nc4.Dataset(tmp_file, mode="r")
            store = backends.NetCDF4DataStore(nc)
            with open_dataset(store) as ds:
                expected = Dataset({"x": ((), 42)})
                assert_identical(expected, ds)

    def test_group(self) -> None:
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                group = nc.createGroup("g")
                v = group.createVariable("x", "int")
                v[...] = 42

            nc = nc4.Dataset(tmp_file, mode="r")
            store = backends.NetCDF4DataStore(nc.groups["g"])
            with open_dataset(store) as ds:
                expected = Dataset({"x": ((), 42)})
                assert_identical(expected, ds)

            nc = nc4.Dataset(tmp_file, mode="r")
            store = backends.NetCDF4DataStore(nc, group="g")
            with open_dataset(store) as ds:
                expected = Dataset({"x": ((), 42)})
                assert_identical(expected, ds)

            with nc4.Dataset(tmp_file, mode="r") as nc:
                with pytest.raises(ValueError, match="must supply a root"):
                    backends.NetCDF4DataStore(nc.groups["g"], group="g")

    def test_deepcopy(self) -> None:
        # regression test for https://github.com/pydata/xarray/issues/4425
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("x", 10)
                v = nc.createVariable("y", np.int32, ("x",))
                v[:] = np.arange(10)

            h5 = nc4.Dataset(tmp_file, mode="r")
            store = backends.NetCDF4DataStore(h5)
            with open_dataset(store) as ds:
                copied = ds.copy(deep=True)
                expected = Dataset({"y": ("x", np.arange(10))})
                assert_identical(expected, copied)


@requires_netCDF4
@requires_dask
@pytest.mark.filterwarnings("ignore:deallocating CachingFileManager")
class TestNetCDF4ViaDaskData(TestNetCDF4Data):
    @contextlib.contextmanager
    def roundtrip(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        if open_kwargs is None:
            open_kwargs = {}
        if save_kwargs is None:
            save_kwargs = {}
        open_kwargs.setdefault("chunks", -1)
        with TestNetCDF4Data.roundtrip(
            self, data, save_kwargs, open_kwargs, allow_cleanup_failure
        ) as ds:
            yield ds

    def test_unsorted_index_raises(self) -> None:
        # Skip when using dask because dask rewrites indexers to getitem,
        # dask first pulls items by block.
        pass

    @pytest.mark.skip(reason="caching behavior differs for dask")
    def test_dataset_caching(self) -> None:
        pass

    def test_write_inconsistent_chunks(self) -> None:
        # Construct two variables with the same dimensions, but different
        # chunk sizes.
        x = da.zeros((100, 100), dtype="f4", chunks=(50, 100))
        x = DataArray(data=x, dims=("lat", "lon"), name="x")
        x.encoding["chunksizes"] = (50, 100)
        x.encoding["original_shape"] = (100, 100)
        y = da.ones((100, 100), dtype="f4", chunks=(100, 50))
        y = DataArray(data=y, dims=("lat", "lon"), name="y")
        y.encoding["chunksizes"] = (100, 50)
        y.encoding["original_shape"] = (100, 100)
        # Put them both into the same dataset
        ds = Dataset({"x": x, "y": y})
        with self.roundtrip(ds) as actual:
            assert actual["x"].encoding["chunksizes"] == (50, 100)
            assert actual["y"].encoding["chunksizes"] == (100, 50)


@requires_zarr
class ZarrBase(CFEncodedBase):
    DIMENSION_KEY = "_ARRAY_DIMENSIONS"
    zarr_version = 2
    version_kwargs: dict[str, Any] = {}

    def create_zarr_target(self):
        raise NotImplementedError

    @contextlib.contextmanager
    def create_store(self):
        with self.create_zarr_target() as store_target:
            yield backends.ZarrStore.open_group(
                store_target, mode="w", **self.version_kwargs
            )

    def save(self, dataset, store_target, **kwargs):
        return dataset.to_zarr(store=store_target, **kwargs, **self.version_kwargs)

    @contextlib.contextmanager
    def open(self, store_target, **kwargs):
        with xr.open_dataset(
            store_target, engine="zarr", **kwargs, **self.version_kwargs
        ) as ds:
            yield ds

    @contextlib.contextmanager
    def roundtrip(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        if save_kwargs is None:
            save_kwargs = {}
        if open_kwargs is None:
            open_kwargs = {}
        with self.create_zarr_target() as store_target:
            self.save(data, store_target, **save_kwargs)
            with self.open(store_target, **open_kwargs) as ds:
                yield ds

    @pytest.mark.parametrize("consolidated", [False, True, None])
    def test_roundtrip_consolidated(self, consolidated) -> None:
        if consolidated and self.zarr_version > 2:
            pytest.xfail("consolidated metadata is not supported for zarr v3 yet")
        expected = create_test_data()
        with self.roundtrip(
            expected,
            save_kwargs={"consolidated": consolidated},
            open_kwargs={"backend_kwargs": {"consolidated": consolidated}},
        ) as actual:
            self.check_dtypes_roundtripped(expected, actual)
            assert_identical(expected, actual)

    def test_read_non_consolidated_warning(self) -> None:
        if self.zarr_version > 2:
            pytest.xfail("consolidated metadata is not supported for zarr v3 yet")

        expected = create_test_data()
        with self.create_zarr_target() as store:
            expected.to_zarr(store, consolidated=False, **self.version_kwargs)
            with pytest.warns(
                RuntimeWarning,
                match="Failed to open Zarr store with consolidated",
            ):
                with xr.open_zarr(store, **self.version_kwargs) as ds:
                    assert_identical(ds, expected)

    def test_non_existent_store(self) -> None:
        with pytest.raises(FileNotFoundError, match=r"No such file or directory:"):
            xr.open_zarr(f"{uuid.uuid4()}")

    def test_with_chunkstore(self) -> None:
        expected = create_test_data()
        with (
            self.create_zarr_target() as store_target,
            self.create_zarr_target() as chunk_store,
        ):
            save_kwargs = {"chunk_store": chunk_store}
            self.save(expected, store_target, **save_kwargs)
            # the chunk store must have been populated with some entries
            assert len(chunk_store) > 0
            open_kwargs = {"backend_kwargs": {"chunk_store": chunk_store}}
            with self.open(store_target, **open_kwargs) as ds:
                assert_equal(ds, expected)

    @requires_dask
    def test_auto_chunk(self) -> None:
        original = create_test_data().chunk()

        with self.roundtrip(original, open_kwargs={"chunks": None}) as actual:
            for k, v in actual.variables.items():
                # only index variables should be in memory
                assert v._in_memory == (k in actual.dims)
                # there should be no chunks
                assert v.chunks is None

        with self.roundtrip(original, open_kwargs={"chunks": {}}) as actual:
            for k, v in actual.variables.items():
                # only index variables should be in memory
                assert v._in_memory == (k in actual.dims)
                # chunk size should be the same as original
                assert v.chunks == original[k].chunks

    @requires_dask
    @pytest.mark.filterwarnings("ignore:The specified chunks separate:UserWarning")
    def test_manual_chunk(self) -> None:
        original = create_test_data().chunk({"dim1": 3, "dim2": 4, "dim3": 3})

        # Using chunks = None should return non-chunked arrays
        open_kwargs: dict[str, Any] = {"chunks": None}
        with self.roundtrip(original, open_kwargs=open_kwargs) as actual:
            for k, v in actual.variables.items():
                # only index variables should be in memory
                assert v._in_memory == (k in actual.dims)
                # there should be no chunks
                assert v.chunks is None

        # uniform arrays
        for i in range(2, 6):
            rechunked = original.chunk(chunks=i)
            open_kwargs = {"chunks": i}
            with self.roundtrip(original, open_kwargs=open_kwargs) as actual:
                for k, v in actual.variables.items():
                    # only index variables should be in memory
                    assert v._in_memory == (k in actual.dims)
                    # chunk size should be the same as rechunked
                    assert v.chunks == rechunked[k].chunks

        chunks = {"dim1": 2, "dim2": 3, "dim3": 5}
        rechunked = original.chunk(chunks=chunks)

        open_kwargs = {
            "chunks": chunks,
            "backend_kwargs": {"overwrite_encoded_chunks": True},
        }
        with self.roundtrip(original, open_kwargs=open_kwargs) as actual:
            for k, v in actual.variables.items():
                assert v.chunks == rechunked[k].chunks

            with self.roundtrip(actual) as auto:
                # encoding should have changed
                for k, v in actual.variables.items():
                    assert v.chunks == rechunked[k].chunks

                assert_identical(actual, auto)
                assert_identical(actual.load(), auto.load())

    @requires_dask
    def test_warning_on_bad_chunks(self) -> None:
        original = create_test_data().chunk({"dim1": 4, "dim2": 3, "dim3": 3})

        bad_chunks = (2, {"dim2": (3, 3, 2, 1)})
        for chunks in bad_chunks:
            kwargs = {"chunks": chunks}
            with pytest.warns(UserWarning):
                with self.roundtrip(original, open_kwargs=kwargs) as actual:
                    for k, v in actual.variables.items():
                        # only index variables should be in memory
                        assert v._in_memory == (k in actual.dims)

        good_chunks: tuple[dict[str, Any], ...] = ({"dim2": 3}, {"dim3": (6, 4)}, {})
        for chunks in good_chunks:
            kwargs = {"chunks": chunks}
            with assert_no_warnings():
                with self.roundtrip(original, open_kwargs=kwargs) as actual:
                    for k, v in actual.variables.items():
                        # only index variables should be in memory
                        assert v._in_memory == (k in actual.dims)

    @requires_dask
    def test_deprecate_auto_chunk(self) -> None:
        original = create_test_data().chunk()
        with pytest.raises(TypeError):
            with self.roundtrip(original, open_kwargs={"auto_chunk": True}) as actual:
                for k, v in actual.variables.items():
                    # only index variables should be in memory
                    assert v._in_memory == (k in actual.dims)
                    # chunk size should be the same as original
                    assert v.chunks == original[k].chunks

        with pytest.raises(TypeError):
            with self.roundtrip(original, open_kwargs={"auto_chunk": False}) as actual:
                for k, v in actual.variables.items():
                    # only index variables should be in memory
                    assert v._in_memory == (k in actual.dims)
                    # there should be no chunks
                    assert v.chunks is None

    @requires_dask
    def test_write_uneven_dask_chunks(self) -> None:
        # regression for GH#2225
        original = create_test_data().chunk({"dim1": 3, "dim2": 4, "dim3": 3})
        with self.roundtrip(original, open_kwargs={"chunks": {}}) as actual:
            for k, v in actual.data_vars.items():
                print(k)
                assert v.chunks == actual[k].chunks

    def test_chunk_encoding(self) -> None:
        # These datasets have no dask chunks. All chunking specified in
        # encoding
        data = create_test_data()
        chunks = (5, 5)
        data["var2"].encoding.update({"chunks": chunks})

        with self.roundtrip(data) as actual:
            assert chunks == actual["var2"].encoding["chunks"]

        # expect an error with non-integer chunks
        data["var2"].encoding.update({"chunks": (5, 4.5)})
        with pytest.raises(TypeError):
            with self.roundtrip(data) as actual:
                pass

    @requires_dask
    @pytest.mark.skipif(
        ON_WINDOWS,
        reason="Very flaky on Windows CI. Can re-enable assuming it starts consistently passing.",
    )
    def test_chunk_encoding_with_dask(self) -> None:
        # These datasets DO have dask chunks. Need to check for various
        # interactions between dask and zarr chunks
        ds = xr.DataArray((np.arange(12)), dims="x", name="var1").to_dataset()

        # - no encoding specified -
        # zarr automatically gets chunk information from dask chunks
        ds_chunk4 = ds.chunk({"x": 4})
        with self.roundtrip(ds_chunk4) as actual:
            assert (4,) == actual["var1"].encoding["chunks"]

        # should fail if dask_chunks are irregular...
        ds_chunk_irreg = ds.chunk({"x": (5, 4, 3)})
        with pytest.raises(ValueError, match=r"uniform chunk sizes."):
            with self.roundtrip(ds_chunk_irreg) as actual:
                pass

        # should fail if encoding["chunks"] clashes with dask_chunks
        badenc = ds.chunk({"x": 4})
        badenc.var1.encoding["chunks"] = (6,)
        with pytest.raises(NotImplementedError, match=r"named 'var1' would overlap"):
            with self.roundtrip(badenc) as actual:
                pass

        # unless...
        with self.roundtrip(badenc, save_kwargs={"safe_chunks": False}) as actual:
            # don't actually check equality because the data could be corrupted
            pass

        # if dask chunks (4) are an integer multiple of zarr chunks (2) it should not fail...
        goodenc = ds.chunk({"x": 4})
        goodenc.var1.encoding["chunks"] = (2,)
        with self.roundtrip(goodenc) as actual:
            pass

        # if initial dask chunks are aligned, size of last dask chunk doesn't matter
        goodenc = ds.chunk({"x": (3, 3, 6)})
        goodenc.var1.encoding["chunks"] = (3,)
        with self.roundtrip(goodenc) as actual:
            pass

        goodenc = ds.chunk({"x": (3, 6, 3)})
        goodenc.var1.encoding["chunks"] = (3,)
        with self.roundtrip(goodenc) as actual:
            pass

        # ... also if the last chunk is irregular
        ds_chunk_irreg = ds.chunk({"x": (5, 5, 2)})
        with self.roundtrip(ds_chunk_irreg) as actual:
            assert (5,) == actual["var1"].encoding["chunks"]
        # re-save Zarr arrays
        with self.roundtrip(ds_chunk_irreg) as original:
            with self.roundtrip(original) as actual:
                assert_identical(original, actual)

        # but itermediate unaligned chunks are bad
        badenc = ds.chunk({"x": (3, 5, 3, 1)})
        badenc.var1.encoding["chunks"] = (3,)
        with pytest.raises(
            NotImplementedError, match=r"would overlap multiple dask chunks"
        ):
            with self.roundtrip(badenc) as actual:
                pass

        # - encoding specified  -
        # specify compatible encodings
        for chunk_enc in 4, (4,):
            ds_chunk4["var1"].encoding.update({"chunks": chunk_enc})
            with self.roundtrip(ds_chunk4) as actual:
                assert (4,) == actual["var1"].encoding["chunks"]

        # TODO: remove this failure once synchronized overlapping writes are
        # supported by xarray
        ds_chunk4["var1"].encoding.update({"chunks": 5})
        with pytest.raises(NotImplementedError, match=r"named 'var1' would overlap"):
            with self.roundtrip(ds_chunk4) as actual:
                pass
        # override option
        with self.roundtrip(ds_chunk4, save_kwargs={"safe_chunks": False}) as actual:
            # don't actually check equality because the data could be corrupted
            pass

    def test_drop_encoding(self):
        with open_example_dataset("example_1.nc") as ds:
            encodings = {v: {**ds[v].encoding} for v in ds.data_vars}
            with self.create_zarr_target() as store:
                ds.to_zarr(store, encoding=encodings)

    def test_hidden_zarr_keys(self) -> None:
        expected = create_test_data()
        with self.create_store() as store:
            expected.dump_to_store(store)
            zarr_group = store.ds

            # check that a variable hidden attribute is present and correct
            # JSON only has a single array type, which maps to list in Python.
            # In contrast, dims in xarray is always a tuple.
            for var in expected.variables.keys():
                dims = zarr_group[var].attrs[self.DIMENSION_KEY]
                assert dims == list(expected[var].dims)

            with xr.decode_cf(store):
                # make sure it is hidden
                for var in expected.variables.keys():
                    assert self.DIMENSION_KEY not in expected[var].attrs

            # put it back and try removing from a variable
            del zarr_group.var2.attrs[self.DIMENSION_KEY]
            with pytest.raises(KeyError):
                with xr.decode_cf(store):
                    pass

    @pytest.mark.parametrize("group", [None, "group1"])
    def test_write_persistence_modes(self, group) -> None:
        original = create_test_data()

        # overwrite mode
        with self.roundtrip(
            original,
            save_kwargs={"mode": "w", "group": group},
            open_kwargs={"group": group},
        ) as actual:
            assert_identical(original, actual)

        # don't overwrite mode
        with self.roundtrip(
            original,
            save_kwargs={"mode": "w-", "group": group},
            open_kwargs={"group": group},
        ) as actual:
            assert_identical(original, actual)

        # make sure overwriting works as expected
        with self.create_zarr_target() as store:
            self.save(original, store)
            # should overwrite with no error
            self.save(original, store, mode="w", group=group)
            with self.open(store, group=group) as actual:
                assert_identical(original, actual)
                with pytest.raises(ValueError):
                    self.save(original, store, mode="w-")

        # check append mode for normal write
        with self.roundtrip(
            original,
            save_kwargs={"mode": "a", "group": group},
            open_kwargs={"group": group},
        ) as actual:
            assert_identical(original, actual)

