pycoupler.LPJmLData

class pycoupler.LPJmLData(*args, **kwargs)

Bases: DataArray

Class for single LPJmL data arrays (input, output, etc.) with meta data and defined dimensions (cell, band, time).

Parameters:

• *args (tuple) – Arguments for the xarray.DataArray constructor.

• **kwargs (dict) – Keyword arguments for the xarray.DataArray constructor.

Variables:

• attrs (dict) – Attributes of the data array.

• coords (dict) – Coordinates of the data array.

add_meta(meta_data)

Add meta data to the data array.

Parameters:

meta_data (LPJmLMetaData) – Meta data to be added to the data array.

get_neighbourhood(id=True, cellsize=0.5)

Get the IDs of all neighbouring cells within a given size of cells.

Parameters:

• id (bool, default True) – If True, return cell ids, else return cell indices

• cellsize (float, default 0.5) – Size of cells in degrees.

Returns:

• return: Array with the IDs of all neighbouring cells.

• rtype: numpy.ndarray

transform()

TODO: implement function to convert cell into lon/lat format

property T

__getitem__()

all()

Reduce this DataArray’s data by applying all along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply all. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating all on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with all applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.all, dask.array.all, Dataset.all

Aggregation

User guide on reduction or aggregation operations.

Examples

>>> da = xr.DataArray(
...     np.array([True, True, True, True, True, False], dtype=bool),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 6B
array([ True,  True,  True,  True,  True, False])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.all()
<xarray.DataArray ()> Size: 1B
array(False)

any()

Reduce this DataArray’s data by applying any along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply any. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating any on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with any applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.any, dask.array.any, Dataset.any

Aggregation

User guide on reduction or aggregation operations.

Examples

>>> da = xr.DataArray(
...     np.array([True, True, True, True, True, False], dtype=bool),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 6B
array([ True,  True,  True,  True,  True, False])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.any()
<xarray.DataArray ()> Size: 1B
array(True)

argmax()

Index or indices of the maximum of the DataArray over one or more dimensions.

If a sequence is passed to ‘dim’, then result returned as dict of DataArrays, which can be passed directly to isel(). If a single str is passed to ‘dim’ then returns a DataArray with dtype int.

If there are multiple maxima, the indices of the first one found will be returned.

Parameters:

• dim (”…”, str, Iterable of Hashable or None, optional) – The dimensions over which to find the maximum. By default, finds maximum over all dimensions - for now returning an int for backward compatibility, but this is deprecated, in future will return a dict with indices for all dimensions; to return a dict with all dimensions now, pass ‘…’.

• axis (int or None, optional) – Axis over which to apply argmax. Only one of the ‘dim’ and ‘axis’ arguments can be supplied.

• keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

Returns:

result

Return type:

DataArray or dict of DataArray

See also

Variable.argmax, DataArray.idxmax

Examples

>>> array = xr.DataArray([0, 2, -1, 3], dims="x")
>>> array.max()
<xarray.DataArray ()> Size: 8B
array(3)
>>> array.argmax(...)
{'x': <xarray.DataArray ()> Size: 8B
array(3)}
>>> array.isel(array.argmax(...))
<xarray.DataArray ()> Size: 8B
array(3)

>>> array = xr.DataArray(
...     [[[3, 2, 1], [3, 1, 2], [2, 1, 3]], [[1, 3, 2], [2, 5, 1], [2, 3, 1]]],
...     dims=("x", "y", "z"),
... )
>>> array.max(dim="x")
<xarray.DataArray (y: 3, z: 3)> Size: 72B
array([[3, 3, 2],
       [3, 5, 2],
       [2, 3, 3]])
Dimensions without coordinates: y, z
>>> array.argmax(dim="x")
<xarray.DataArray (y: 3, z: 3)> Size: 72B
array([[0, 1, 1],
       [0, 1, 0],
       [0, 1, 0]])
Dimensions without coordinates: y, z
>>> array.argmax(dim=["x"])
{'x': <xarray.DataArray (y: 3, z: 3)> Size: 72B
array([[0, 1, 1],
       [0, 1, 0],
       [0, 1, 0]])
Dimensions without coordinates: y, z}
>>> array.max(dim=("x", "z"))
<xarray.DataArray (y: 3)> Size: 24B
array([3, 5, 3])
Dimensions without coordinates: y
>>> array.argmax(dim=["x", "z"])
{'x': <xarray.DataArray (y: 3)> Size: 24B
array([0, 1, 0])
Dimensions without coordinates: y, 'z': <xarray.DataArray (y: 3)> Size: 24B
array([0, 1, 2])
Dimensions without coordinates: y}
>>> array.isel(array.argmax(dim=["x", "z"]))
<xarray.DataArray (y: 3)> Size: 24B
array([3, 5, 3])
Dimensions without coordinates: y

