Implement DataArray.to_dask_dataframe() (pydata#7635)

* Add feature  to convert dataarray to dask dataframe. This is for the issue pydata#7409

* Add test for new method  dataarray.to_dask_dataframe()

* Changes after review

* Corrections in docstring and import

* docstring correction

* Remove name parameter

* Add feature  to convert dataarray to dask dataframe. This is for the issue pydata#7409

* Add test for new method  dataarray.to_dask_dataframe()

* Changes after review

* Corrections in docstring and import

* docstring correction

* Remove name parameter

* Corrected doc/whats-new.rst

* Update whats-new.rst

* Space corrections in docstring

* Whitespace correction in docstring

* Add white space in docstring line

* Whitespace correction

* Update line npartitions=1

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Revert "Update line npartitions=1"

This reverts commit 4bae82c.

Reverting commit .

* Add whitespace in npartitions=1

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Change example in docstring

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Change example in docstring

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update xarray/core/dataarray.py

Co-authored-by: Illviljan <[email protected]>

* Update doc/whats-new.rst

Co-authored-by: Illviljan <[email protected]>

* Add name check

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Add test for unnamed dataarray.

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Remove scalar array test

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Change error message

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update xarray/tests/test_dataarray.py

* Update whats-new.rst

* Update doc/whats-new.rst

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Illviljan <[email protected]>
Co-authored-by: Deepak Cherian <[email protected]>

Author: 4 people

doc/api.rst

@@ -632,6 +632,7 @@ DataArray methods
    DataArray.from_iris
    DataArray.from_series
    DataArray.to_cdms2
+   DataArray.to_dask_dataframe
    DataArray.to_dataframe
    DataArray.to_dataset
    DataArray.to_dict

doc/whats-new.rst

@@ -22,6 +22,8 @@ v2023.05.0 (unreleased)
 
 New Features
 ~~~~~~~~~~~~
+- Added new method :py:meth:`DataArray.to_dask_dataframe`, convert a dataarray into a dask dataframe (:issue:`7409`).
+  By `Deeksha <https://github.com/dsgreen2>`_.
 - Add support for lshift and rshift binary operators (`<<`, `>>`) on
   :py:class:`xr.DataArray` of type :py:class:`int` (:issue:`7727` , :pull:`7741`).
   By `Alan Brammer <https://github.com/abrammer>`_.

xarray/core/dataarray.py

@@ -57,6 +57,10 @@
 
     from numpy.typing import ArrayLike
 
+    try:
+        from dask.dataframe import DataFrame as DaskDataFrame
+    except ImportError:
+        DaskDataFrame = None  # type: ignore
     try:
         from dask.delayed import Delayed
     except ImportError:
@@ -6888,6 +6892,71 @@ def resample(
             **indexer_kwargs,
         )
 
+    def to_dask_dataframe(
+        self,
+        dim_order: Sequence[Hashable] | None = None,
+        set_index: bool = False,
+    ) -> DaskDataFrame:
+        """Convert this array into a dask.dataframe.DataFrame.
+
+        Parameters
+        ----------
+        dim_order : Sequence of Hashable or None , optional
+            Hierarchical dimension order for the resulting dataframe.
+            Array content is transposed to this order and then written out as flat
+            vectors in contiguous order, so the last dimension in this list
+            will be contiguous in the resulting DataFrame. This has a major influence
+            on which operations are efficient on the resulting dask dataframe.
+        set_index : bool, default: False
+            If set_index=True, the dask DataFrame is indexed by this dataset's
+            coordinate. Since dask DataFrames do not support multi-indexes,
+            set_index only works if the dataset only contains one dimension.
+
+        Returns
+        -------
+        dask.dataframe.DataFrame
+
+        Examples
+        --------
+        >>> da = xr.DataArray(
+        ...     np.arange(4 * 2 * 2).reshape(4, 2, 2),
+        ...     dims=("time", "lat", "lon"),
+        ...     coords={
+        ...         "time": np.arange(4),
+        ...         "lat": [-30, -20],
+        ...         "lon": [120, 130],
+        ...     },
+        ...     name="eg_dataarray",
+        ...     attrs={"units": "Celsius", "description": "Random temperature data"},
+        ... )
+        >>> da.to_dask_dataframe(["lat", "lon", "time"]).compute()
+            lat  lon  time  eg_dataarray
+        0   -30  120     0             0
+        1   -30  120     1             4
+        2   -30  120     2             8
+        3   -30  120     3            12
+        4   -30  130     0             1
+        5   -30  130     1             5
+        6   -30  130     2             9
+        7   -30  130     3            13
+        8   -20  120     0             2
+        9   -20  120     1             6
+        10  -20  120     2            10
+        11  -20  120     3            14
+        12  -20  130     0             3
+        13  -20  130     1             7
+        14  -20  130     2            11
+        15  -20  130     3            15
+        """
+        if self.name is None:
+            raise ValueError(
+                "Cannot convert an unnamed DataArray to a "
+                "dask dataframe : use the ``.rename`` method to assign a name."
+            )
+        name = self.name
+        ds = self._to_dataset_whole(name, shallow_copy=False)
+        return ds.to_dask_dataframe(dim_order, set_index)
+
     # this needs to be at the end, or mypy will confuse with `str`
     # https://mypy.readthedocs.io/en/latest/common_issues.html#dealing-with-conflicting-names
     str = utils.UncachedAccessor(StringAccessor["DataArray"])

xarray/tests/test_dataarray.py

@@ -3245,6 +3245,39 @@ def test_to_dataframe_0length(self) -> None:
         assert len(actual) == 0
         assert_array_equal(actual.index.names, list("ABC"))
 
+    @requires_dask
+    def test_to_dask_dataframe(self) -> None:
+        arr_np = np.arange(3 * 4).reshape(3, 4)
+        arr = DataArray(arr_np, [("B", [1, 2, 3]), ("A", list("cdef"))], name="foo")
+        expected = arr.to_series()
+        actual = arr.to_dask_dataframe()["foo"]
+
+        assert_array_equal(actual.values, expected.values)
+
+        actual = arr.to_dask_dataframe(dim_order=["A", "B"])["foo"]
+        assert_array_equal(arr_np.transpose().reshape(-1), actual.values)
+
+        # regression test for coords with different dimensions
+
+        arr.coords["C"] = ("B", [-1, -2, -3])
+        expected = arr.to_series().to_frame()
+        expected["C"] = [-1] * 4 + [-2] * 4 + [-3] * 4
+        expected = expected[["C", "foo"]]
+        actual = arr.to_dask_dataframe()[["C", "foo"]]
+
+        assert_array_equal(expected.values, actual.values)
+        assert_array_equal(expected.columns.values, actual.columns.values)
+
+        with pytest.raises(ValueError, match="does not match the set of dimensions"):
+            arr.to_dask_dataframe(dim_order=["B", "A", "C"])
+
+        arr.name = None
+        with pytest.raises(
+            ValueError,
+            match="Cannot convert an unnamed DataArray",
+        ):
+            arr.to_dask_dataframe()
+
     def test_to_pandas_name_matches_coordinate(self) -> None:
         # coordinate with same name as array
         arr = DataArray([1, 2, 3], dims="x", name="x")