        # check append mode for append write
        ds, ds_to_append, _ = create_append_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", group=group, **self.version_kwargs)
            ds_to_append.to_zarr(
                store_target, append_dim="time", group=group, **self.version_kwargs
            )
            original = xr.concat([ds, ds_to_append], dim="time")
            actual = xr.open_dataset(
                store_target, group=group, engine="zarr", **self.version_kwargs
            )
            assert_identical(original, actual)

    def test_compressor_encoding(self) -> None:
        original = create_test_data()
        # specify a custom compressor
        import zarr

        blosc_comp = zarr.Blosc(cname="zstd", clevel=3, shuffle=2)
        save_kwargs = dict(encoding={"var1": {"compressor": blosc_comp}})
        with self.roundtrip(original, save_kwargs=save_kwargs) as ds:
            actual = ds["var1"].encoding["compressor"]
            # get_config returns a dictionary of compressor attributes
            assert actual.get_config() == blosc_comp.get_config()

    def test_group(self) -> None:
        original = create_test_data()
        group = "some/random/path"
        with self.roundtrip(
            original, save_kwargs={"group": group}, open_kwargs={"group": group}
        ) as actual:
            assert_identical(original, actual)

    def test_encoding_kwarg_fixed_width_string(self) -> None:
        # not relevant for zarr, since we don't use EncodedStringCoder
        pass

    def test_dataset_caching(self) -> None:
        super().test_dataset_caching()

    def test_append_write(self) -> None:
        super().test_append_write()

    def test_append_with_mode_rplus_success(self) -> None:
        original = Dataset({"foo": ("x", [1])})
        modified = Dataset({"foo": ("x", [2])})
        with self.create_zarr_target() as store:
            original.to_zarr(store, **self.version_kwargs)
            modified.to_zarr(store, mode="r+", **self.version_kwargs)
            with self.open(store) as actual:
                assert_identical(actual, modified)

    def test_append_with_mode_rplus_fails(self) -> None:
        original = Dataset({"foo": ("x", [1])})
        modified = Dataset({"bar": ("x", [2])})
        with self.create_zarr_target() as store:
            original.to_zarr(store, **self.version_kwargs)
            with pytest.raises(
                ValueError, match="dataset contains non-pre-existing variables"
            ):
                modified.to_zarr(store, mode="r+", **self.version_kwargs)

    def test_append_with_invalid_dim_raises(self) -> None:
        ds, ds_to_append, _ = create_append_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
            with pytest.raises(
                ValueError, match="does not match any existing dataset dimensions"
            ):
                ds_to_append.to_zarr(
                    store_target, append_dim="notvalid", **self.version_kwargs
                )

    def test_append_with_no_dims_raises(self) -> None:
        with self.create_zarr_target() as store_target:
            Dataset({"foo": ("x", [1])}).to_zarr(
                store_target, mode="w", **self.version_kwargs
            )
            with pytest.raises(ValueError, match="different dimension names"):
                Dataset({"foo": ("y", [2])}).to_zarr(
                    store_target, mode="a", **self.version_kwargs
                )

    def test_append_with_append_dim_not_set_raises(self) -> None:
        ds, ds_to_append, _ = create_append_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
            with pytest.raises(ValueError, match="different dimension sizes"):
                ds_to_append.to_zarr(store_target, mode="a", **self.version_kwargs)

    def test_append_with_mode_not_a_raises(self) -> None:
        ds, ds_to_append, _ = create_append_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
            with pytest.raises(ValueError, match="cannot set append_dim unless"):
                ds_to_append.to_zarr(
                    store_target, mode="w", append_dim="time", **self.version_kwargs
                )

    def test_append_with_existing_encoding_raises(self) -> None:
        ds, ds_to_append, _ = create_append_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
            with pytest.raises(ValueError, match="but encoding was provided"):
                ds_to_append.to_zarr(
                    store_target,
                    append_dim="time",
                    encoding={"da": {"compressor": None}},
                    **self.version_kwargs,
                )

    @pytest.mark.parametrize("dtype", ["U", "S"])
    def test_append_string_length_mismatch_raises(self, dtype) -> None:
        ds, ds_to_append = create_append_string_length_mismatch_test_data(dtype)
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
            with pytest.raises(ValueError, match="Mismatched dtypes for variable"):
                ds_to_append.to_zarr(
                    store_target, append_dim="time", **self.version_kwargs
                )

    def test_check_encoding_is_consistent_after_append(self) -> None:
        ds, ds_to_append, _ = create_append_test_data()

        # check encoding consistency
        with self.create_zarr_target() as store_target:
            import zarr

            compressor = zarr.Blosc()
            encoding = {"da": {"compressor": compressor}}
            ds.to_zarr(store_target, mode="w", encoding=encoding, **self.version_kwargs)
            ds_to_append.to_zarr(store_target, append_dim="time", **self.version_kwargs)
            actual_ds = xr.open_dataset(
                store_target, engine="zarr", **self.version_kwargs
            )
            actual_encoding = actual_ds["da"].encoding["compressor"]
            assert actual_encoding.get_config() == compressor.get_config()
            assert_identical(
                xr.open_dataset(
                    store_target, engine="zarr", **self.version_kwargs
                ).compute(),
                xr.concat([ds, ds_to_append], dim="time"),
            )

    def test_append_with_new_variable(self) -> None:
        ds, ds_to_append, ds_with_new_var = create_append_test_data()

        # check append mode for new variable
        with self.create_zarr_target() as store_target:
            xr.concat([ds, ds_to_append], dim="time").to_zarr(
                store_target, mode="w", **self.version_kwargs
            )
            ds_with_new_var.to_zarr(store_target, mode="a", **self.version_kwargs)
            combined = xr.concat([ds, ds_to_append], dim="time")
            combined["new_var"] = ds_with_new_var["new_var"]
            assert_identical(
                combined,
                xr.open_dataset(store_target, engine="zarr", **self.version_kwargs),
            )

    def test_append_with_append_dim_no_overwrite(self) -> None:
        ds, ds_to_append, _ = create_append_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, mode="w", **self.version_kwargs)
            original = xr.concat([ds, ds_to_append], dim="time")
            original2 = xr.concat([original, ds_to_append], dim="time")

            # overwrite a coordinate;
            # for mode='a-', this will not get written to the store
            # because it does not have the append_dim as a dim
            lon = ds_to_append.lon.to_numpy().copy()
            lon[:] = -999
            ds_to_append["lon"] = lon
            ds_to_append.to_zarr(
                store_target, mode="a-", append_dim="time", **self.version_kwargs
            )
            actual = xr.open_dataset(store_target, engine="zarr", **self.version_kwargs)
            assert_identical(original, actual)

            # by default, mode="a" will overwrite all coordinates.
            ds_to_append.to_zarr(store_target, append_dim="time", **self.version_kwargs)
            actual = xr.open_dataset(store_target, engine="zarr", **self.version_kwargs)
            lon = original2.lon.to_numpy().copy()
            lon[:] = -999
            original2["lon"] = lon
            assert_identical(original2, actual)

    @requires_dask
    def test_to_zarr_compute_false_roundtrip(self) -> None:
        from dask.delayed import Delayed

        original = create_test_data().chunk()

        with self.create_zarr_target() as store:
            delayed_obj = self.save(original, store, compute=False)
            assert isinstance(delayed_obj, Delayed)

            # make sure target store has not been written to yet
            with pytest.raises(AssertionError):
                with self.open(store) as actual:
                    assert_identical(original, actual)

            delayed_obj.compute()

            with self.open(store) as actual:
                assert_identical(original, actual)

    @requires_dask
    def test_to_zarr_append_compute_false_roundtrip(self) -> None:
        from dask.delayed import Delayed

        ds, ds_to_append, _ = create_append_test_data()
        ds, ds_to_append = ds.chunk(), ds_to_append.chunk()

        with pytest.warns(SerializationWarning):
            with self.create_zarr_target() as store:
                delayed_obj = self.save(ds, store, compute=False, mode="w")
                assert isinstance(delayed_obj, Delayed)

                with pytest.raises(AssertionError):
                    with self.open(store) as actual:
                        assert_identical(ds, actual)

                delayed_obj.compute()

                with self.open(store) as actual:
                    assert_identical(ds, actual)

                delayed_obj = self.save(
                    ds_to_append, store, compute=False, append_dim="time"
                )
                assert isinstance(delayed_obj, Delayed)

                with pytest.raises(AssertionError):
                    with self.open(store) as actual:
                        assert_identical(
                            xr.concat([ds, ds_to_append], dim="time"), actual
                        )

                delayed_obj.compute()

                with self.open(store) as actual:
                    assert_identical(xr.concat([ds, ds_to_append], dim="time"), actual)

    @pytest.mark.parametrize("chunk", [False, True])
    def test_save_emptydim(self, chunk) -> None:
        if chunk and not has_dask:
            pytest.skip("requires dask")
        ds = Dataset({"x": (("a", "b"), np.empty((5, 0))), "y": ("a", [1, 2, 5, 8, 9])})
        if chunk:
            ds = ds.chunk({})  # chunk dataset to save dask array
        with self.roundtrip(ds) as ds_reload:
            assert_identical(ds, ds_reload)

    @requires_dask
    def test_no_warning_from_open_emptydim_with_chunks(self) -> None:
        ds = Dataset({"x": (("a", "b"), np.empty((5, 0)))}).chunk({"a": 1})
        with assert_no_warnings():
            with self.roundtrip(ds, open_kwargs=dict(chunks={"a": 1})) as ds_reload:
                assert_identical(ds, ds_reload)

    @pytest.mark.parametrize("consolidated", [False, True, None])
    @pytest.mark.parametrize("compute", [False, True])
    @pytest.mark.parametrize("use_dask", [False, True])
    @pytest.mark.parametrize("write_empty", [False, True, None])
    def test_write_region(self, consolidated, compute, use_dask, write_empty) -> None:
        if (use_dask or not compute) and not has_dask:
            pytest.skip("requires dask")
        if consolidated and self.zarr_version > 2:
            pytest.xfail("consolidated metadata is not supported for zarr v3 yet")

        zeros = Dataset({"u": (("x",), np.zeros(10))})
        nonzeros = Dataset({"u": (("x",), np.arange(1, 11))})

        if use_dask:
            zeros = zeros.chunk(2)
            nonzeros = nonzeros.chunk(2)

        with self.create_zarr_target() as store:
            zeros.to_zarr(
                store,
                consolidated=consolidated,
                compute=compute,
                encoding={"u": dict(chunks=2)},
                **self.version_kwargs,
            )
            if compute:
                with xr.open_zarr(
                    store, consolidated=consolidated, **self.version_kwargs
                ) as actual:
                    assert_identical(actual, zeros)
            for i in range(0, 10, 2):
                region = {"x": slice(i, i + 2)}
                nonzeros.isel(region).to_zarr(
                    store,
                    region=region,
                    consolidated=consolidated,
                    write_empty_chunks=write_empty,
                    **self.version_kwargs,
                )
            with xr.open_zarr(
                store, consolidated=consolidated, **self.version_kwargs
            ) as actual:
                assert_identical(actual, nonzeros)

    @pytest.mark.parametrize("mode", [None, "r+", "a"])
    def test_write_region_mode(self, mode) -> None:
        zeros = Dataset({"u": (("x",), np.zeros(10))})
        nonzeros = Dataset({"u": (("x",), np.arange(1, 11))})
        with self.create_zarr_target() as store:
            zeros.to_zarr(store, **self.version_kwargs)
            for region in [{"x": slice(5)}, {"x": slice(5, 10)}]:
                nonzeros.isel(region).to_zarr(
                    store, region=region, mode=mode, **self.version_kwargs
                )
            with xr.open_zarr(store, **self.version_kwargs) as actual:
                assert_identical(actual, nonzeros)

    @requires_dask
    def test_write_preexisting_override_metadata(self) -> None:
        """Metadata should be overridden if mode="a" but not in mode="r+"."""
        original = Dataset(
            {"u": (("x",), np.zeros(10), {"variable": "original"})},
            attrs={"global": "original"},
        )
        both_modified = Dataset(
            {"u": (("x",), np.ones(10), {"variable": "modified"})},
            attrs={"global": "modified"},
        )
        global_modified = Dataset(
            {"u": (("x",), np.ones(10), {"variable": "original"})},
            attrs={"global": "modified"},
        )
        only_new_data = Dataset(
            {"u": (("x",), np.ones(10), {"variable": "original"})},
            attrs={"global": "original"},
        )

        with self.create_zarr_target() as store:
            original.to_zarr(store, compute=False, **self.version_kwargs)
            both_modified.to_zarr(store, mode="a", **self.version_kwargs)
            with self.open(store) as actual:
                # NOTE: this arguably incorrect -- we should probably be
                # overriding the variable metadata, too. See the TODO note in
                # ZarrStore.set_variables.
                assert_identical(actual, global_modified)

        with self.create_zarr_target() as store:
            original.to_zarr(store, compute=False, **self.version_kwargs)
            both_modified.to_zarr(store, mode="r+", **self.version_kwargs)
            with self.open(store) as actual:
                assert_identical(actual, only_new_data)

        with self.create_zarr_target() as store:
            original.to_zarr(store, compute=False, **self.version_kwargs)
            # with region, the default mode becomes r+
            both_modified.to_zarr(
                store, region={"x": slice(None)}, **self.version_kwargs
            )
            with self.open(store) as actual:
                assert_identical(actual, only_new_data)

    def test_write_region_errors(self) -> None:
        data = Dataset({"u": (("x",), np.arange(5))})
        data2 = Dataset({"u": (("x",), np.array([10, 11]))})

        @contextlib.contextmanager
        def setup_and_verify_store(expected=data):
            with self.create_zarr_target() as store:
                data.to_zarr(store, **self.version_kwargs)
                yield store
                with self.open(store) as actual:
                    assert_identical(actual, expected)

        # verify the base case works
        expected = Dataset({"u": (("x",), np.array([10, 11, 2, 3, 4]))})
        with setup_and_verify_store(expected) as store:
            data2.to_zarr(store, region={"x": slice(2)}, **self.version_kwargs)

        with setup_and_verify_store() as store:
            with pytest.raises(
                ValueError,
                match=re.escape(
                    "cannot set region unless mode='a', mode='a-', mode='r+' or mode=None"
                ),
            ):
                data.to_zarr(
                    store, region={"x": slice(None)}, mode="w", **self.version_kwargs
                )

        with setup_and_verify_store() as store:
            with pytest.raises(TypeError, match=r"must be a dict"):
                data.to_zarr(store, region=slice(None), **self.version_kwargs)  # type: ignore[call-overload]

        with setup_and_verify_store() as store:
            with pytest.raises(TypeError, match=r"must be slice objects"):
                data2.to_zarr(store, region={"x": [0, 1]}, **self.version_kwargs)  # type: ignore[dict-item]

        with setup_and_verify_store() as store:
            with pytest.raises(ValueError, match=r"step on all slices"):
                data2.to_zarr(
                    store, region={"x": slice(None, None, 2)}, **self.version_kwargs
                )

        with setup_and_verify_store() as store:
            with pytest.raises(
                ValueError,
                match=r"all keys in ``region`` are not in Dataset dimensions",
            ):
                data.to_zarr(store, region={"y": slice(None)}, **self.version_kwargs)

        with setup_and_verify_store() as store:
            with pytest.raises(
                ValueError,
                match=r"all variables in the dataset to write must have at least one dimension in common",
            ):
                data2.assign(v=2).to_zarr(
                    store, region={"x": slice(2)}, **self.version_kwargs
                )

        with setup_and_verify_store() as store:
            with pytest.raises(
                ValueError, match=r"cannot list the same dimension in both"
            ):
                data.to_zarr(
                    store,
                    region={"x": slice(None)},
                    append_dim="x",
                    **self.version_kwargs,
                )

        with setup_and_verify_store() as store:
            with pytest.raises(
                ValueError,
                match=r"variable 'u' already exists with different dimension sizes",
            ):
                data2.to_zarr(store, region={"x": slice(3)}, **self.version_kwargs)

    @requires_dask
    def test_encoding_chunksizes(self) -> None:
        # regression test for GH2278
        # see also test_encoding_chunksizes_unlimited
        nx, ny, nt = 4, 4, 5
        original = xr.Dataset(
            {}, coords={"x": np.arange(nx), "y": np.arange(ny), "t": np.arange(nt)}
        )
        original["v"] = xr.Variable(("x", "y", "t"), np.zeros((nx, ny, nt)))
        original = original.chunk({"t": 1, "x": 2, "y": 2})

        with self.roundtrip(original) as ds1:
            assert_equal(ds1, original)
            with self.roundtrip(ds1.isel(t=0)) as ds2:
                assert_equal(ds2, original.isel(t=0))