argmin()

Index or indices of the minimum of the DataArray over one or more dimensions.

If a sequence is passed to ‘dim’, then result returned as dict of DataArrays, which can be passed directly to isel(). If a single str is passed to ‘dim’ then returns a DataArray with dtype int.

If there are multiple minima, the indices of the first one found will be returned.

Parameters:

• dim (”…”, str, Iterable of Hashable or None, optional) – The dimensions over which to find the minimum. By default, finds minimum over all dimensions - for now returning an int for backward compatibility, but this is deprecated, in future will return a dict with indices for all dimensions; to return a dict with all dimensions now, pass ‘…’.

• axis (int or None, optional) – Axis over which to apply argmin. Only one of the ‘dim’ and ‘axis’ arguments can be supplied.

• keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

Returns:

result

Return type:

DataArray or dict of DataArray

See also

Variable.argmin, DataArray.idxmin

Examples

>>> array = xr.DataArray([0, 2, -1, 3], dims="x")
>>> array.min()
<xarray.DataArray ()> Size: 8B
array(-1)
>>> array.argmin(...)
{'x': <xarray.DataArray ()> Size: 8B
array(2)}
>>> array.isel(array.argmin(...))
<xarray.DataArray ()> Size: 8B
array(-1)

>>> array = xr.DataArray(
...     [[[3, 2, 1], [3, 1, 2], [2, 1, 3]], [[1, 3, 2], [2, -5, 1], [2, 3, 1]]],
...     dims=("x", "y", "z"),
... )
>>> array.min(dim="x")
<xarray.DataArray (y: 3, z: 3)> Size: 72B
array([[ 1,  2,  1],
       [ 2, -5,  1],
       [ 2,  1,  1]])
Dimensions without coordinates: y, z
>>> array.argmin(dim="x")
<xarray.DataArray (y: 3, z: 3)> Size: 72B
array([[1, 0, 0],
       [1, 1, 1],
       [0, 0, 1]])
Dimensions without coordinates: y, z
>>> array.argmin(dim=["x"])
{'x': <xarray.DataArray (y: 3, z: 3)> Size: 72B
array([[1, 0, 0],
       [1, 1, 1],
       [0, 0, 1]])
Dimensions without coordinates: y, z}
>>> array.min(dim=("x", "z"))
<xarray.DataArray (y: 3)> Size: 24B
array([ 1, -5,  1])
Dimensions without coordinates: y
>>> array.argmin(dim=["x", "z"])
{'x': <xarray.DataArray (y: 3)> Size: 24B
array([0, 1, 0])
Dimensions without coordinates: y, 'z': <xarray.DataArray (y: 3)> Size: 24B
array([2, 1, 1])
Dimensions without coordinates: y}
>>> array.isel(array.argmin(dim=["x", "z"]))
<xarray.DataArray (y: 3)> Size: 24B
array([ 1, -5,  1])
Dimensions without coordinates: y

argsort(axis=-1, kind=None, order=None)

Returns the indices that would sort this array.

Refer to numpy.argsort for full documentation.

See also

numpy.argsort

equivalent function

as_numpy()

Coerces wrapped data and coordinates into numpy arrays, returning a DataArray.

See also

DataArray.to_numpy

Same but returns only the data as a numpy.ndarray object.

Dataset.as_numpy

Converts all variables in a Dataset.

DataArray.values, DataArray.data

assign_attrs()

Assign new attrs to this object.

Returns a new object equivalent to self.attrs.update(*args, **kwargs).

Parameters:

• *args – positional arguments passed into attrs.update.

• **kwargs – keyword arguments passed into attrs.update.