    @requires_dask
    def test_chunk_encoding_with_partial_dask_chunks(self) -> None:
        original = xr.Dataset(
            {"x": xr.DataArray(np.random.random(size=(6, 8)), dims=("a", "b"))}
        ).chunk({"a": 3})

        with self.roundtrip(
            original, save_kwargs={"encoding": {"x": {"chunks": [3, 2]}}}
        ) as ds1:
            assert_equal(ds1, original)

    @requires_dask
    def test_chunk_encoding_with_larger_dask_chunks(self) -> None:
        original = xr.Dataset({"a": ("x", [1, 2, 3, 4])}).chunk({"x": 2})

        with self.roundtrip(
            original, save_kwargs={"encoding": {"a": {"chunks": [1]}}}
        ) as ds1:
            assert_equal(ds1, original)

    @requires_cftime
    def test_open_zarr_use_cftime(self) -> None:
        ds = create_test_data()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, **self.version_kwargs)
            ds_a = xr.open_zarr(store_target, **self.version_kwargs)
            assert_identical(ds, ds_a)
            ds_b = xr.open_zarr(store_target, use_cftime=True, **self.version_kwargs)
            assert xr.coding.times.contains_cftime_datetimes(ds_b.time.variable)

    def test_write_read_select_write(self) -> None:
        # Test for https://github.com/pydata/xarray/issues/4084
        ds = create_test_data()

        # NOTE: using self.roundtrip, which uses open_dataset, will not trigger the bug.
        with self.create_zarr_target() as initial_store:
            ds.to_zarr(initial_store, mode="w", **self.version_kwargs)
            ds1 = xr.open_zarr(initial_store, **self.version_kwargs)

            # Combination of where+squeeze triggers error on write.
            ds_sel = ds1.where(ds1.coords["dim3"] == "a", drop=True).squeeze("dim3")
            with self.create_zarr_target() as final_store:
                ds_sel.to_zarr(final_store, mode="w", **self.version_kwargs)

    @pytest.mark.parametrize("obj", [Dataset(), DataArray(name="foo")])
    def test_attributes(self, obj) -> None:
        obj = obj.copy()

        obj.attrs["good"] = {"key": "value"}
        ds = obj if isinstance(obj, Dataset) else obj.to_dataset()
        with self.create_zarr_target() as store_target:
            ds.to_zarr(store_target, **self.version_kwargs)
            assert_identical(ds, xr.open_zarr(store_target, **self.version_kwargs))

        obj.attrs["bad"] = DataArray()
        ds = obj if isinstance(obj, Dataset) else obj.to_dataset()
        with self.create_zarr_target() as store_target:
            with pytest.raises(TypeError, match=r"Invalid attribute in Dataset.attrs."):
                ds.to_zarr(store_target, **self.version_kwargs)

    @requires_dask
    @pytest.mark.parametrize("dtype", ["datetime64[ns]", "timedelta64[ns]"])
    def test_chunked_datetime64_or_timedelta64(self, dtype) -> None:
        # Generalized from @malmans2's test in PR #8253
        original = create_test_data().astype(dtype).chunk(1)
        with self.roundtrip(original, open_kwargs={"chunks": {}}) as actual:
            for name, actual_var in actual.variables.items():
                assert original[name].chunks == actual_var.chunks
            assert original.chunks == actual.chunks

    @requires_cftime
    @requires_dask
    def test_chunked_cftime_datetime(self) -> None:
        # Based on @malmans2's test in PR #8253
        times = cftime_range("2000", freq="D", periods=3)
        original = xr.Dataset(data_vars={"chunked_times": (["time"], times)})
        original = original.chunk({"time": 1})
        with self.roundtrip(original, open_kwargs={"chunks": {}}) as actual:
            for name, actual_var in actual.variables.items():
                assert original[name].chunks == actual_var.chunks
            assert original.chunks == actual.chunks

    def test_vectorized_indexing_negative_step(self) -> None:
        if not has_dask:
            pytest.xfail(
                reason="zarr without dask handles negative steps in slices incorrectly"
            )

        super().test_vectorized_indexing_negative_step()


@requires_zarr
class TestZarrDictStore(ZarrBase):
    @contextlib.contextmanager
    def create_zarr_target(self):
        if have_zarr_kvstore:
            yield KVStore({})
        else:
            yield {}


@requires_zarr
@pytest.mark.skipif(
    ON_WINDOWS,
    reason="Very flaky on Windows CI. Can re-enable assuming it starts consistently passing.",
)
class TestZarrDirectoryStore(ZarrBase):
    @contextlib.contextmanager
    def create_zarr_target(self):
        with create_tmp_file(suffix=".zarr") as tmp:
            yield tmp

    @contextlib.contextmanager
    def create_store(self):
        with self.create_zarr_target() as store_target:
            group = backends.ZarrStore.open_group(store_target, mode="w")
            # older Zarr versions do not have the _store_version attribute
            if have_zarr_v3:
                # verify that a v2 store was created
                assert group.zarr_group.store._store_version == 2
            yield group


@requires_zarr
class TestZarrWriteEmpty(TestZarrDirectoryStore):
    @contextlib.contextmanager
    def temp_dir(self) -> Iterator[tuple[str, str]]:
        with tempfile.TemporaryDirectory() as d:
            store = os.path.join(d, "test.zarr")
            yield d, store

    @contextlib.contextmanager
    def roundtrip_dir(
        self,
        data,
        store,
        save_kwargs=None,
        open_kwargs=None,
        allow_cleanup_failure=False,
    ) -> Iterator[Dataset]:
        if save_kwargs is None:
            save_kwargs = {}
        if open_kwargs is None:
            open_kwargs = {}

        data.to_zarr(store, **save_kwargs, **self.version_kwargs)
        with xr.open_dataset(
            store, engine="zarr", **open_kwargs, **self.version_kwargs
        ) as ds:
            yield ds

    @pytest.mark.parametrize("consolidated", [True, False, None])
    @pytest.mark.parametrize("write_empty", [True, False, None])
    def test_write_empty(
        self, consolidated: bool | None, write_empty: bool | None
    ) -> None:
        if write_empty is False:
            expected = ["0.1.0", "1.1.0"]
        else:
            expected = [
                "0.0.0",
                "0.0.1",
                "0.1.0",
                "0.1.1",
                "1.0.0",
                "1.0.1",
                "1.1.0",
                "1.1.1",
            ]

        ds = xr.Dataset(
            data_vars={
                "test": (
                    ("Z", "Y", "X"),
                    np.array([np.nan, np.nan, 1.0, np.nan]).reshape((1, 2, 2)),
                )
            }
        )

        if has_dask:
            ds["test"] = ds["test"].chunk(1)
            encoding = None
        else:
            encoding = {"test": {"chunks": (1, 1, 1)}}

        with self.temp_dir() as (d, store):
            ds.to_zarr(
                store,
                mode="w",
                encoding=encoding,
                write_empty_chunks=write_empty,
            )

            with self.roundtrip_dir(
                ds,
                store,
                {"mode": "a", "append_dim": "Z", "write_empty_chunks": write_empty},
            ) as a_ds:
                expected_ds = xr.concat([ds, ds], dim="Z")

                assert_identical(a_ds, expected_ds)

                ls = listdir(os.path.join(store, "test"))
                assert set(expected) == set([file for file in ls if file[0] != "."])

    def test_avoid_excess_metadata_calls(self) -> None:
        """Test that chunk requests do not trigger redundant metadata requests.

        This test targets logic in backends.zarr.ZarrArrayWrapper, asserting that calls
        to retrieve chunk data after initialization do not trigger additional
        metadata requests.

        https://github.com/pydata/xarray/issues/8290
        """

        import zarr

        ds = xr.Dataset(data_vars={"test": (("Z",), np.array([123]).reshape(1))})

        # The call to retrieve metadata performs a group lookup. We patch Group.__getitem__
        # so that we can inspect calls to this method - specifically count of calls.
        # Use of side_effect means that calls are passed through to the original method
        # rather than a mocked method.
        Group = zarr.hierarchy.Group
        with (
            self.create_zarr_target() as store,
            patch.object(
                Group, "__getitem__", side_effect=Group.__getitem__, autospec=True
            ) as mock,
        ):
            ds.to_zarr(store, mode="w")

            # We expect this to request array metadata information, so call_count should be == 1,
            xrds = xr.open_zarr(store)
            call_count = mock.call_count
            assert call_count == 1

            # compute() requests array data, which should not trigger additional metadata requests
            # we assert that the number of calls has not increased after fetchhing the array
            xrds.test.compute(scheduler="sync")
            assert mock.call_count == call_count


class ZarrBaseV3(ZarrBase):
    zarr_version = 3

    def test_roundtrip_coordinates_with_space(self):
        original = Dataset(coords={"x": 0, "y z": 1})
        with pytest.warns(SerializationWarning):
            # v3 stores do not allow spaces in the key name
            with pytest.raises(ValueError):
                with self.roundtrip(original):
                    pass


@pytest.mark.skipif(not have_zarr_v3, reason="requires zarr version 3")
class TestZarrKVStoreV3(ZarrBaseV3):
    @contextlib.contextmanager
    def create_zarr_target(self):
        yield KVStoreV3({})


@pytest.mark.skipif(not have_zarr_v3, reason="requires zarr version 3")
class TestZarrDirectoryStoreV3(ZarrBaseV3):
    @contextlib.contextmanager
    def create_zarr_target(self):
        with create_tmp_file(suffix=".zr3") as tmp:
            yield DirectoryStoreV3(tmp)


@pytest.mark.skipif(not have_zarr_v3, reason="requires zarr version 3")
class TestZarrDirectoryStoreV3FromPath(TestZarrDirectoryStoreV3):
    # Must specify zarr_version=3 to get a v3 store because create_zarr_target
    # is a string path.
    version_kwargs = {"zarr_version": 3}

    @contextlib.contextmanager
    def create_zarr_target(self):
        with create_tmp_file(suffix=".zr3") as tmp:
            yield tmp


@requires_zarr
@requires_fsspec
def test_zarr_storage_options() -> None:
    pytest.importorskip("aiobotocore")
    ds = create_test_data()
    store_target = "memory://test.zarr"
    ds.to_zarr(store_target, storage_options={"test": "zarr_write"})
    ds_a = xr.open_zarr(store_target, storage_options={"test": "zarr_read"})
    assert_identical(ds, ds_a)


@requires_scipy
class TestScipyInMemoryData(CFEncodedBase, NetCDF3Only):
    engine: T_NetcdfEngine = "scipy"

    @contextlib.contextmanager
    def create_store(self):
        fobj = BytesIO()
        yield backends.ScipyDataStore(fobj, "w")

    def test_to_netcdf_explicit_engine(self) -> None:
        # regression test for GH1321
        Dataset({"foo": 42}).to_netcdf(engine="scipy")

    def test_bytes_pickle(self) -> None:
        data = Dataset({"foo": ("x", [1, 2, 3])})
        fobj = data.to_netcdf()
        with self.open(fobj) as ds:
            unpickled = pickle.loads(pickle.dumps(ds))
            assert_identical(unpickled, data)


@requires_scipy
class TestScipyFileObject(CFEncodedBase, NetCDF3Only):
    engine: T_NetcdfEngine = "scipy"

    @contextlib.contextmanager
    def create_store(self):
        fobj = BytesIO()
        yield backends.ScipyDataStore(fobj, "w")

    @contextlib.contextmanager
    def roundtrip(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        if save_kwargs is None:
            save_kwargs = {}
        if open_kwargs is None:
            open_kwargs = {}
        with create_tmp_file() as tmp_file:
            with open(tmp_file, "wb") as f:
                self.save(data, f, **save_kwargs)
            with open(tmp_file, "rb") as f:
                with self.open(f, **open_kwargs) as ds:
                    yield ds

    @pytest.mark.skip(reason="cannot pickle file objects")
    def test_pickle(self) -> None:
        pass

    @pytest.mark.skip(reason="cannot pickle file objects")
    def test_pickle_dataarray(self) -> None:
        pass


@requires_scipy
class TestScipyFilePath(CFEncodedBase, NetCDF3Only):
    engine: T_NetcdfEngine = "scipy"

    @contextlib.contextmanager
    def create_store(self):
        with create_tmp_file() as tmp_file:
            with backends.ScipyDataStore(tmp_file, mode="w") as store:
                yield store

    def test_array_attrs(self) -> None:
        ds = Dataset(attrs={"foo": [[1, 2], [3, 4]]})
        with pytest.raises(ValueError, match=r"must be 1-dimensional"):
            with self.roundtrip(ds):
                pass

    def test_roundtrip_example_1_netcdf_gz(self) -> None:
        with open_example_dataset("example_1.nc.gz") as expected:
            with open_example_dataset("example_1.nc") as actual:
                assert_identical(expected, actual)

    def test_netcdf3_endianness(self) -> None:
        # regression test for GH416
        with open_example_dataset("bears.nc", engine="scipy") as expected:
            for var in expected.variables.values():
                assert var.dtype.isnative

    @requires_netCDF4
    def test_nc4_scipy(self) -> None:
        with create_tmp_file(allow_cleanup_failure=True) as tmp_file:
            with nc4.Dataset(tmp_file, "w", format="NETCDF4") as rootgrp:
                rootgrp.createGroup("foo")

            with pytest.raises(TypeError, match=r"pip install netcdf4"):
                open_dataset(tmp_file, engine="scipy")


@requires_netCDF4
class TestNetCDF3ViaNetCDF4Data(CFEncodedBase, NetCDF3Only):
    engine: T_NetcdfEngine = "netcdf4"
    file_format: T_NetcdfTypes = "NETCDF3_CLASSIC"

    @contextlib.contextmanager
    def create_store(self):
        with create_tmp_file() as tmp_file:
            with backends.NetCDF4DataStore.open(
                tmp_file, mode="w", format="NETCDF3_CLASSIC"
            ) as store:
                yield store

    def test_encoding_kwarg_vlen_string(self) -> None:
        original = Dataset({"x": ["foo", "bar", "baz"]})
        kwargs = dict(encoding={"x": {"dtype": str}})
        with pytest.raises(ValueError, match=r"encoding dtype=str for vlen"):
            with self.roundtrip(original, save_kwargs=kwargs):
                pass


@requires_netCDF4
class TestNetCDF4ClassicViaNetCDF4Data(CFEncodedBase, NetCDF3Only):
    engine: T_NetcdfEngine = "netcdf4"
    file_format: T_NetcdfTypes = "NETCDF4_CLASSIC"

    @contextlib.contextmanager
    def create_store(self):
        with create_tmp_file() as tmp_file:
            with backends.NetCDF4DataStore.open(
                tmp_file, mode="w", format="NETCDF4_CLASSIC"
            ) as store:
                yield store


@requires_scipy_or_netCDF4
class TestGenericNetCDFData(CFEncodedBase, NetCDF3Only):
    # verify that we can read and write netCDF3 files as long as we have scipy
    # or netCDF4-python installed
    file_format: T_NetcdfTypes = "NETCDF3_64BIT"

    def test_write_store(self) -> None:
        # there's no specific store to test here
        pass

    @requires_scipy
    def test_engine(self) -> None:
        data = create_test_data()
        with pytest.raises(ValueError, match=r"unrecognized engine"):
            data.to_netcdf("foo.nc", engine="foobar")  # type: ignore[call-overload]
        with pytest.raises(ValueError, match=r"invalid engine"):
            data.to_netcdf(engine="netcdf4")

        with create_tmp_file() as tmp_file:
            data.to_netcdf(tmp_file)
            with pytest.raises(ValueError, match=r"unrecognized engine"):
                open_dataset(tmp_file, engine="foobar")

        netcdf_bytes = data.to_netcdf()
        with pytest.raises(ValueError, match=r"unrecognized engine"):
            open_dataset(BytesIO(netcdf_bytes), engine="foobar")

    def test_cross_engine_read_write_netcdf3(self) -> None:
        data = create_test_data()
        valid_engines: set[T_NetcdfEngine] = set()
        if has_netCDF4:
            valid_engines.add("netcdf4")
        if has_scipy:
            valid_engines.add("scipy")

        for write_engine in valid_engines:
            for format in self.netcdf3_formats:
                with create_tmp_file() as tmp_file:
                    data.to_netcdf(tmp_file, format=format, engine=write_engine)
                    for read_engine in valid_engines:
                        with open_dataset(tmp_file, engine=read_engine) as actual:
                            # hack to allow test to work:
                            # coord comes back as DataArray rather than coord,
                            # and so need to loop through here rather than in
                            # the test function (or we get recursion)
                            [
                                assert_allclose(data[k].variable, actual[k].variable)
                                for k in data.variables
                            ]

    def test_encoding_unlimited_dims(self) -> None:
        ds = Dataset({"x": ("y", np.arange(10.0))})
        with self.roundtrip(ds, save_kwargs=dict(unlimited_dims=["y"])) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)

        # Regression test for https://github.com/pydata/xarray/issues/2134
        with self.roundtrip(ds, save_kwargs=dict(unlimited_dims="y")) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)

        ds.encoding = {"unlimited_dims": ["y"]}
        with self.roundtrip(ds) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)

        # Regression test for https://github.com/pydata/xarray/issues/2134
        ds.encoding = {"unlimited_dims": "y"}
        with self.roundtrip(ds) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)


@requires_h5netcdf
@requires_netCDF4
@pytest.mark.filterwarnings("ignore:use make_scale(name) instead")
class TestH5NetCDFData(NetCDF4Base):
    engine: T_NetcdfEngine = "h5netcdf"

    @contextlib.contextmanager
    def create_store(self):
        with create_tmp_file() as tmp_file:
            yield backends.H5NetCDFStore.open(tmp_file, "w")

    def test_complex(self) -> None:
        expected = Dataset({"x": ("y", np.ones(5) + 1j * np.ones(5))})
        save_kwargs = {"invalid_netcdf": True}
        with pytest.warns(UserWarning, match="You are writing invalid netcdf features"):
            with self.roundtrip(expected, save_kwargs=save_kwargs) as actual:
                assert_equal(expected, actual)

    @pytest.mark.parametrize("invalid_netcdf", [None, False])
    def test_complex_error(self, invalid_netcdf) -> None:
        import h5netcdf

        expected = Dataset({"x": ("y", np.ones(5) + 1j * np.ones(5))})
        save_kwargs = {"invalid_netcdf": invalid_netcdf}
        with pytest.raises(
            h5netcdf.CompatibilityError, match="are not a supported NetCDF feature"
        ):
            with self.roundtrip(expected, save_kwargs=save_kwargs) as actual:
                assert_equal(expected, actual)

    def test_numpy_bool_(self) -> None:
        # h5netcdf loads booleans as numpy.bool_, this type needs to be supported
        # when writing invalid_netcdf datasets in order to support a roundtrip
        expected = Dataset({"x": ("y", np.ones(5), {"numpy_bool": np.bool_(True)})})
        save_kwargs = {"invalid_netcdf": True}
        with pytest.warns(UserWarning, match="You are writing invalid netcdf features"):
            with self.roundtrip(expected, save_kwargs=save_kwargs) as actual:
                assert_identical(expected, actual)

    def test_cross_engine_read_write_netcdf4(self) -> None:
        # Drop dim3, because its labels include strings. These appear to be
        # not properly read with python-netCDF4, which converts them into
        # unicode instead of leaving them as bytes.
        data = create_test_data().drop_vars("dim3")
        data.attrs["foo"] = "bar"
        valid_engines: list[T_NetcdfEngine] = ["netcdf4", "h5netcdf"]
        for write_engine in valid_engines:
            with create_tmp_file() as tmp_file:
                data.to_netcdf(tmp_file, engine=write_engine)
                for read_engine in valid_engines:
                    with open_dataset(tmp_file, engine=read_engine) as actual:
                        assert_identical(data, actual)

    def test_read_byte_attrs_as_unicode(self) -> None:
        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, "w") as nc:
                nc.foo = b"bar"
            with open_dataset(tmp_file) as actual:
                expected = Dataset(attrs={"foo": "bar"})
                assert_identical(expected, actual)

    def test_encoding_unlimited_dims(self) -> None:
        ds = Dataset({"x": ("y", np.arange(10.0))})
        with self.roundtrip(ds, save_kwargs=dict(unlimited_dims=["y"])) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)
        ds.encoding = {"unlimited_dims": ["y"]}
        with self.roundtrip(ds) as actual:
            assert actual.encoding["unlimited_dims"] == set("y")
            assert_equal(ds, actual)

    def test_compression_encoding_h5py(self) -> None:
        ENCODINGS: tuple[tuple[dict[str, Any], dict[str, Any]], ...] = (
            # h5py style compression with gzip codec will be converted to
            # NetCDF4-Python style on round-trip
            (
                {"compression": "gzip", "compression_opts": 9},
                {"zlib": True, "complevel": 9},
            ),
            # What can't be expressed in NetCDF4-Python style is
            # round-tripped unaltered
            (
                {"compression": "lzf", "compression_opts": None},
                {"compression": "lzf", "compression_opts": None},
            ),
            # If both styles are used together, h5py format takes precedence
            (
                {
                    "compression": "lzf",
                    "compression_opts": None,
                    "zlib": True,
                    "complevel": 9,
                },
                {"compression": "lzf", "compression_opts": None},
            ),
        )

        for compr_in, compr_out in ENCODINGS:
            data = create_test_data()
            compr_common = {
                "chunksizes": (5, 5),
                "fletcher32": True,
                "shuffle": True,
                "original_shape": data.var2.shape,
            }
            data["var2"].encoding.update(compr_in)
            data["var2"].encoding.update(compr_common)
            compr_out.update(compr_common)
            data["scalar"] = ("scalar_dim", np.array([2.0]))
            data["scalar"] = data["scalar"][0]
            with self.roundtrip(data) as actual:
                for k, v in compr_out.items():
                    assert v == actual["var2"].encoding[k]

    def test_compression_check_encoding_h5py(self) -> None:
        """When mismatched h5py and NetCDF4-Python encodings are expressed
        in to_netcdf(encoding=...), must raise ValueError
        """
        data = Dataset({"x": ("y", np.arange(10.0))})
        # Compatible encodings are graciously supported
        with create_tmp_file() as tmp_file:
            data.to_netcdf(
                tmp_file,
                engine="h5netcdf",
                encoding={
                    "x": {
                        "compression": "gzip",
                        "zlib": True,
                        "compression_opts": 6,
                        "complevel": 6,
                    }
                },
            )
            with open_dataset(tmp_file, engine="h5netcdf") as actual:
                assert actual.x.encoding["zlib"] is True
                assert actual.x.encoding["complevel"] == 6

        # Incompatible encodings cause a crash
        with create_tmp_file() as tmp_file:
            with pytest.raises(
                ValueError, match=r"'zlib' and 'compression' encodings mismatch"
            ):
                data.to_netcdf(
                    tmp_file,
                    engine="h5netcdf",
                    encoding={"x": {"compression": "lzf", "zlib": True}},
                )

        with create_tmp_file() as tmp_file:
            with pytest.raises(
                ValueError,
                match=r"'complevel' and 'compression_opts' encodings mismatch",
            ):
                data.to_netcdf(
                    tmp_file,
                    engine="h5netcdf",
                    encoding={
                        "x": {
                            "compression": "gzip",
                            "compression_opts": 5,
                            "complevel": 6,
                        }
                    },
                )

    def test_dump_encodings_h5py(self) -> None:
        # regression test for #709
        ds = Dataset({"x": ("y", np.arange(10.0))})

        kwargs = {"encoding": {"x": {"compression": "gzip", "compression_opts": 9}}}
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert actual.x.encoding["zlib"]
            assert actual.x.encoding["complevel"] == 9

        kwargs = {"encoding": {"x": {"compression": "lzf", "compression_opts": None}}}
        with self.roundtrip(ds, save_kwargs=kwargs) as actual:
            assert actual.x.encoding["compression"] == "lzf"
            assert actual.x.encoding["compression_opts"] is None


@requires_h5netcdf
@requires_netCDF4
class TestH5NetCDFAlreadyOpen:
    def test_open_dataset_group(self) -> None:
        import h5netcdf

        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                group = nc.createGroup("g")
                v = group.createVariable("x", "int")
                v[...] = 42

            kwargs = {"decode_vlen_strings": True}

            h5 = h5netcdf.File(tmp_file, mode="r", **kwargs)
            store = backends.H5NetCDFStore(h5["g"])
            with open_dataset(store) as ds:
                expected = Dataset({"x": ((), 42)})
                assert_identical(expected, ds)

            h5 = h5netcdf.File(tmp_file, mode="r", **kwargs)
            store = backends.H5NetCDFStore(h5, group="g")
            with open_dataset(store) as ds:
                expected = Dataset({"x": ((), 42)})
                assert_identical(expected, ds)

    def test_deepcopy(self) -> None:
        import h5netcdf

        with create_tmp_file() as tmp_file:
            with nc4.Dataset(tmp_file, mode="w") as nc:
                nc.createDimension("x", 10)
                v = nc.createVariable("y", np.int32, ("x",))
                v[:] = np.arange(10)

            kwargs = {"decode_vlen_strings": True}

            h5 = h5netcdf.File(tmp_file, mode="r", **kwargs)
            store = backends.H5NetCDFStore(h5)
            with open_dataset(store) as ds:
                copied = ds.copy(deep=True)
                expected = Dataset({"y": ("x", np.arange(10))})
                assert_identical(expected, copied)


@requires_h5netcdf
class TestH5NetCDFFileObject(TestH5NetCDFData):
    engine: T_NetcdfEngine = "h5netcdf"

    def test_open_badbytes(self) -> None:
        with pytest.raises(ValueError, match=r"HDF5 as bytes"):
            with open_dataset(b"\211HDF\r\n\032\n", engine="h5netcdf"):  # type: ignore[arg-type]
                pass
        with pytest.raises(
            ValueError, match=r"match in any of xarray's currently installed IO"
        ):
            with open_dataset(b"garbage"):  # type: ignore[arg-type]
                pass
        with pytest.raises(ValueError, match=r"can only read bytes"):
            with open_dataset(b"garbage", engine="netcdf4"):  # type: ignore[arg-type]
                pass
        with pytest.raises(
            ValueError, match=r"not the signature of a valid netCDF4 file"
        ):
            with open_dataset(BytesIO(b"garbage"), engine="h5netcdf"):
                pass

    def test_open_twice(self) -> None:
        expected = create_test_data()
        expected.attrs["foo"] = "bar"
        with create_tmp_file() as tmp_file:
            expected.to_netcdf(tmp_file, engine="h5netcdf")
            with open(tmp_file, "rb") as f:
                with open_dataset(f, engine="h5netcdf"):
                    with open_dataset(f, engine="h5netcdf"):
                        pass

    @requires_scipy
    def test_open_fileobj(self) -> None:
        # open in-memory datasets instead of local file paths
        expected = create_test_data().drop_vars("dim3")
        expected.attrs["foo"] = "bar"
        with create_tmp_file() as tmp_file:
            expected.to_netcdf(tmp_file, engine="h5netcdf")

            with open(tmp_file, "rb") as f:
                with open_dataset(f, engine="h5netcdf") as actual:
                    assert_identical(expected, actual)

                f.seek(0)
                with open_dataset(f) as actual:
                    assert_identical(expected, actual)

                f.seek(0)
                with BytesIO(f.read()) as bio:
                    with open_dataset(bio, engine="h5netcdf") as actual:
                        assert_identical(expected, actual)

                f.seek(0)
                with pytest.raises(TypeError, match="not a valid NetCDF 3"):
                    open_dataset(f, engine="scipy")

            # TODO: this additional open is required since scipy seems to close the file
            # when it fails on the TypeError (though didn't when we used
            # `raises_regex`?). Ref https://github.com/pydata/xarray/pull/5191
            with open(tmp_file, "rb") as f:
                f.seek(8)
                open_dataset(f)


@requires_h5netcdf
@requires_dask
@pytest.mark.filterwarnings("ignore:deallocating CachingFileManager")
class TestH5NetCDFViaDaskData(TestH5NetCDFData):
    @contextlib.contextmanager
    def roundtrip(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        if save_kwargs is None:
            save_kwargs = {}
        if open_kwargs is None:
            open_kwargs = {}
        open_kwargs.setdefault("chunks", -1)
        with TestH5NetCDFData.roundtrip(
            self, data, save_kwargs, open_kwargs, allow_cleanup_failure
        ) as ds:
            yield ds

    @pytest.mark.skip(reason="caching behavior differs for dask")
    def test_dataset_caching(self) -> None:
        pass

    def test_write_inconsistent_chunks(self) -> None:
        # Construct two variables with the same dimensions, but different
        # chunk sizes.
        x = da.zeros((100, 100), dtype="f4", chunks=(50, 100))
        x = DataArray(data=x, dims=("lat", "lon"), name="x")
        x.encoding["chunksizes"] = (50, 100)
        x.encoding["original_shape"] = (100, 100)
        y = da.ones((100, 100), dtype="f4", chunks=(100, 50))
        y = DataArray(data=y, dims=("lat", "lon"), name="y")
        y.encoding["chunksizes"] = (100, 50)
        y.encoding["original_shape"] = (100, 100)
        # Put them both into the same dataset
        ds = Dataset({"x": x, "y": y})
        with self.roundtrip(ds) as actual:
            assert actual["x"].encoding["chunksizes"] == (50, 100)
            assert actual["y"].encoding["chunksizes"] == (100, 50)


@requires_h5netcdf_ros3
class TestH5NetCDFDataRos3Driver(TestCommon):
    engine: T_NetcdfEngine = "h5netcdf"
    test_remote_dataset: str = (
        "https://www.unidata.ucar.edu/software/netcdf/examples/OMI-Aura_L2-example.nc"
    )

    @pytest.mark.filterwarnings("ignore:Duplicate dimension names")
    def test_get_variable_list(self) -> None:
        with open_dataset(
            self.test_remote_dataset,
            engine="h5netcdf",
            backend_kwargs={"driver": "ros3"},
        ) as actual:
            assert "Temperature" in list(actual)

    @pytest.mark.filterwarnings("ignore:Duplicate dimension names")
    def test_get_variable_list_empty_driver_kwds(self) -> None:
        driver_kwds = {
            "secret_id": b"",
            "secret_key": b"",
        }
        backend_kwargs = {"driver": "ros3", "driver_kwds": driver_kwds}

        with open_dataset(
            self.test_remote_dataset, engine="h5netcdf", backend_kwargs=backend_kwargs
        ) as actual:
            assert "Temperature" in list(actual)


@pytest.fixture(params=["scipy", "netcdf4", "h5netcdf", "pynio", "zarr"])
def readengine(request):
    return request.param


@pytest.fixture(params=[1, 20])
def nfiles(request):
    return request.param


@pytest.fixture(params=[5, None])
def file_cache_maxsize(request):
    maxsize = request.param
    if maxsize is not None:
        with set_options(file_cache_maxsize=maxsize):
            yield maxsize
    else:
        yield maxsize


@pytest.fixture(params=[True, False])
def parallel(request):
    return request.param


@pytest.fixture(params=[None, 5])
def chunks(request):
    return request.param


@pytest.fixture(params=["tmp_path", "ZipStore", "Dict"])
def tmp_store(request, tmp_path):
    if request.param == "tmp_path":
        return tmp_path
    elif request.param == "ZipStore":
        from zarr.storage import ZipStore

        path = tmp_path / "store.zip"
        return ZipStore(path)
    elif request.param == "Dict":
        return dict()
    else:
        raise ValueError("not supported")


# using pytest.mark.skipif does not work so this a work around
def skip_if_not_engine(engine):
    if engine == "netcdf4":
        pytest.importorskip("netCDF4")
    elif engine == "pynio":
        pytest.importorskip("Nio")
    else:
        pytest.importorskip(engine)


@requires_dask
@pytest.mark.filterwarnings("ignore:use make_scale(name) instead")
@pytest.mark.xfail(reason="Flaky test. Very open to contributions on fixing this")
def test_open_mfdataset_manyfiles(
    readengine, nfiles, parallel, chunks, file_cache_maxsize
):
    # skip certain combinations
    skip_if_not_engine(readengine)

    if ON_WINDOWS:
        pytest.skip("Skipping on Windows")

    randdata = np.random.randn(nfiles)
    original = Dataset({"foo": ("x", randdata)})
    # test standard open_mfdataset approach with too many files
    with create_tmp_files(nfiles) as tmpfiles:
        writeengine = readengine if readengine != "pynio" else "netcdf4"
        # split into multiple sets of temp files
        for ii in original.x.values:
            subds = original.isel(x=slice(ii, ii + 1))
            if writeengine != "zarr":
                subds.to_netcdf(tmpfiles[ii], engine=writeengine)
            else:  # if writeengine == "zarr":
                subds.to_zarr(store=tmpfiles[ii])

        # check that calculation on opened datasets works properly
        with open_mfdataset(
            tmpfiles,
            combine="nested",
            concat_dim="x",
            engine=readengine,
            parallel=parallel,
            chunks=chunks if (not chunks and readengine != "zarr") else "auto",
        ) as actual:
            # check that using open_mfdataset returns dask arrays for variables
            assert isinstance(actual["foo"].data, dask_array_type)

            assert_identical(original, actual)


@requires_netCDF4
@requires_dask
def test_open_mfdataset_can_open_path_objects() -> None:
    dataset = os.path.join(os.path.dirname(__file__), "data", "example_1.nc")
    with open_mfdataset(Path(dataset)) as actual:
        assert isinstance(actual, Dataset)