Examples

>>> dataset = xr.Dataset({"temperature": [25, 30, 27]})
>>> dataset
<xarray.Dataset> Size: 24B
Dimensions:      (temperature: 3)
Coordinates:
  * temperature  (temperature) int64 24B 25 30 27
Data variables:
    *empty*

>>> new_dataset = dataset.assign_attrs(
...     units="Celsius", description="Temperature data"
... )
>>> new_dataset
<xarray.Dataset> Size: 24B
Dimensions:      (temperature: 3)
Coordinates:
  * temperature  (temperature) int64 24B 25 30 27
Data variables:
    *empty*
:ivar units: Celsius
:ivar description: Temperature data

# Attributes of the new dataset

>>> new_dataset.attrs
{'units': 'Celsius', 'description': 'Temperature data'}

Returns:

assigned – A new object with the new attrs in addition to the existing data.

Return type:

same type as caller

See also

Dataset.assign

assign_coords()

Assign new coordinates to this object.

Returns a new object with all the original data in addition to the new coordinates.

Parameters:

• coords (mapping of dim to coord, optional) – A mapping whose keys are the names of the coordinates and values are the coordinates to assign. The mapping will generally be a dict or Coordinates.

  • If a value is a standard data value — for example, a DataArray, scalar, or array — the data is simply assigned as a coordinate.

  • If a value is callable, it is called with this object as the only parameter, and the return value is used as new coordinate variables.

  • A coordinate can also be defined and attached to an existing dimension using a tuple with the first element the dimension name and the second element the values for this new coordinate.

• **coords_kwargs (optional) – The keyword arguments form of coords. One of coords or coords_kwargs must be provided.

Returns:

assigned – A new object with the new coordinates in addition to the existing data.

Return type:

same type as caller

Examples

Convert DataArray longitude coordinates from 0-359 to -180-179:

>>> da = xr.DataArray(
...     np.random.rand(4),
...     coords=[np.array([358, 359, 0, 1])],
...     dims="lon",
... )
>>> da
<xarray.DataArray (lon: 4)> Size: 32B
array([0.5488135 , 0.71518937, 0.60276338, 0.54488318])
Coordinates:
  * lon      (lon) int64 32B 358 359 0 1
>>> da.assign_coords(lon=(((da.lon + 180) % 360) - 180))
<xarray.DataArray (lon: 4)> Size: 32B
array([0.5488135 , 0.71518937, 0.60276338, 0.54488318])
Coordinates:
  * lon      (lon) int64 32B -2 -1 0 1

The function also accepts dictionary arguments:

>>> da.assign_coords({"lon": (((da.lon + 180) % 360) - 180)})
<xarray.DataArray (lon: 4)> Size: 32B
array([0.5488135 , 0.71518937, 0.60276338, 0.54488318])
Coordinates:
  * lon      (lon) int64 32B -2 -1 0 1

New coordinate can also be attached to an existing dimension:

>>> lon_2 = np.array([300, 289, 0, 1])
>>> da.assign_coords(lon_2=("lon", lon_2))
<xarray.DataArray (lon: 4)> Size: 32B
array([0.5488135 , 0.71518937, 0.60276338, 0.54488318])
Coordinates:
  * lon      (lon) int64 32B 358 359 0 1
    lon_2    (lon) int64 32B 300 289 0 1

Note that the same result can also be obtained with a dict e.g.

>>> _ = da.assign_coords({"lon_2": ("lon", lon_2)})

Note the same method applies to Dataset objects.

Convert Dataset longitude coordinates from 0-359 to -180-179:

>>> temperature = np.linspace(20, 32, num=16).reshape(2, 2, 4)
>>> precipitation = 2 * np.identity(4).reshape(2, 2, 4)
>>> ds = xr.Dataset(
...     data_vars=dict(
...         temperature=(["x", "y", "time"], temperature),
...         precipitation=(["x", "y", "time"], precipitation),
...     ),
...     coords=dict(
...         lon=(["x", "y"], [[260.17, 260.68], [260.21, 260.77]]),
...         lat=(["x", "y"], [[42.25, 42.21], [42.63, 42.59]]),
...         time=pd.date_range("2014-09-06", periods=4),
...         reference_time=pd.Timestamp("2014-09-05"),
...     ),
...     attrs=dict(description="Weather-related data"),
... )
>>> ds
<xarray.Dataset> Size: 360B
Dimensions:         (x: 2, y: 2, time: 4)
Coordinates:
    lon             (x, y) float64 32B 260.2 260.7 260.2 260.8
    lat             (x, y) float64 32B 42.25 42.21 42.63 42.59
  * time            (time) datetime64[ns] 32B 2014-09-06 ... 2014-09-09
    reference_time  datetime64[ns] 8B 2014-09-05
Dimensions without coordinates: x, y
Data variables:
    temperature     (x, y, time) float64 128B 20.0 20.8 21.6 ... 30.4 31.2 32.0
    precipitation   (x, y, time) float64 128B 2.0 0.0 0.0 0.0 ... 0.0 0.0 2.0
:ivar description: Weather-related data

>>> ds.assign_coords(lon=(((ds.lon + 180) % 360) - 180))
<xarray.Dataset> Size: 360B
Dimensions:         (x: 2, y: 2, time: 4)
Coordinates:
    lon             (x, y) float64 32B -99.83 -99.32 -99.79 -99.23
    lat             (x, y) float64 32B 42.25 42.21 42.63 42.59
  * time            (time) datetime64[ns] 32B 2014-09-06 ... 2014-09-09
    reference_time  datetime64[ns] 8B 2014-09-05
Dimensions without coordinates: x, y
Data variables:
    temperature     (x, y, time) float64 128B 20.0 20.8 21.6 ... 30.4 31.2 32.0
    precipitation   (x, y, time) float64 128B 2.0 0.0 0.0 0.0 ... 0.0 0.0 2.0
:ivar description: Weather-related data

See also

Dataset.assign, Dataset.swap_dims, Dataset.set_coords

astype()

Copy of the xarray object, with data cast to a specified type. Leaves coordinate dtype unchanged.

Parameters:

• dtype (str or dtype) – Typecode or data-type to which the array is cast.

• order ({‘C’, ‘F’, ‘A’, ‘K’}, optional) – Controls the memory layout order of the result. ‘C’ means C order, ‘F’ means Fortran order, ‘A’ means ‘F’ order if all the arrays are Fortran contiguous, ‘C’ order otherwise, and ‘K’ means as close to the order the array elements appear in memory as possible.

• casting ({‘no’, ‘equiv’, ‘safe’, ‘same_kind’, ‘unsafe’}, optional) – Controls what kind of data casting may occur.

  • ‘no’ means the data types should not be cast at all.

  • ‘equiv’ means only byte-order changes are allowed.

  • ‘safe’ means only casts which can preserve values are allowed.

  • ‘same_kind’ means only safe casts or casts within a kind, like float64 to float32, are allowed.

  • ‘unsafe’ means any data conversions may be done.

• subok (bool, optional) – If True, then sub-classes will be passed-through, otherwise the returned array will be forced to be a base-class array.

• copy (bool, optional) – By default, astype always returns a newly allocated array. If this is set to False and the dtype requirement is satisfied, the input array is returned instead of a copy.

• keep_attrs (bool, optional) – By default, astype keeps attributes. Set to False to remove attributes in the returned object.

Returns:

out – New object with data cast to the specified type.

Return type:

same as object

Notes

The order, casting, subok and copy arguments are only passed through to the astype method of the underlying array when a value different than None is supplied. Make sure to only supply these arguments if the underlying array class supports them.

See also

numpy.ndarray.astype, dask.array.Array.astype, sparse.COO.astype

property attrs

Dictionary storing arbitrary metadata with this array.

bfill()

Fill NaN values by propagating values backward

Requires bottleneck.

Parameters:

• dim (str) – Specifies the dimension along which to propagate values when filling.

• limit (int or None, default: None) – The maximum number of consecutive NaN values to backward fill. In other words, if there is a gap with more than this number of consecutive NaNs, it will only be partially filled. Must be greater than 0 or None for no limit. Must be None or greater than or equal to axis length if filling along chunked axes (dimensions).