@requires_netCDF4
@requires_dask
def test_open_mfdataset_list_attr() -> None:
    """
    Case when an attribute of type list differs across the multiple files
    """
    from netCDF4 import Dataset

    with create_tmp_files(2) as nfiles:
        for i in range(2):
            with Dataset(nfiles[i], "w") as f:
                f.createDimension("x", 3)
                vlvar = f.createVariable("test_var", np.int32, ("x"))
                # here create an attribute as a list
                vlvar.test_attr = [f"string a {i}", f"string b {i}"]
                vlvar[:] = np.arange(3)

        with open_dataset(nfiles[0]) as ds1:
            with open_dataset(nfiles[1]) as ds2:
                original = xr.concat([ds1, ds2], dim="x")
                with xr.open_mfdataset(
                    [nfiles[0], nfiles[1]], combine="nested", concat_dim="x"
                ) as actual:
                    assert_identical(actual, original)


@requires_scipy_or_netCDF4
@requires_dask
class TestOpenMFDatasetWithDataVarsAndCoordsKw:
    coord_name = "lon"
    var_name = "v1"

    @contextlib.contextmanager
    def setup_files_and_datasets(self, fuzz=0):
        ds1, ds2 = self.gen_datasets_with_common_coord_and_time()

        # to test join='exact'
        ds1["x"] = ds1.x + fuzz

        with create_tmp_file() as tmpfile1:
            with create_tmp_file() as tmpfile2:
                # save data to the temporary files
                ds1.to_netcdf(tmpfile1)
                ds2.to_netcdf(tmpfile2)

                yield [tmpfile1, tmpfile2], [ds1, ds2]

    def gen_datasets_with_common_coord_and_time(self):
        # create coordinate data
        nx = 10
        nt = 10
        x = np.arange(nx)
        t1 = np.arange(nt)
        t2 = np.arange(nt, 2 * nt, 1)

        v1 = np.random.randn(nt, nx)
        v2 = np.random.randn(nt, nx)

        ds1 = Dataset(
            data_vars={self.var_name: (["t", "x"], v1), self.coord_name: ("x", 2 * x)},
            coords={"t": (["t"], t1), "x": (["x"], x)},
        )

        ds2 = Dataset(
            data_vars={self.var_name: (["t", "x"], v2), self.coord_name: ("x", 2 * x)},
            coords={"t": (["t"], t2), "x": (["x"], x)},
        )

        return ds1, ds2

    @pytest.mark.parametrize(
        "combine, concat_dim", [("nested", "t"), ("by_coords", None)]
    )
    @pytest.mark.parametrize("opt", ["all", "minimal", "different"])
    @pytest.mark.parametrize("join", ["outer", "inner", "left", "right"])
    def test_open_mfdataset_does_same_as_concat(
        self, combine, concat_dim, opt, join
    ) -> None:
        with self.setup_files_and_datasets() as (files, [ds1, ds2]):
            if combine == "by_coords":
                files.reverse()
            with open_mfdataset(
                files, data_vars=opt, combine=combine, concat_dim=concat_dim, join=join
            ) as ds:
                ds_expect = xr.concat([ds1, ds2], data_vars=opt, dim="t", join=join)
                assert_identical(ds, ds_expect)

    @pytest.mark.parametrize(
        ["combine_attrs", "attrs", "expected", "expect_error"],
        (
            pytest.param("drop", [{"a": 1}, {"a": 2}], {}, False, id="drop"),
            pytest.param(
                "override", [{"a": 1}, {"a": 2}], {"a": 1}, False, id="override"
            ),
            pytest.param(
                "no_conflicts", [{"a": 1}, {"a": 2}], None, True, id="no_conflicts"
            ),
            pytest.param(
                "identical",
                [{"a": 1, "b": 2}, {"a": 1, "c": 3}],
                None,
                True,
                id="identical",
            ),
            pytest.param(
                "drop_conflicts",
                [{"a": 1, "b": 2}, {"b": -1, "c": 3}],
                {"a": 1, "c": 3},
                False,
                id="drop_conflicts",
            ),
        ),
    )
    def test_open_mfdataset_dataset_combine_attrs(
        self, combine_attrs, attrs, expected, expect_error
    ):
        with self.setup_files_and_datasets() as (files, [ds1, ds2]):
            # Give the files an inconsistent attribute
            for i, f in enumerate(files):
                ds = open_dataset(f).load()
                ds.attrs = attrs[i]
                ds.close()
                ds.to_netcdf(f)

            if expect_error:
                with pytest.raises(xr.MergeError):
                    xr.open_mfdataset(
                        files,
                        combine="nested",
                        concat_dim="t",
                        combine_attrs=combine_attrs,
                    )
            else:
                with xr.open_mfdataset(
                    files,
                    combine="nested",
                    concat_dim="t",
                    combine_attrs=combine_attrs,
                ) as ds:
                    assert ds.attrs == expected

    def test_open_mfdataset_dataset_attr_by_coords(self) -> None:
        """
        Case when an attribute differs across the multiple files
        """
        with self.setup_files_and_datasets() as (files, [ds1, ds2]):
            # Give the files an inconsistent attribute
            for i, f in enumerate(files):
                ds = open_dataset(f).load()
                ds.attrs["test_dataset_attr"] = 10 + i
                ds.close()
                ds.to_netcdf(f)

            with xr.open_mfdataset(files, combine="nested", concat_dim="t") as ds:
                assert ds.test_dataset_attr == 10

    def test_open_mfdataset_dataarray_attr_by_coords(self) -> None:
        """
        Case when an attribute of a member DataArray differs across the multiple files
        """
        with self.setup_files_and_datasets() as (files, [ds1, ds2]):
            # Give the files an inconsistent attribute
            for i, f in enumerate(files):
                ds = open_dataset(f).load()
                ds["v1"].attrs["test_dataarray_attr"] = i
                ds.close()
                ds.to_netcdf(f)

            with xr.open_mfdataset(files, combine="nested", concat_dim="t") as ds:
                assert ds["v1"].test_dataarray_attr == 0

    @pytest.mark.parametrize(
        "combine, concat_dim", [("nested", "t"), ("by_coords", None)]
    )
    @pytest.mark.parametrize("opt", ["all", "minimal", "different"])
    def test_open_mfdataset_exact_join_raises_error(
        self, combine, concat_dim, opt
    ) -> None:
        with self.setup_files_and_datasets(fuzz=0.1) as (files, [ds1, ds2]):
            if combine == "by_coords":
                files.reverse()
            with pytest.raises(
                ValueError, match=r"cannot align objects.*join.*exact.*"
            ):
                open_mfdataset(
                    files,
                    data_vars=opt,
                    combine=combine,
                    concat_dim=concat_dim,
                    join="exact",
                )

    def test_common_coord_when_datavars_all(self) -> None:
        opt: Final = "all"

        with self.setup_files_and_datasets() as (files, [ds1, ds2]):
            # open the files with the data_var option
            with open_mfdataset(
                files, data_vars=opt, combine="nested", concat_dim="t"
            ) as ds:
                coord_shape = ds[self.coord_name].shape
                coord_shape1 = ds1[self.coord_name].shape
                coord_shape2 = ds2[self.coord_name].shape

                var_shape = ds[self.var_name].shape

                assert var_shape == coord_shape
                assert coord_shape1 != coord_shape
                assert coord_shape2 != coord_shape

    def test_common_coord_when_datavars_minimal(self) -> None:
        opt: Final = "minimal"

        with self.setup_files_and_datasets() as (files, [ds1, ds2]):
            # open the files using data_vars option
            with open_mfdataset(
                files, data_vars=opt, combine="nested", concat_dim="t"
            ) as ds:
                coord_shape = ds[self.coord_name].shape
                coord_shape1 = ds1[self.coord_name].shape
                coord_shape2 = ds2[self.coord_name].shape

                var_shape = ds[self.var_name].shape

                assert var_shape != coord_shape
                assert coord_shape1 == coord_shape
                assert coord_shape2 == coord_shape

    def test_invalid_data_vars_value_should_fail(self) -> None:
        with self.setup_files_and_datasets() as (files, _):
            with pytest.raises(ValueError):
                with open_mfdataset(files, data_vars="minimum", combine="by_coords"):  # type: ignore[arg-type]
                    pass

            # test invalid coord parameter
            with pytest.raises(ValueError):
                with open_mfdataset(files, coords="minimum", combine="by_coords"):
                    pass


@requires_dask
@requires_scipy
@requires_netCDF4
class TestDask(DatasetIOBase):
    @contextlib.contextmanager
    def create_store(self):
        yield Dataset()

    @contextlib.contextmanager
    def roundtrip(
        self, data, save_kwargs=None, open_kwargs=None, allow_cleanup_failure=False
    ):
        yield data.chunk()

    # Override methods in DatasetIOBase - not applicable to dask
    def test_roundtrip_string_encoded_characters(self) -> None:
        pass

    def test_roundtrip_coordinates_with_space(self) -> None:
        pass

    def test_roundtrip_numpy_datetime_data(self) -> None:
        # Override method in DatasetIOBase - remove not applicable
        # save_kwargs
        times = pd.to_datetime(["2000-01-01", "2000-01-02", "NaT"])
        expected = Dataset({"t": ("t", times), "t0": times[0]})
        with self.roundtrip(expected) as actual:
            assert_identical(expected, actual)

    def test_roundtrip_cftime_datetime_data(self) -> None:
        # Override method in DatasetIOBase - remove not applicable
        # save_kwargs
        from xarray.tests.test_coding_times import _all_cftime_date_types

        date_types = _all_cftime_date_types()
        for date_type in date_types.values():
            times = [date_type(1, 1, 1), date_type(1, 1, 2)]
            expected = Dataset({"t": ("t", times), "t0": times[0]})
            expected_decoded_t = np.array(times)
            expected_decoded_t0 = np.array([date_type(1, 1, 1)])

            with self.roundtrip(expected) as actual:
                abs_diff = abs(actual.t.values - expected_decoded_t)
                assert (abs_diff <= np.timedelta64(1, "s")).all()

                abs_diff = abs(actual.t0.values - expected_decoded_t0)
                assert (abs_diff <= np.timedelta64(1, "s")).all()

    def test_write_store(self) -> None:
        # Override method in DatasetIOBase - not applicable to dask
        pass

    def test_dataset_caching(self) -> None:
        expected = Dataset({"foo": ("x", [5, 6, 7])})
        with self.roundtrip(expected) as actual:
            assert not actual.foo.variable._in_memory
            actual.foo.values  # no caching
            assert not actual.foo.variable._in_memory

    def test_open_mfdataset(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                original.isel(x=slice(5)).to_netcdf(tmp1)
                original.isel(x=slice(5, 10)).to_netcdf(tmp2)
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested"
                ) as actual:
                    assert isinstance(actual.foo.variable.data, da.Array)
                    assert actual.foo.variable.data.chunks == ((5, 5),)
                    assert_identical(original, actual)
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested", chunks={"x": 3}
                ) as actual:
                    assert actual.foo.variable.data.chunks == ((3, 2, 3, 2),)

        with pytest.raises(OSError, match=r"no files to open"):
            open_mfdataset("foo-bar-baz-*.nc")
        with pytest.raises(ValueError, match=r"wild-card"):
            open_mfdataset("http://some/remote/uri")

    @requires_fsspec
    def test_open_mfdataset_no_files(self) -> None:
        pytest.importorskip("aiobotocore")

        # glob is attempted as of #4823, but finds no files
        with pytest.raises(OSError, match=r"no files"):
            open_mfdataset("http://some/remote/uri", engine="zarr")

    def test_open_mfdataset_2d(self) -> None:
        original = Dataset({"foo": (["x", "y"], np.random.randn(10, 8))})
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                with create_tmp_file() as tmp3:
                    with create_tmp_file() as tmp4:
                        original.isel(x=slice(5), y=slice(4)).to_netcdf(tmp1)
                        original.isel(x=slice(5, 10), y=slice(4)).to_netcdf(tmp2)
                        original.isel(x=slice(5), y=slice(4, 8)).to_netcdf(tmp3)
                        original.isel(x=slice(5, 10), y=slice(4, 8)).to_netcdf(tmp4)
                        with open_mfdataset(
                            [[tmp1, tmp2], [tmp3, tmp4]],
                            combine="nested",
                            concat_dim=["y", "x"],
                        ) as actual:
                            assert isinstance(actual.foo.variable.data, da.Array)
                            assert actual.foo.variable.data.chunks == ((5, 5), (4, 4))
                            assert_identical(original, actual)
                        with open_mfdataset(
                            [[tmp1, tmp2], [tmp3, tmp4]],
                            combine="nested",
                            concat_dim=["y", "x"],
                            chunks={"x": 3, "y": 2},
                        ) as actual:
                            assert actual.foo.variable.data.chunks == (
                                (3, 2, 3, 2),
                                (2, 2, 2, 2),
                            )

    def test_open_mfdataset_pathlib(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmps1:
            with create_tmp_file() as tmps2:
                tmp1 = Path(tmps1)
                tmp2 = Path(tmps2)
                original.isel(x=slice(5)).to_netcdf(tmp1)
                original.isel(x=slice(5, 10)).to_netcdf(tmp2)
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested"
                ) as actual:
                    assert_identical(original, actual)

    def test_open_mfdataset_2d_pathlib(self) -> None:
        original = Dataset({"foo": (["x", "y"], np.random.randn(10, 8))})
        with create_tmp_file() as tmps1:
            with create_tmp_file() as tmps2:
                with create_tmp_file() as tmps3:
                    with create_tmp_file() as tmps4:
                        tmp1 = Path(tmps1)
                        tmp2 = Path(tmps2)
                        tmp3 = Path(tmps3)
                        tmp4 = Path(tmps4)
                        original.isel(x=slice(5), y=slice(4)).to_netcdf(tmp1)
                        original.isel(x=slice(5, 10), y=slice(4)).to_netcdf(tmp2)
                        original.isel(x=slice(5), y=slice(4, 8)).to_netcdf(tmp3)
                        original.isel(x=slice(5, 10), y=slice(4, 8)).to_netcdf(tmp4)
                        with open_mfdataset(
                            [[tmp1, tmp2], [tmp3, tmp4]],
                            combine="nested",
                            concat_dim=["y", "x"],
                        ) as actual:
                            assert_identical(original, actual)

    def test_open_mfdataset_2(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                original.isel(x=slice(5)).to_netcdf(tmp1)
                original.isel(x=slice(5, 10)).to_netcdf(tmp2)

                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested"
                ) as actual:
                    assert_identical(original, actual)

    def test_attrs_mfdataset(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                ds1 = original.isel(x=slice(5))
                ds2 = original.isel(x=slice(5, 10))
                ds1.attrs["test1"] = "foo"
                ds2.attrs["test2"] = "bar"
                ds1.to_netcdf(tmp1)
                ds2.to_netcdf(tmp2)
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested"
                ) as actual:
                    # presumes that attributes inherited from
                    # first dataset loaded
                    assert actual.test1 == ds1.test1
                    # attributes from ds2 are not retained, e.g.,
                    with pytest.raises(AttributeError, match=r"no attribute"):
                        actual.test2

    def test_open_mfdataset_attrs_file(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_files(2) as (tmp1, tmp2):
            ds1 = original.isel(x=slice(5))
            ds2 = original.isel(x=slice(5, 10))
            ds1.attrs["test1"] = "foo"
            ds2.attrs["test2"] = "bar"
            ds1.to_netcdf(tmp1)
            ds2.to_netcdf(tmp2)
            with open_mfdataset(
                [tmp1, tmp2], concat_dim="x", combine="nested", attrs_file=tmp2
            ) as actual:
                # attributes are inherited from the master file
                assert actual.attrs["test2"] == ds2.attrs["test2"]
                # attributes from ds1 are not retained, e.g.,
                assert "test1" not in actual.attrs

    def test_open_mfdataset_attrs_file_path(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_files(2) as (tmps1, tmps2):
            tmp1 = Path(tmps1)
            tmp2 = Path(tmps2)
            ds1 = original.isel(x=slice(5))
            ds2 = original.isel(x=slice(5, 10))
            ds1.attrs["test1"] = "foo"
            ds2.attrs["test2"] = "bar"
            ds1.to_netcdf(tmp1)
            ds2.to_netcdf(tmp2)
            with open_mfdataset(
                [tmp1, tmp2], concat_dim="x", combine="nested", attrs_file=tmp2
            ) as actual:
                # attributes are inherited from the master file
                assert actual.attrs["test2"] == ds2.attrs["test2"]
                # attributes from ds1 are not retained, e.g.,
                assert "test1" not in actual.attrs

    def test_open_mfdataset_auto_combine(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10)), "x": np.arange(10)})
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                original.isel(x=slice(5)).to_netcdf(tmp1)
                original.isel(x=slice(5, 10)).to_netcdf(tmp2)

                with open_mfdataset([tmp2, tmp1], combine="by_coords") as actual:
                    assert_identical(original, actual)

    def test_open_mfdataset_raise_on_bad_combine_args(self) -> None:
        # Regression test for unhelpful error shown in #5230
        original = Dataset({"foo": ("x", np.random.randn(10)), "x": np.arange(10)})
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                original.isel(x=slice(5)).to_netcdf(tmp1)
                original.isel(x=slice(5, 10)).to_netcdf(tmp2)
                with pytest.raises(ValueError, match="`concat_dim` has no effect"):
                    open_mfdataset([tmp1, tmp2], concat_dim="x")

    def test_encoding_mfdataset(self) -> None:
        original = Dataset(
            {
                "foo": ("t", np.random.randn(10)),
                "t": ("t", pd.date_range(start="2010-01-01", periods=10, freq="1D")),
            }
        )
        original.t.encoding["units"] = "days since 2010-01-01"