Returns:

filled

Return type:

DataArray

Examples

>>> temperature = np.array(
...     [
...         [0, 1, 3],
...         [0, np.nan, 5],
...         [5, np.nan, np.nan],
...         [3, np.nan, np.nan],
...         [np.nan, 2, 0],
...     ]
... )
>>> da = xr.DataArray(
...     data=temperature,
...     dims=["Y", "X"],
...     coords=dict(
...         lat=("Y", np.array([-20.0, -20.25, -20.50, -20.75, -21.0])),
...         lon=("X", np.array([10.0, 10.25, 10.5])),
...     ),
... )
>>> da
<xarray.DataArray (Y: 5, X: 3)> Size: 120B
array([[ 0.,  1.,  3.],
       [ 0., nan,  5.],
       [ 5., nan, nan],
       [ 3., nan, nan],
       [nan,  2.,  0.]])
Coordinates:
    lat      (Y) float64 40B -20.0 -20.25 -20.5 -20.75 -21.0
    lon      (X) float64 24B 10.0 10.25 10.5
Dimensions without coordinates: Y, X

Fill all NaN values:

>>> da.bfill(dim="Y", limit=None)
<xarray.DataArray (Y: 5, X: 3)> Size: 120B
array([[ 0.,  1.,  3.],
       [ 0.,  2.,  5.],
       [ 5.,  2.,  0.],
       [ 3.,  2.,  0.],
       [nan,  2.,  0.]])
Coordinates:
    lat      (Y) float64 40B -20.0 -20.25 -20.5 -20.75 -21.0
    lon      (X) float64 24B 10.0 10.25 10.5
Dimensions without coordinates: Y, X

Fill only the first of consecutive NaN values:

>>> da.bfill(dim="Y", limit=1)
<xarray.DataArray (Y: 5, X: 3)> Size: 120B
array([[ 0.,  1.,  3.],
       [ 0., nan,  5.],
       [ 5., nan, nan],
       [ 3.,  2.,  0.],
       [nan,  2.,  0.]])
Coordinates:
    lat      (Y) float64 40B -20.0 -20.25 -20.5 -20.75 -21.0
    lon      (X) float64 24B 10.0 10.25 10.5
Dimensions without coordinates: Y, X

broadcast_equals()

Two DataArrays are broadcast equal if they are equal after broadcasting them against each other such that they have the same dimensions.

Parameters:

other (DataArray) – DataArray to compare to.

Returns:

equal – True if the two DataArrays are broadcast equal.

Return type:

bool

See also

DataArray.equals, DataArray.identical

Examples

>>> a = xr.DataArray([1, 2], dims="X")
>>> b = xr.DataArray([[1, 1], [2, 2]], dims=["X", "Y"])
>>> a
<xarray.DataArray (X: 2)> Size: 16B
array([1, 2])
Dimensions without coordinates: X
>>> b
<xarray.DataArray (X: 2, Y: 2)> Size: 32B
array([[1, 1],
       [2, 2]])
Dimensions without coordinates: X, Y

.equals returns True if two DataArrays have the same values, dimensions, and coordinates. .broadcast_equals returns True if the results of broadcasting two DataArrays against each other have the same values, dimensions, and coordinates.

>>> a.equals(b)
False
>>> a2, b2 = xr.broadcast(a, b)
>>> a2.equals(b2)
True
>>> a.broadcast_equals(b)
True

broadcast_like()

Broadcast this DataArray against another Dataset or DataArray.

This is equivalent to xr.broadcast(other, self)[1]

xarray objects are broadcast against each other in arithmetic operations, so this method is not be necessary for most uses.

If no change is needed, the input data is returned to the output without being copied.

If new coords are added by the broadcast, their values are NaN filled.

Parameters:

• other (Dataset or DataArray) – Object against which to broadcast this array.

• exclude (iterable of Hashable, optional) – Dimensions that must not be broadcasted

Returns:

new_da – The caller broadcasted against other.