        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                ds1 = original.isel(t=slice(5))
                ds2 = original.isel(t=slice(5, 10))
                ds1.t.encoding["units"] = "days since 2010-01-01"
                ds2.t.encoding["units"] = "days since 2000-01-01"
                ds1.to_netcdf(tmp1)
                ds2.to_netcdf(tmp2)
                with open_mfdataset([tmp1, tmp2], combine="nested") as actual:
                    assert actual.t.encoding["units"] == original.t.encoding["units"]
                    assert actual.t.encoding["units"] == ds1.t.encoding["units"]
                    assert actual.t.encoding["units"] != ds2.t.encoding["units"]

    def test_preprocess_mfdataset(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmp:
            original.to_netcdf(tmp)

            def preprocess(ds):
                return ds.assign_coords(z=0)

            expected = preprocess(original)
            with open_mfdataset(
                tmp, preprocess=preprocess, combine="by_coords"
            ) as actual:
                assert_identical(expected, actual)

    def test_save_mfdataset_roundtrip(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        datasets = [original.isel(x=slice(5)), original.isel(x=slice(5, 10))]
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                save_mfdataset(datasets, [tmp1, tmp2])
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested"
                ) as actual:
                    assert_identical(actual, original)

    def test_save_mfdataset_invalid(self) -> None:
        ds = Dataset()
        with pytest.raises(ValueError, match=r"cannot use mode"):
            save_mfdataset([ds, ds], ["same", "same"])
        with pytest.raises(ValueError, match=r"same length"):
            save_mfdataset([ds, ds], ["only one path"])

    def test_save_mfdataset_invalid_dataarray(self) -> None:
        # regression test for GH1555
        da = DataArray([1, 2])
        with pytest.raises(TypeError, match=r"supports writing Dataset"):
            save_mfdataset([da], ["dataarray"])

    def test_save_mfdataset_pathlib_roundtrip(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        datasets = [original.isel(x=slice(5)), original.isel(x=slice(5, 10))]
        with create_tmp_file() as tmps1:
            with create_tmp_file() as tmps2:
                tmp1 = Path(tmps1)
                tmp2 = Path(tmps2)
                save_mfdataset(datasets, [tmp1, tmp2])
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim="x", combine="nested"
                ) as actual:
                    assert_identical(actual, original)

    def test_save_mfdataset_pass_kwargs(self) -> None:
        # create a timeseries to store in a netCDF file
        times = [0, 1]
        time = xr.DataArray(times, dims=("time",))

        # create a simple dataset to write using save_mfdataset
        test_ds = xr.Dataset()
        test_ds["time"] = time

        # make sure the times are written as double and
        # turn off fill values
        encoding = dict(time=dict(dtype="double"))
        unlimited_dims = ["time"]

        # set the output file name
        output_path = "test.nc"

        # attempt to write the dataset with the encoding and unlimited args
        # passed through
        xr.save_mfdataset(
            [test_ds], [output_path], encoding=encoding, unlimited_dims=unlimited_dims
        )

    def test_open_and_do_math(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmp:
            original.to_netcdf(tmp)
            with open_mfdataset(tmp, combine="by_coords") as ds:
                actual = 1.0 * ds
                assert_allclose(original, actual, decode_bytes=False)

    def test_open_mfdataset_concat_dim_none(self) -> None:
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                data = Dataset({"x": 0})
                data.to_netcdf(tmp1)
                Dataset({"x": np.nan}).to_netcdf(tmp2)
                with open_mfdataset(
                    [tmp1, tmp2], concat_dim=None, combine="nested"
                ) as actual:
                    assert_identical(data, actual)

    def test_open_mfdataset_concat_dim_default_none(self) -> None:
        with create_tmp_file() as tmp1:
            with create_tmp_file() as tmp2:
                data = Dataset({"x": 0})
                data.to_netcdf(tmp1)
                Dataset({"x": np.nan}).to_netcdf(tmp2)
                with open_mfdataset([tmp1, tmp2], combine="nested") as actual:
                    assert_identical(data, actual)

    def test_open_dataset(self) -> None:
        original = Dataset({"foo": ("x", np.random.randn(10))})
        with create_tmp_file() as tmp:
            original.to_netcdf(tmp)
            with open_dataset(tmp, chunks={"x": 5}) as actual:
                assert isinstance(actual.foo.variable.data, da.Array)
                assert actual.foo.variable.data.chunks == ((5, 5),)
                assert_identical(original, actual)
            with open_dataset(tmp, chunks=5) as actual:
                assert_identical(original, actual)
            with open_dataset(tmp) as actual:
                assert isinstance(actual.foo.variable.data, np.ndarray)
                assert_identical(original, actual)

    def test_open_single_dataset(self) -> None:
        # Test for issue GH #1988. This makes sure that the
        # concat_dim is utilized when specified in open_mfdataset().
        rnddata = np.random.randn(10)
        original = Dataset({"foo": ("x", rnddata)})
        dim = DataArray([100], name="baz", dims="baz")
        expected = Dataset(
            {"foo": (("baz", "x"), rnddata[np.newaxis, :])}, {"baz": [100]}
        )
        with create_tmp_file() as tmp:
            original.to_netcdf(tmp)
            with open_mfdataset([tmp], concat_dim=dim, combine="nested") as actual:
                assert_identical(expected, actual)

    def test_open_multi_dataset(self) -> None:
        # Test for issue GH #1988 and #2647. This makes sure that the
        # concat_dim is utilized when specified in open_mfdataset().
        # The additional wrinkle is to ensure that a length greater
        # than one is tested as well due to numpy's implicit casting
        # of 1-length arrays to booleans in tests, which allowed
        # #2647 to still pass the test_open_single_dataset(),
        # which is itself still needed as-is because the original
        # bug caused one-length arrays to not be used correctly
        # in concatenation.
        rnddata = np.random.randn(10)
        original = Dataset({"foo": ("x", rnddata)})
        dim = DataArray([100, 150], name="baz", dims="baz")
        expected = Dataset(
            {"foo": (("baz", "x"), np.tile(rnddata[np.newaxis, :], (2, 1)))},
            {"baz": [100, 150]},
        )
        with create_tmp_file() as tmp1, create_tmp_file() as tmp2:
            original.to_netcdf(tmp1)
            original.to_netcdf(tmp2)
            with open_mfdataset(
                [tmp1, tmp2], concat_dim=dim, combine="nested"
            ) as actual:
                assert_identical(expected, actual)

    @pytest.mark.xfail(reason="Flaky test. Very open to contributions on fixing this")
    def test_dask_roundtrip(self) -> None:
        with create_tmp_file() as tmp:
            data = create_test_data()
            data.to_netcdf(tmp)
            chunks = {"dim1": 4, "dim2": 4, "dim3": 4, "time": 10}
            with open_dataset(tmp, chunks=chunks) as dask_ds:
                assert_identical(data, dask_ds)
                with create_tmp_file() as tmp2:
                    dask_ds.to_netcdf(tmp2)
                    with open_dataset(tmp2) as on_disk:
                        assert_identical(data, on_disk)

    def test_deterministic_names(self) -> None:
        with create_tmp_file() as tmp:
            data = create_test_data()
            data.to_netcdf(tmp)
            with open_mfdataset(tmp, combine="by_coords") as ds:
                original_names = {k: v.data.name for k, v in ds.data_vars.items()}
            with open_mfdataset(tmp, combine="by_coords") as ds:
                repeat_names = {k: v.data.name for k, v in ds.data_vars.items()}
            for var_name, dask_name in original_names.items():
                assert var_name in dask_name
                assert dask_name[:13] == "open_dataset-"
            assert original_names == repeat_names

    def test_dataarray_compute(self) -> None:
        # Test DataArray.compute() on dask backend.
        # The test for Dataset.compute() is already in DatasetIOBase;
        # however dask is the only tested backend which supports DataArrays
        actual = DataArray([1, 2]).chunk()
        computed = actual.compute()
        assert not actual._in_memory
        assert computed._in_memory
        assert_allclose(actual, computed, decode_bytes=False)

    def test_save_mfdataset_compute_false_roundtrip(self) -> None:
        from dask.delayed import Delayed

        original = Dataset({"foo": ("x", np.random.randn(10))}).chunk()
        datasets = [original.isel(x=slice(5)), original.isel(x=slice(5, 10))]
        with create_tmp_file(allow_cleanup_failure=ON_WINDOWS) as tmp1:
            with create_tmp_file(allow_cleanup_failure=ON_WINDOWS) as tmp2:
                delayed_obj = save_mfdataset(
                    datasets, [tmp1, tmp2], engine=self.engine, compute=False
                )
                assert isinstance(delayed_obj, Delayed)
                delayed_obj.compute()
                with open_mfdataset(
                    [tmp1, tmp2], combine="nested", concat_dim="x"
                ) as actual:
                    assert_identical(actual, original)

    def test_load_dataset(self) -> None:
        with create_tmp_file() as tmp:
            original = Dataset({"foo": ("x", np.random.randn(10))})
            original.to_netcdf(tmp)
            ds = load_dataset(tmp)
            # this would fail if we used open_dataset instead of load_dataset
            ds.to_netcdf(tmp)

    def test_load_dataarray(self) -> None:
        with create_tmp_file() as tmp:
            original = Dataset({"foo": ("x", np.random.randn(10))})
            original.to_netcdf(tmp)
            ds = load_dataarray(tmp)
            # this would fail if we used open_dataarray instead of
            # load_dataarray
            ds.to_netcdf(tmp)

    @pytest.mark.skipif(
        ON_WINDOWS,
        reason="counting number of tasks in graph fails on windows for some reason",
    )
    def test_inline_array(self) -> None:
        with create_tmp_file() as tmp:
            original = Dataset({"foo": ("x", np.random.randn(10))})
            original.to_netcdf(tmp)
            chunks = {"time": 10}

            def num_graph_nodes(obj):
                return len(obj.__dask_graph__())

            with (
                open_dataset(tmp, inline_array=False, chunks=chunks) as not_inlined_ds,
                open_dataset(tmp, inline_array=True, chunks=chunks) as inlined_ds,
            ):
                assert num_graph_nodes(inlined_ds) < num_graph_nodes(not_inlined_ds)

            with (
                open_dataarray(
                    tmp, inline_array=False, chunks=chunks
                ) as not_inlined_da,
                open_dataarray(tmp, inline_array=True, chunks=chunks) as inlined_da,
            ):
                assert num_graph_nodes(inlined_da) < num_graph_nodes(not_inlined_da)


@requires_scipy_or_netCDF4
@requires_pydap
@pytest.mark.filterwarnings("ignore:The binary mode of fromstring is deprecated")
class TestPydap:
    def convert_to_pydap_dataset(self, original):
        from pydap.model import BaseType, DatasetType, GridType

        ds = DatasetType("bears", **original.attrs)
        for key, var in original.data_vars.items():
            v = GridType(key)
            v[key] = BaseType(key, var.values, dimensions=var.dims, **var.attrs)
            for d in var.dims:
                v[d] = BaseType(d, var[d].values)
            ds[key] = v
        # check all dims are stored in ds
        for d in original.coords:
            ds[d] = BaseType(
                d, original[d].values, dimensions=(d,), **original[d].attrs
            )
        return ds

    @contextlib.contextmanager
    def create_datasets(self, **kwargs):
        with open_example_dataset("bears.nc") as expected:
            pydap_ds = self.convert_to_pydap_dataset(expected)
            actual = open_dataset(PydapDataStore(pydap_ds))
            # TODO solve this workaround:
            # netcdf converts string to byte not unicode
            expected["bears"] = expected["bears"].astype(str)
            yield actual, expected

    def test_cmp_local_file(self) -> None:
        with self.create_datasets() as (actual, expected):
            assert_equal(actual, expected)

            # global attributes should be global attributes on the dataset
            assert "NC_GLOBAL" not in actual.attrs
            assert "history" in actual.attrs

            # we don't check attributes exactly with assertDatasetIdentical()
            # because the test DAP server seems to insert some extra
            # attributes not found in the netCDF file.
            assert actual.attrs.keys() == expected.attrs.keys()

        with self.create_datasets() as (actual, expected):
            assert_equal(actual[{"l": 2}], expected[{"l": 2}])

        with self.create_datasets() as (actual, expected):
            assert_equal(actual.isel(i=0, j=-1), expected.isel(i=0, j=-1))

        with self.create_datasets() as (actual, expected):
            assert_equal(actual.isel(j=slice(1, 2)), expected.isel(j=slice(1, 2)))

        with self.create_datasets() as (actual, expected):
            indexers = {"i": [1, 0, 0], "j": [1, 2, 0, 1]}
            assert_equal(actual.isel(**indexers), expected.isel(**indexers))

        with self.create_datasets() as (actual, expected):
            indexers2 = {
                "i": DataArray([0, 1, 0], dims="a"),
                "j": DataArray([0, 2, 1], dims="a"),
            }
            assert_equal(actual.isel(**indexers2), expected.isel(**indexers2))

    def test_compatible_to_netcdf(self) -> None:
        # make sure it can be saved as a netcdf
        with self.create_datasets() as (actual, expected):
            with create_tmp_file() as tmp_file:
                actual.to_netcdf(tmp_file)
                with open_dataset(tmp_file) as actual2:
                    actual2["bears"] = actual2["bears"].astype(str)
                    assert_equal(actual2, expected)

    @requires_dask
    def test_dask(self) -> None:
        with self.create_datasets(chunks={"j": 2}) as (actual, expected):
            assert_equal(actual, expected)


@network
@requires_scipy_or_netCDF4
@requires_pydap
class TestPydapOnline(TestPydap):
    @contextlib.contextmanager
    def create_datasets(self, **kwargs):
        url = "http://test.opendap.org/opendap/hyrax/data/nc/bears.nc"
        actual = open_dataset(url, engine="pydap", **kwargs)
        with open_example_dataset("bears.nc") as expected:
            # workaround to restore string which is converted to byte
            expected["bears"] = expected["bears"].astype(str)
            yield actual, expected

    def test_session(self) -> None:
        from pydap.cas.urs import setup_session

        session = setup_session("XarrayTestUser", "Xarray2017")
        with mock.patch("pydap.client.open_url") as mock_func:
            xr.backends.PydapDataStore.open("http://test.url", session=session)
        mock_func.assert_called_with(
            url="http://test.url",
            application=None,
            session=session,
            output_grid=True,
            timeout=120,
        )


@requires_scipy
@requires_pynio
class TestPyNio(CFEncodedBase, NetCDF3Only):
    def test_write_store(self) -> None:
        # pynio is read-only for now
        pass

    @contextlib.contextmanager
    def open(self, path, **kwargs):
        with open_dataset(path, engine="pynio", **kwargs) as ds:
            yield ds

    def test_kwargs(self) -> None:
        kwargs = {"format": "grib"}
        path = os.path.join(os.path.dirname(__file__), "data", "example")
        with backends.NioDataStore(path, **kwargs) as store:
            assert store._manager._kwargs["format"] == "grib"

    def save(self, dataset, path, **kwargs):
        return dataset.to_netcdf(path, engine="scipy", **kwargs)

    def test_weakrefs(self) -> None:
        example = Dataset({"foo": ("x", np.arange(5.0))})
        expected = example.rename({"foo": "bar", "x": "y"})

        with create_tmp_file() as tmp_file:
            example.to_netcdf(tmp_file, engine="scipy")
            on_disk = open_dataset(tmp_file, engine="pynio")
            actual = on_disk.rename({"foo": "bar", "x": "y"})
            del on_disk  # trigger garbage collection
            assert_identical(actual, expected)


class TestEncodingInvalid:
    def test_extract_nc4_variable_encoding(self) -> None:
        var = xr.Variable(("x",), [1, 2, 3], {}, {"foo": "bar"})
        with pytest.raises(ValueError, match=r"unexpected encoding"):
            _extract_nc4_variable_encoding(var, raise_on_invalid=True)

        var = xr.Variable(("x",), [1, 2, 3], {}, {"chunking": (2, 1)})
        encoding = _extract_nc4_variable_encoding(var)
        assert {} == encoding