Return type:

DataArray

Examples

>>> arr1 = xr.DataArray(
...     np.random.randn(2, 3),
...     dims=("x", "y"),
...     coords={"x": ["a", "b"], "y": ["a", "b", "c"]},
... )
>>> arr2 = xr.DataArray(
...     np.random.randn(3, 2),
...     dims=("x", "y"),
...     coords={"x": ["a", "b", "c"], "y": ["a", "b"]},
... )
>>> arr1
<xarray.DataArray (x: 2, y: 3)> Size: 48B
array([[ 1.76405235,  0.40015721,  0.97873798],
       [ 2.2408932 ,  1.86755799, -0.97727788]])
Coordinates:
  * x        (x) <U1 8B 'a' 'b'
  * y        (y) <U1 12B 'a' 'b' 'c'
>>> arr2
<xarray.DataArray (x: 3, y: 2)> Size: 48B
array([[ 0.95008842, -0.15135721],
       [-0.10321885,  0.4105985 ],
       [ 0.14404357,  1.45427351]])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'
  * y        (y) <U1 8B 'a' 'b'
>>> arr1.broadcast_like(arr2)
<xarray.DataArray (x: 3, y: 3)> Size: 72B
array([[ 1.76405235,  0.40015721,  0.97873798],
       [ 2.2408932 ,  1.86755799, -0.97727788],
       [        nan,         nan,         nan]])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'
  * y        (y) <U1 12B 'a' 'b' 'c'

chunk()

Coerce this array’s data into a dask arrays with the given chunks.

If this variable is a non-dask array, it will be converted to dask array. If it’s a dask array, it will be rechunked to the given chunk sizes.

If neither chunks is not provided for one or more dimensions, chunk sizes along that dimension will not be updated; non-dask arrays will be converted into dask arrays with a single block.

Along datetime-like dimensions, a pandas frequency string is also accepted.

Parameters:

• chunks (int, “auto”, tuple of int or mapping of hashable to int or a pandas frequency string, optional) – Chunk sizes along each dimension, e.g., 5, "auto", (5, 5) or {"x": 5, "y": 5} or {"x": 5, "time": "YE"}.

• name_prefix (str, optional) – Prefix for the name of the new dask array.

• token (str, optional) – Token uniquely identifying this array.

• lock (bool, default: False) – Passed on to dask.array.from_array(), if the array is not already as dask array.

• inline_array (bool, default: False) – Passed on to dask.array.from_array(), if the array is not already as dask array.

• chunked_array_type (str, optional) – Which chunked array type to coerce the underlying data array to. Defaults to ‘dask’ if installed, else whatever is registered via the ChunkManagerEntryPoint system. Experimental API that should not be relied upon.

• from_array_kwargs (dict, optional) – Additional keyword arguments passed on to the ChunkManagerEntrypoint.from_array method used to create chunked arrays, via whichever chunk manager is specified through the chunked_array_type kwarg. For example, with dask as the default chunked array type, this method would pass additional kwargs to dask.array.from_array(). Experimental API that should not be relied upon.

• **chunks_kwargs ({dim: chunks, …}, optional) – The keyword arguments form of chunks. One of chunks or chunks_kwargs must be provided.

Returns:

chunked

Return type:

xarray.DataArray

See also

DataArray.chunks, DataArray.chunksizes, xarray.unify_chunks, dask.array.from_array

property chunks

Tuple of block lengths for this dataarray’s data, in order of dimensions, or None if the underlying data is not a dask array.

See also

DataArray.chunk, DataArray.chunksizes, xarray.unify_chunks

property chunksizes

Mapping from dimension names to block lengths for this dataarray’s data.

If this dataarray does not contain chunked arrays, the mapping will be empty.

Cannot be modified directly, but can be modified by calling .chunk().

Differs from DataArray.chunks because it returns a mapping of dimensions to chunk shapes instead of a tuple of chunk shapes.

See also

DataArray.chunk, DataArray.chunks, xarray.unify_chunks

clip()

Return an array whose values are limited to [min, max]. At least one of max or min must be given.

Parameters:

• min (None or Hashable, optional) – Minimum value. If None, no lower clipping is performed.

• max (None or Hashable, optional) – Maximum value. If None, no upper clipping is performed.

• keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

Returns:

clipped – This object, but with with values < min are replaced with min, and those > max with max.

Return type:

same type as caller

See also

numpy.clip

equivalent function

close()

Release any resources linked to this object.

combine_first()

Combine two DataArray objects, with union of coordinates.

This operation follows the normal broadcasting and alignment rules of join='outer'. Default to non-null values of array calling the method. Use np.nan to fill in vacant cells after alignment.