        # regression test
        var = xr.Variable(("x",), [1, 2, 3], {}, {"shuffle": True})
        encoding = _extract_nc4_variable_encoding(var, raise_on_invalid=True)
        assert {"shuffle": True} == encoding

        # Variables with unlim dims must be chunked on output.
        var = xr.Variable(("x",), [1, 2, 3], {}, {"contiguous": True})
        encoding = _extract_nc4_variable_encoding(var, unlimited_dims=("x",))
        assert {} == encoding

    @requires_netCDF4
    def test_extract_nc4_variable_encoding_netcdf4(self):
        # New netCDF4 1.6.0 compression argument.
        var = xr.Variable(("x",), [1, 2, 3], {}, {"compression": "szlib"})
        _extract_nc4_variable_encoding(var, backend="netCDF4", raise_on_invalid=True)

    def test_extract_h5nc_encoding(self) -> None:
        # not supported with h5netcdf (yet)
        var = xr.Variable(("x",), [1, 2, 3], {}, {"least_sigificant_digit": 2})
        with pytest.raises(ValueError, match=r"unexpected encoding"):
            _extract_nc4_variable_encoding(var, raise_on_invalid=True)


class MiscObject:
    pass


@requires_netCDF4
class TestValidateAttrs:
    def test_validating_attrs(self) -> None:
        def new_dataset():
            return Dataset({"data": ("y", np.arange(10.0))}, {"y": np.arange(10)})

        def new_dataset_and_dataset_attrs():
            ds = new_dataset()
            return ds, ds.attrs

        def new_dataset_and_data_attrs():
            ds = new_dataset()
            return ds, ds.data.attrs

        def new_dataset_and_coord_attrs():
            ds = new_dataset()
            return ds, ds.coords["y"].attrs

        for new_dataset_and_attrs in [
            new_dataset_and_dataset_attrs,
            new_dataset_and_data_attrs,
            new_dataset_and_coord_attrs,
        ]:
            ds, attrs = new_dataset_and_attrs()

            attrs[123] = "test"
            with pytest.raises(TypeError, match=r"Invalid name for attr: 123"):
                ds.to_netcdf("test.nc")

            ds, attrs = new_dataset_and_attrs()
            attrs[MiscObject()] = "test"
            with pytest.raises(TypeError, match=r"Invalid name for attr: "):
                ds.to_netcdf("test.nc")

            ds, attrs = new_dataset_and_attrs()
            attrs[""] = "test"
            with pytest.raises(ValueError, match=r"Invalid name for attr '':"):
                ds.to_netcdf("test.nc")

            # This one should work
            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = "test"
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = {"a": 5}
            with pytest.raises(TypeError, match=r"Invalid value for attr 'test'"):
                ds.to_netcdf("test.nc")

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = MiscObject()
            with pytest.raises(TypeError, match=r"Invalid value for attr 'test'"):
                ds.to_netcdf("test.nc")

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = 5
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = 3.14
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = [1, 2, 3, 4]
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = (1.9, 2.5)
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = np.arange(5)
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = "This is a string"
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)

            ds, attrs = new_dataset_and_attrs()
            attrs["test"] = ""
            with create_tmp_file() as tmp_file:
                ds.to_netcdf(tmp_file)


@requires_scipy_or_netCDF4
class TestDataArrayToNetCDF:
    def test_dataarray_to_netcdf_no_name(self) -> None:
        original_da = DataArray(np.arange(12).reshape((3, 4)))

        with create_tmp_file() as tmp:
            original_da.to_netcdf(tmp)

            with open_dataarray(tmp) as loaded_da:
                assert_identical(original_da, loaded_da)

    def test_dataarray_to_netcdf_with_name(self) -> None:
        original_da = DataArray(np.arange(12).reshape((3, 4)), name="test")

        with create_tmp_file() as tmp:
            original_da.to_netcdf(tmp)

            with open_dataarray(tmp) as loaded_da:
                assert_identical(original_da, loaded_da)

    def test_dataarray_to_netcdf_coord_name_clash(self) -> None:
        original_da = DataArray(
            np.arange(12).reshape((3, 4)), dims=["x", "y"], name="x"
        )

        with create_tmp_file() as tmp:
            original_da.to_netcdf(tmp)

            with open_dataarray(tmp) as loaded_da:
                assert_identical(original_da, loaded_da)

    def test_open_dataarray_options(self) -> None:
        data = DataArray(np.arange(5), coords={"y": ("x", range(5))}, dims=["x"])

        with create_tmp_file() as tmp:
            data.to_netcdf(tmp)

            expected = data.drop_vars("y")
            with open_dataarray(tmp, drop_variables=["y"]) as loaded:
                assert_identical(expected, loaded)

    @requires_scipy
    def test_dataarray_to_netcdf_return_bytes(self) -> None:
        # regression test for GH1410
        data = xr.DataArray([1, 2, 3])
        output = data.to_netcdf()
        assert isinstance(output, bytes)

    def test_dataarray_to_netcdf_no_name_pathlib(self) -> None:
        original_da = DataArray(np.arange(12).reshape((3, 4)))

        with create_tmp_file() as tmps:
            tmp = Path(tmps)
            original_da.to_netcdf(tmp)

            with open_dataarray(tmp) as loaded_da:
                assert_identical(original_da, loaded_da)


@requires_zarr
class TestDataArrayToZarr:
    def test_dataarray_to_zarr_no_name(self, tmp_store) -> None:
        original_da = DataArray(np.arange(12).reshape((3, 4)))

        original_da.to_zarr(tmp_store)

        with open_dataarray(tmp_store, engine="zarr") as loaded_da:
            assert_identical(original_da, loaded_da)

    def test_dataarray_to_zarr_with_name(self, tmp_store) -> None:
        original_da = DataArray(np.arange(12).reshape((3, 4)), name="test")

        original_da.to_zarr(tmp_store)

        with open_dataarray(tmp_store, engine="zarr") as loaded_da:
            assert_identical(original_da, loaded_da)

    def test_dataarray_to_zarr_coord_name_clash(self, tmp_store) -> None:
        original_da = DataArray(
            np.arange(12).reshape((3, 4)), dims=["x", "y"], name="x"
        )

        original_da.to_zarr(tmp_store)

        with open_dataarray(tmp_store, engine="zarr") as loaded_da:
            assert_identical(original_da, loaded_da)

    def test_open_dataarray_options(self, tmp_store) -> None:
        data = DataArray(np.arange(5), coords={"y": ("x", range(5))}, dims=["x"])

        data.to_zarr(tmp_store)

        expected = data.drop_vars("y")
        with open_dataarray(tmp_store, engine="zarr", drop_variables=["y"]) as loaded:
            assert_identical(expected, loaded)

    @requires_dask
    def test_dataarray_to_zarr_compute_false(self, tmp_store) -> None:
        from dask.delayed import Delayed

        original_da = DataArray(np.arange(12).reshape((3, 4)))

        output = original_da.to_zarr(tmp_store, compute=False)
        assert isinstance(output, Delayed)
        output.compute()
        with open_dataarray(tmp_store, engine="zarr") as loaded_da:
            assert_identical(original_da, loaded_da)


@requires_scipy_or_netCDF4
def test_no_warning_from_dask_effective_get() -> None:
    with create_tmp_file() as tmpfile:
        with assert_no_warnings():
            ds = Dataset()
            ds.to_netcdf(tmpfile)


@requires_scipy_or_netCDF4
def test_source_encoding_always_present() -> None:
    # Test for GH issue #2550.
    rnddata = np.random.randn(10)
    original = Dataset({"foo": ("x", rnddata)})
    with create_tmp_file() as tmp:
        original.to_netcdf(tmp)
        with open_dataset(tmp) as ds:
            assert ds.encoding["source"] == tmp


@requires_scipy_or_netCDF4
def test_source_encoding_always_present_with_pathlib() -> None:
    # Test for GH issue #5888.
    rnddata = np.random.randn(10)
    original = Dataset({"foo": ("x", rnddata)})
    with create_tmp_file() as tmp:
        original.to_netcdf(tmp)
        with open_dataset(Path(tmp)) as ds:
            assert ds.encoding["source"] == tmp


def _assert_no_dates_out_of_range_warning(record):
    undesired_message = "dates out of range"
    for warning in record:
        assert undesired_message not in str(warning.message)


@requires_scipy_or_netCDF4
@pytest.mark.parametrize("calendar", _STANDARD_CALENDARS)
def test_use_cftime_standard_calendar_default_in_range(calendar) -> None:
    x = [0, 1]
    time = [0, 720]
    units_date = "2000-01-01"
    units = "days since 2000-01-01"
    original = DataArray(x, [("time", time)], name="x").to_dataset()
    for v in ["x", "time"]:
        original[v].attrs["units"] = units
        original[v].attrs["calendar"] = calendar

    x_timedeltas = np.array(x).astype("timedelta64[D]")
    time_timedeltas = np.array(time).astype("timedelta64[D]")
    decoded_x = np.datetime64(units_date, "ns") + x_timedeltas
    decoded_time = np.datetime64(units_date, "ns") + time_timedeltas
    expected_x = DataArray(decoded_x, [("time", decoded_time)], name="x")
    expected_time = DataArray(decoded_time, [("time", decoded_time)], name="time")

    with create_tmp_file() as tmp_file:
        original.to_netcdf(tmp_file)
        with warnings.catch_warnings(record=True) as record:
            with open_dataset(tmp_file) as ds:
                assert_identical(expected_x, ds.x)
                assert_identical(expected_time, ds.time)
            _assert_no_dates_out_of_range_warning(record)


@requires_cftime
@requires_scipy_or_netCDF4
@pytest.mark.parametrize("calendar", _STANDARD_CALENDARS)
@pytest.mark.parametrize("units_year", [1500, 2500])
def test_use_cftime_standard_calendar_default_out_of_range(
    calendar, units_year
) -> None:
    import cftime

    x = [0, 1]
    time = [0, 720]
    units = f"days since {units_year}-01-01"
    original = DataArray(x, [("time", time)], name="x").to_dataset()
    for v in ["x", "time"]:
        original[v].attrs["units"] = units
        original[v].attrs["calendar"] = calendar

    decoded_x = cftime.num2date(x, units, calendar, only_use_cftime_datetimes=True)
    decoded_time = cftime.num2date(
        time, units, calendar, only_use_cftime_datetimes=True
    )
    expected_x = DataArray(decoded_x, [("time", decoded_time)], name="x")
    expected_time = DataArray(decoded_time, [("time", decoded_time)], name="time")

    with create_tmp_file() as tmp_file:
        original.to_netcdf(tmp_file)
        with pytest.warns(SerializationWarning):
            with open_dataset(tmp_file) as ds:
                assert_identical(expected_x, ds.x)
                assert_identical(expected_time, ds.time)


@requires_cftime
@requires_scipy_or_netCDF4
@pytest.mark.parametrize("calendar", _ALL_CALENDARS)
@pytest.mark.parametrize("units_year", [1500, 2000, 2500])
def test_use_cftime_true(calendar, units_year) -> None:
    import cftime

    x = [0, 1]
    time = [0, 720]
    units = f"days since {units_year}-01-01"
    original = DataArray(x, [("time", time)], name="x").to_dataset()
    for v in ["x", "time"]:
        original[v].attrs["units"] = units
        original[v].attrs["calendar"] = calendar

    decoded_x = cftime.num2date(x, units, calendar, only_use_cftime_datetimes=True)
    decoded_time = cftime.num2date(
        time, units, calendar, only_use_cftime_datetimes=True
    )
    expected_x = DataArray(decoded_x, [("time", decoded_time)], name="x")
    expected_time = DataArray(decoded_time, [("time", decoded_time)], name="time")

    with create_tmp_file() as tmp_file:
        original.to_netcdf(tmp_file)
        with warnings.catch_warnings(record=True) as record:
            with open_dataset(tmp_file, use_cftime=True) as ds:
                assert_identical(expected_x, ds.x)
                assert_identical(expected_time, ds.time)
            _assert_no_dates_out_of_range_warning(record)


@requires_scipy_or_netCDF4
@pytest.mark.parametrize("calendar", _STANDARD_CALENDARS)
def test_use_cftime_false_standard_calendar_in_range(calendar) -> None:
    x = [0, 1]
    time = [0, 720]
    units_date = "2000-01-01"
    units = "days since 2000-01-01"
    original = DataArray(x, [("time", time)], name="x").to_dataset()
    for v in ["x", "time"]:
        original[v].attrs["units"] = units
        original[v].attrs["calendar"] = calendar

    x_timedeltas = np.array(x).astype("timedelta64[D]")
    time_timedeltas = np.array(time).astype("timedelta64[D]")
    decoded_x = np.datetime64(units_date, "ns") + x_timedeltas
    decoded_time = np.datetime64(units_date, "ns") + time_timedeltas
    expected_x = DataArray(decoded_x, [("time", decoded_time)], name="x")
    expected_time = DataArray(decoded_time, [("time", decoded_time)], name="time")

    with create_tmp_file() as tmp_file:
        original.to_netcdf(tmp_file)
        with warnings.catch_warnings(record=True) as record:
            with open_dataset(tmp_file, use_cftime=False) as ds:
                assert_identical(expected_x, ds.x)
                assert_identical(expected_time, ds.time)
            _assert_no_dates_out_of_range_warning(record)


@requires_scipy_or_netCDF4
@pytest.mark.parametrize("calendar", _STANDARD_CALENDARS)
@pytest.mark.parametrize("units_year", [1500, 2500])
def test_use_cftime_false_standard_calendar_out_of_range(calendar, units_year) -> None:
    x = [0, 1]
    time = [0, 720]
    units = f"days since {units_year}-01-01"
    original = DataArray(x, [("time", time)], name="x").to_dataset()
    for v in ["x", "time"]:
        original[v].attrs["units"] = units
        original[v].attrs["calendar"] = calendar

    with create_tmp_file() as tmp_file:
        original.to_netcdf(tmp_file)
        with pytest.raises((OutOfBoundsDatetime, ValueError)):
            open_dataset(tmp_file, use_cftime=False)


@requires_scipy_or_netCDF4
@pytest.mark.parametrize("calendar", _NON_STANDARD_CALENDARS)
@pytest.mark.parametrize("units_year", [1500, 2000, 2500])
def test_use_cftime_false_nonstandard_calendar(calendar, units_year) -> None:
    x = [0, 1]
    time = [0, 720]
    units = f"days since {units_year}"
    original = DataArray(x, [("time", time)], name="x").to_dataset()
    for v in ["x", "time"]:
        original[v].attrs["units"] = units
        original[v].attrs["calendar"] = calendar

    with create_tmp_file() as tmp_file:
        original.to_netcdf(tmp_file)
        with pytest.raises((OutOfBoundsDatetime, ValueError)):
            open_dataset(tmp_file, use_cftime=False)


@pytest.mark.parametrize("engine", ["netcdf4", "scipy"])
def test_invalid_netcdf_raises(engine) -> None:
    data = create_test_data()
    with pytest.raises(ValueError, match=r"unrecognized option 'invalid_netcdf'"):
        data.to_netcdf("foo.nc", engine=engine, invalid_netcdf=True)


@requires_zarr
def test_encode_zarr_attr_value() -> None:
    # array -> list
    arr = np.array([1, 2, 3])
    expected1 = [1, 2, 3]
    actual1 = backends.zarr.encode_zarr_attr_value(arr)
    assert isinstance(actual1, list)
    assert actual1 == expected1

    # scalar array -> scalar
    sarr = np.array(1)[()]
    expected2 = 1
    actual2 = backends.zarr.encode_zarr_attr_value(sarr)
    assert isinstance(actual2, int)
    assert actual2 == expected2

    # string -> string (no change)
    expected3 = "foo"
    actual3 = backends.zarr.encode_zarr_attr_value(expected3)
    assert isinstance(actual3, str)
    assert actual3 == expected3


@requires_zarr
def test_extract_zarr_variable_encoding() -> None:
    var = xr.Variable("x", [1, 2])
    actual = backends.zarr.extract_zarr_variable_encoding(var)
    assert "chunks" in actual
    assert actual["chunks"] is None

    var = xr.Variable("x", [1, 2], encoding={"chunks": (1,)})
    actual = backends.zarr.extract_zarr_variable_encoding(var)
    assert actual["chunks"] == (1,)

    # does not raise on invalid
    var = xr.Variable("x", [1, 2], encoding={"foo": (1,)})
    actual = backends.zarr.extract_zarr_variable_encoding(var)

    # raises on invalid
    var = xr.Variable("x", [1, 2], encoding={"foo": (1,)})
    with pytest.raises(ValueError, match=r"unexpected encoding parameters"):
        actual = backends.zarr.extract_zarr_variable_encoding(
            var, raise_on_invalid=True
        )


@requires_zarr
@requires_fsspec
@pytest.mark.filterwarnings("ignore:deallocating CachingFileManager")
def test_open_fsspec() -> None:
    import fsspec
    import zarr

    if not hasattr(zarr.storage, "FSStore") or not hasattr(
        zarr.storage.FSStore, "getitems"
    ):
        pytest.skip("zarr too old")

    ds = open_dataset(os.path.join(os.path.dirname(__file__), "data", "example_1.nc"))

    m = fsspec.filesystem("memory")
    mm = m.get_mapper("out1.zarr")
    ds.to_zarr(mm)  # old interface
    ds0 = ds.copy()
    ds0["time"] = ds.time + pd.to_timedelta("1 day")
    mm = m.get_mapper("out2.zarr")
    ds0.to_zarr(mm)  # old interface

    # single dataset
    url = "memory://out2.zarr"
    ds2 = open_dataset(url, engine="zarr")
    xr.testing.assert_equal(ds0, ds2)

    # single dataset with caching
    url = "simplecache::memory://out2.zarr"
    ds2 = open_dataset(url, engine="zarr")
    xr.testing.assert_equal(ds0, ds2)

    # open_mfdataset requires dask
    if has_dask:
        # multi dataset
        url = "memory://out*.zarr"
        ds2 = open_mfdataset(url, engine="zarr")
        xr.testing.assert_equal(xr.concat([ds, ds0], dim="time"), ds2)

        # multi dataset with caching
        url = "simplecache::memory://out*.zarr"
        ds2 = open_mfdataset(url, engine="zarr")
        xr.testing.assert_equal(xr.concat([ds, ds0], dim="time"), ds2)


@requires_h5netcdf
@requires_netCDF4
def test_load_single_value_h5netcdf(tmp_path: Path) -> None:
    """Test that numeric single-element vector attributes are handled fine.