Parameters:

other (DataArray) – Used to fill all matching missing values in this array.

Return type:

DataArray

compute()

Trigger loading data into memory and return a new dataarray.

Data will be computed and/or loaded from disk or a remote source.

Unlike .load, the original dataarray is left unaltered.

Normally, it should not be necessary to call this method in user code, because all xarray functions should either work on deferred data or load data automatically. However, this method can be necessary when working with many file objects on disk.

Parameters:

**kwargs (dict) – Additional keyword arguments passed on to dask.compute.

Returns:

object – New object with the data and all coordinates as in-memory arrays.

Return type:

DataArray

See also

dask.compute, DataArray.load, DataArray.load_async, Dataset.compute, Variable.compute

conj()

Complex-conjugate all elements.

Refer to numpy.conjugate for full documentation.

See also

numpy.conjugate

equivalent function

conjugate()

Complex-conjugate all elements.

Refer to numpy.conjugate for full documentation.

See also

numpy.conjugate

equivalent function

convert_calendar()

Convert the DataArray to another calendar.

Only converts the individual timestamps, does not modify any data except in dropping invalid/surplus dates or inserting missing dates.

If the source and target calendars are either no_leap, all_leap or a standard type, only the type of the time array is modified. When converting to a leap year from a non-leap year, the 29th of February is removed from the array. In the other direction the 29th of February will be missing in the output, unless missing is specified, in which case that value is inserted.

For conversions involving 360_day calendars, see Notes.

This method is safe to use with sub-daily data as it doesn’t touch the time part of the timestamps.

Parameters:

• calendar (str) – The target calendar name.

• dim (str) – Name of the time coordinate.

• align_on ({None, ‘date’, ‘year’}) – Must be specified when either source or target is a 360_day calendar, ignored otherwise. See Notes.

• missing (Optional[any]) – By default, i.e. if the value is None, this method will simply attempt to convert the dates in the source calendar to the same dates in the target calendar, and drop any of those that are not possible to represent. If a value is provided, a new time coordinate will be created in the target calendar with the same frequency as the original time coordinate; for any dates that are not present in the source, the data will be filled with this value. Note that using this mode requires that the source data have an inferable frequency; for more information see xarray.infer_freq(). For certain frequency, source, and target calendar combinations, this could result in many missing values, see notes.

• use_cftime (boolean, optional) – Whether to use cftime objects in the output, only used if calendar is one of {“proleptic_gregorian”, “gregorian” or “standard”}. If True, the new time axis uses cftime objects. If None (default), it uses numpy.datetime64 values if the date range permits it, and cftime.datetime objects if not. If False, it uses numpy.datetime64 or fails.

Returns:

Copy of the dataarray with the time coordinate converted to the target calendar. If ‘missing’ was None (default), invalid dates in the new calendar are dropped, but missing dates are not inserted. If missing was given, the new data is reindexed to have a time axis with the same frequency as the source, but in the new calendar; any missing datapoints are filled with missing.

Return type:

DataArray

Notes

Passing a value to missing is only usable if the source’s time coordinate as an inferable frequencies (see infer_freq()) and is only appropriate if the target coordinate, generated from this frequency, has dates equivalent to the source. It is usually not appropriate to use this mode with:

• Period-end frequencies : ‘A’, ‘Y’, ‘Q’ or ‘M’, in opposition to ‘AS’ ‘YS’, ‘QS’ and ‘MS’

• Sub-monthly frequencies that do not divide a day evenly‘W’, ‘nD’ where N != 1

  or ‘mH’ where 24 % m != 0).

If one of the source or target calendars is “360_day”, align_on must be specified and two options are offered.

• “year”

  The dates are translated according to their relative position in the year, ignoring their original month and day information, meaning that the missing/surplus days are added/removed at regular intervals.

  From a 360_day to a standard calendar, the output will be missing the following dates (day of year in parentheses):

  To a leap year:

  January 31st (31), March 31st (91), June 1st (153), July 31st (213), September 31st (275) and November 30th (335).

  To a non-leap year:

  February 6th (36), April 19th (109), July 2nd (183), September 12th (255), November 25th (329).