    At present (h5netcdf v0.8.1), the h5netcdf exposes single-valued numeric variable
    attributes as arrays of length 1, as opposed to scalars for the NetCDF4
    backend.  This was leading to a ValueError upon loading a single value from
    a file, see #4471.  Test that loading causes no failure.
    """
    ds = xr.Dataset(
        {
            "test": xr.DataArray(
                np.array([0]), dims=("x",), attrs={"scale_factor": 1, "add_offset": 0}
            )
        }
    )
    ds.to_netcdf(tmp_path / "test.nc")
    with xr.open_dataset(tmp_path / "test.nc", engine="h5netcdf") as ds2:
        ds2["test"][0].load()


@requires_zarr
@requires_dask
@pytest.mark.parametrize(
    "chunks", ["auto", -1, {}, {"x": "auto"}, {"x": -1}, {"x": "auto", "y": -1}]
)
def test_open_dataset_chunking_zarr(chunks, tmp_path: Path) -> None:
    encoded_chunks = 100
    dask_arr = da.from_array(
        np.ones((500, 500), dtype="float64"), chunks=encoded_chunks
    )
    ds = xr.Dataset(
        {
            "test": xr.DataArray(
                dask_arr,
                dims=("x", "y"),
            )
        }
    )
    ds["test"].encoding["chunks"] = encoded_chunks
    ds.to_zarr(tmp_path / "test.zarr")

    with dask.config.set({"array.chunk-size": "1MiB"}):
        expected = ds.chunk(chunks)
        with open_dataset(
            tmp_path / "test.zarr", engine="zarr", chunks=chunks
        ) as actual:
            xr.testing.assert_chunks_equal(actual, expected)


@requires_zarr
@requires_dask
@pytest.mark.parametrize(
    "chunks", ["auto", -1, {}, {"x": "auto"}, {"x": -1}, {"x": "auto", "y": -1}]
)
@pytest.mark.filterwarnings("ignore:The specified chunks separate")
def test_chunking_consintency(chunks, tmp_path: Path) -> None:
    encoded_chunks: dict[str, Any] = {}
    dask_arr = da.from_array(
        np.ones((500, 500), dtype="float64"), chunks=encoded_chunks
    )
    ds = xr.Dataset(
        {
            "test": xr.DataArray(
                dask_arr,
                dims=("x", "y"),
            )
        }
    )
    ds["test"].encoding["chunks"] = encoded_chunks
    ds.to_zarr(tmp_path / "test.zarr")
    ds.to_netcdf(tmp_path / "test.nc")

    with dask.config.set({"array.chunk-size": "1MiB"}):
        expected = ds.chunk(chunks)
        with xr.open_dataset(
            tmp_path / "test.zarr", engine="zarr", chunks=chunks
        ) as actual:
            xr.testing.assert_chunks_equal(actual, expected)

        with xr.open_dataset(tmp_path / "test.nc", chunks=chunks) as actual:
            xr.testing.assert_chunks_equal(actual, expected)


def _check_guess_can_open_and_open(entrypoint, obj, engine, expected):
    assert entrypoint.guess_can_open(obj)
    with open_dataset(obj, engine=engine) as actual:
        assert_identical(expected, actual)


@requires_netCDF4
def test_netcdf4_entrypoint(tmp_path: Path) -> None:
    entrypoint = NetCDF4BackendEntrypoint()
    ds = create_test_data()

    path = tmp_path / "foo"
    ds.to_netcdf(path, format="NETCDF3_CLASSIC")
    _check_guess_can_open_and_open(entrypoint, path, engine="netcdf4", expected=ds)
    _check_guess_can_open_and_open(entrypoint, str(path), engine="netcdf4", expected=ds)

    path = tmp_path / "bar"
    ds.to_netcdf(path, format="NETCDF4_CLASSIC")
    _check_guess_can_open_and_open(entrypoint, path, engine="netcdf4", expected=ds)
    _check_guess_can_open_and_open(entrypoint, str(path), engine="netcdf4", expected=ds)

    assert entrypoint.guess_can_open("http://something/remote")
    assert entrypoint.guess_can_open("something-local.nc")
    assert entrypoint.guess_can_open("something-local.nc4")
    assert entrypoint.guess_can_open("something-local.cdf")
    assert not entrypoint.guess_can_open("not-found-and-no-extension")

    path = tmp_path / "baz"
    with open(path, "wb") as f:
        f.write(b"not-a-netcdf-file")
    assert not entrypoint.guess_can_open(path)


@requires_scipy
def test_scipy_entrypoint(tmp_path: Path) -> None:
    entrypoint = ScipyBackendEntrypoint()
    ds = create_test_data()

    path = tmp_path / "foo"
    ds.to_netcdf(path, engine="scipy")
    _check_guess_can_open_and_open(entrypoint, path, engine="scipy", expected=ds)
    _check_guess_can_open_and_open(entrypoint, str(path), engine="scipy", expected=ds)
    with open(path, "rb") as f:
        _check_guess_can_open_and_open(entrypoint, f, engine="scipy", expected=ds)

    contents = ds.to_netcdf(engine="scipy")
    _check_guess_can_open_and_open(entrypoint, contents, engine="scipy", expected=ds)
    _check_guess_can_open_and_open(
        entrypoint, BytesIO(contents), engine="scipy", expected=ds
    )

    path = tmp_path / "foo.nc.gz"
    with gzip.open(path, mode="wb") as f:
        f.write(contents)
    _check_guess_can_open_and_open(entrypoint, path, engine="scipy", expected=ds)
    _check_guess_can_open_and_open(entrypoint, str(path), engine="scipy", expected=ds)

    assert entrypoint.guess_can_open("something-local.nc")
    assert entrypoint.guess_can_open("something-local.nc.gz")
    assert not entrypoint.guess_can_open("not-found-and-no-extension")
    assert not entrypoint.guess_can_open(b"not-a-netcdf-file")  # type: ignore[arg-type]


@requires_h5netcdf
def test_h5netcdf_entrypoint(tmp_path: Path) -> None:
    entrypoint = H5netcdfBackendEntrypoint()
    ds = create_test_data()

    path = tmp_path / "foo"
    ds.to_netcdf(path, engine="h5netcdf")
    _check_guess_can_open_and_open(entrypoint, path, engine="h5netcdf", expected=ds)
    _check_guess_can_open_and_open(
        entrypoint, str(path), engine="h5netcdf", expected=ds
    )
    with open(path, "rb") as f:
        _check_guess_can_open_and_open(entrypoint, f, engine="h5netcdf", expected=ds)

    assert entrypoint.guess_can_open("something-local.nc")
    assert entrypoint.guess_can_open("something-local.nc4")
    assert entrypoint.guess_can_open("something-local.cdf")
    assert not entrypoint.guess_can_open("not-found-and-no-extension")


@requires_netCDF4
@pytest.mark.parametrize("str_type", (str, np.str_))
def test_write_file_from_np_str(str_type, tmpdir) -> None:
    # https://github.com/pydata/xarray/pull/5264
    scenarios = [str_type(v) for v in ["scenario_a", "scenario_b", "scenario_c"]]
    years = range(2015, 2100 + 1)
    tdf = pd.DataFrame(
        data=np.random.random((len(scenarios), len(years))),
        columns=years,
        index=scenarios,
    )
    tdf.index.name = "scenario"
    tdf.columns.name = "year"
    tdf = tdf.stack()
    tdf.name = "tas"

    txr = tdf.to_xarray()

    txr.to_netcdf(tmpdir.join("test.nc"))


@requires_zarr
@requires_netCDF4
class TestNCZarr:
    @property
    def netcdfc_version(self):
        return Version(nc4.getlibversion().split()[0].split("-development")[0])

    def _create_nczarr(self, filename):
        if self.netcdfc_version < Version("4.8.1"):
            pytest.skip("requires netcdf-c>=4.8.1")
        if platform.system() == "Windows" and self.netcdfc_version == Version("4.8.1"):
            # Bug in netcdf-c==4.8.1 (typo: Nan instead of NaN)
            # https://github.com/Unidata/netcdf-c/issues/2265
            pytest.skip("netcdf-c==4.8.1 has issues on Windows")

        ds = create_test_data()
        # Drop dim3: netcdf-c does not support dtype='<U1'
        # https://github.com/Unidata/netcdf-c/issues/2259
        ds = ds.drop_vars("dim3")

        ds.to_netcdf(f"file://{filename}#mode=nczarr")
        return ds

    def test_open_nczarr(self) -> None:
        with create_tmp_file(suffix=".zarr") as tmp:
            expected = self._create_nczarr(tmp)
            actual = xr.open_zarr(tmp, consolidated=False)
            assert_identical(expected, actual)

    def test_overwriting_nczarr(self) -> None:
        with create_tmp_file(suffix=".zarr") as tmp:
            ds = self._create_nczarr(tmp)
            expected = ds[["var1"]]
            expected.to_zarr(tmp, mode="w")
            actual = xr.open_zarr(tmp, consolidated=False)
            assert_identical(expected, actual)

    @pytest.mark.parametrize("mode", ["a", "r+"])
    @pytest.mark.filterwarnings("ignore:.*non-consolidated metadata.*")
    def test_raise_writing_to_nczarr(self, mode) -> None:
        if self.netcdfc_version > Version("4.8.1"):
            pytest.skip("netcdf-c>4.8.1 adds the _ARRAY_DIMENSIONS attribute")

        with create_tmp_file(suffix=".zarr") as tmp:
            ds = self._create_nczarr(tmp)
            with pytest.raises(
                KeyError, match="missing the attribute `_ARRAY_DIMENSIONS`,"
            ):
                ds.to_zarr(tmp, mode=mode)


@requires_netCDF4
@requires_dask
def test_pickle_open_mfdataset_dataset():
    with open_example_mfdataset(["bears.nc"]) as ds:
        assert_identical(ds, pickle.loads(pickle.dumps(ds)))


@requires_zarr
def test_zarr_closing_internal_zip_store():
    store_name = "tmp.zarr.zip"
    original_da = DataArray(np.arange(12).reshape((3, 4)))
    original_da.to_zarr(store_name, mode="w")

    with open_dataarray(store_name, engine="zarr") as loaded_da:
        assert_identical(original_da, loaded_da)


@requires_zarr
class TestZarrRegionAuto:
    def test_zarr_region_auto_all(self, tmp_path):
        x = np.arange(0, 50, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )
        ds.to_zarr(tmp_path / "test.zarr")

        ds_region = 1 + ds.isel(x=slice(2, 4), y=slice(6, 8))
        ds_region.to_zarr(tmp_path / "test.zarr", region="auto")

        ds_updated = xr.open_zarr(tmp_path / "test.zarr")

        expected = ds.copy()
        expected["test"][2:4, 6:8] += 1
        assert_identical(ds_updated, expected)

    def test_zarr_region_auto_mixed(self, tmp_path):
        x = np.arange(0, 50, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )
        ds.to_zarr(tmp_path / "test.zarr")

        ds_region = 1 + ds.isel(x=slice(2, 4), y=slice(6, 8))
        ds_region.to_zarr(
            tmp_path / "test.zarr", region={"x": "auto", "y": slice(6, 8)}
        )

        ds_updated = xr.open_zarr(tmp_path / "test.zarr")

        expected = ds.copy()
        expected["test"][2:4, 6:8] += 1
        assert_identical(ds_updated, expected)

    def test_zarr_region_auto_noncontiguous(self, tmp_path):
        x = np.arange(0, 50, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )
        ds.to_zarr(tmp_path / "test.zarr")

        ds_region = 1 + ds.isel(x=[0, 2, 3], y=[5, 6])
        with pytest.raises(ValueError):
            ds_region.to_zarr(tmp_path / "test.zarr", region={"x": "auto", "y": "auto"})

    def test_zarr_region_auto_new_coord_vals(self, tmp_path):
        x = np.arange(0, 50, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )
        ds.to_zarr(tmp_path / "test.zarr")

        x = np.arange(5, 55, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )

        ds_region = 1 + ds.isel(x=slice(2, 4), y=slice(6, 8))
        with pytest.raises(KeyError):
            ds_region.to_zarr(tmp_path / "test.zarr", region={"x": "auto", "y": "auto"})

    def test_zarr_region_index_write(self, tmp_path):
        from xarray.backends.zarr import ZarrStore

        x = np.arange(0, 50, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )

        region_slice = dict(x=slice(2, 4), y=slice(6, 8))
        ds_region = 1 + ds.isel(region_slice)

        ds.to_zarr(tmp_path / "test.zarr")

        region: Mapping[str, slice] | Literal["auto"]
        for region in [region_slice, "auto"]:  # type: ignore
            with patch.object(
                ZarrStore,
                "set_variables",
                side_effect=ZarrStore.set_variables,
                autospec=True,
            ) as mock:
                ds_region.to_zarr(tmp_path / "test.zarr", region=region, mode="r+")

                # should write the data vars but never the index vars with auto mode
                for call in mock.call_args_list:
                    written_variables = call.args[1].keys()
                    assert "test" in written_variables
                    assert "x" not in written_variables
                    assert "y" not in written_variables

    def test_zarr_region_append(self, tmp_path):
        x = np.arange(0, 50, 10)
        y = np.arange(0, 20, 2)
        data = np.ones((5, 10))
        ds = xr.Dataset(
            {
                "test": xr.DataArray(
                    data,
                    dims=("x", "y"),
                    coords={"x": x, "y": y},
                )
            }
        )
        ds.to_zarr(tmp_path / "test.zarr")

        x_new = np.arange(40, 70, 10)
        data_new = np.ones((3, 10))
        ds_new = xr.Dataset(
            {
                "test": xr.DataArray(
                    data_new,
                    dims=("x", "y"),
                    coords={"x": x_new, "y": y},
                )
            }
        )

        # Don't allow auto region detection in append mode due to complexities in
        # implementing the overlap logic and lack of safety with parallel writes
        with pytest.raises(ValueError):
            ds_new.to_zarr(
                tmp_path / "test.zarr", mode="a", append_dim="x", region="auto"
            )


@requires_zarr
def test_zarr_region(tmp_path):
    x = np.arange(0, 50, 10)
    y = np.arange(0, 20, 2)
    data = np.ones((5, 10))
    ds = xr.Dataset(
        {
            "test": xr.DataArray(
                data,
                dims=("x", "y"),
                coords={"x": x, "y": y},
            )
        }
    )
    ds.to_zarr(tmp_path / "test.zarr")

    ds_transposed = ds.transpose("y", "x")

    ds_region = 1 + ds_transposed.isel(x=[0], y=[0])
    ds_region.to_zarr(
        tmp_path / "test.zarr", region={"x": slice(0, 1), "y": slice(0, 1)}
    )

    # Write without region
    ds_transposed.to_zarr(tmp_path / "test.zarr", mode="r+")