  From a standard calendar to a “360_day”, the following dates in the source array will be dropped:

  From a leap year:

  January 31st (31), April 1st (92), June 1st (153), August 1st (214), September 31st (275), December 1st (336)

  From a non-leap year:

  February 6th (37), April 20th (110), July 2nd (183), September 13th (256), November 25th (329)

  This option is best used on daily and subdaily data.

• “date”

  The month/day information is conserved and invalid dates are dropped from the output. This means that when converting from a “360_day” to a standard calendar, all 31st (Jan, March, May, July, August, October and December) will be missing as there is no equivalent dates in the “360_day” calendar and the 29th (on non-leap years) and 30th of February will be dropped as there are no equivalent dates in a standard calendar.

  This option is best used with data on a frequency coarser than daily.

property coords

Mapping of DataArray objects corresponding to coordinate variables.

See also

Coordinates

copy()

Returns a copy of this array.

If deep=True, a deep copy is made of the data array. Otherwise, a shallow copy is made, and the returned data array’s values are a new view of this data array’s values.

Use data to create a new object with the same structure as original but entirely new data.

Parameters:

• deep (bool, optional) – Whether the data array and its coordinates are loaded into memory and copied onto the new object. Default is True.

• data (array_like, optional) – Data to use in the new object. Must have same shape as original. When data is used, deep is ignored for all data variables, and only used for coords.

Returns:

copy – New object with dimensions, attributes, coordinates, name, encoding, and optionally data copied from original.

Return type:

DataArray

Examples

Shallow versus deep copy

>>> array = xr.DataArray([1, 2, 3], dims="x", coords={"x": ["a", "b", "c"]})
>>> array.copy()
<xarray.DataArray (x: 3)> Size: 24B
array([1, 2, 3])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'
>>> array_0 = array.copy(deep=False)
>>> array_0[0] = 7
>>> array_0
<xarray.DataArray (x: 3)> Size: 24B
array([7, 2, 3])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'
>>> array
<xarray.DataArray (x: 3)> Size: 24B
array([7, 2, 3])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'

Changing the data using the data argument maintains the structure of the original object, but with the new data. Original object is unaffected.

>>> array.copy(data=[0.1, 0.2, 0.3])
<xarray.DataArray (x: 3)> Size: 24B
array([0.1, 0.2, 0.3])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'
>>> array
<xarray.DataArray (x: 3)> Size: 24B
array([7, 2, 3])
Coordinates:
  * x        (x) <U1 12B 'a' 'b' 'c'

See also

pandas.DataFrame.copy

count()

Reduce this DataArray’s data by applying count along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply count. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating count on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with count applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

pandas.DataFrame.count, dask.dataframe.DataFrame.count, Dataset.count

Aggregation

User guide on reduction or aggregation operations.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.count()
<xarray.DataArray ()> Size: 8B
array(5)

cumprod()

Reduce this DataArray’s data by applying cumprod along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply cumprod. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating cumprod on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with cumprod applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.cumprod, dask.array.cumprod, Dataset.cumprod, DataArray.cumulative

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Note that the methods on the cumulative method are more performant (with numbagg installed) and better supported. cumsum and cumprod may be deprecated in the future.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.cumprod()
<xarray.DataArray (time: 6)> Size: 48B
array([1., 2., 6., 0., 0., 0.])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

Use skipna to control whether NaNs are ignored.

>>> da.cumprod(skipna=False)
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  6.,  0.,  0., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

cumsum()

Reduce this DataArray’s data by applying cumsum along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply cumsum. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating cumsum on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with cumsum applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.cumsum, dask.array.cumsum, Dataset.cumsum, DataArray.cumulative

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Note that the methods on the cumulative method are more performant (with numbagg installed) and better supported. cumsum and cumprod may be deprecated in the future.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.cumsum()
<xarray.DataArray (time: 6)> Size: 48B
array([1., 3., 6., 6., 8., 8.])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

Use skipna to control whether NaNs are ignored.

>>> da.cumsum(skipna=False)
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  3.,  6.,  6.,  8., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

cumulative()

Accumulating object for DataArrays.

Parameters:

• dims (iterable of hashable) – The name(s) of the dimensions to create the cumulative window along

• min_periods (int, default: 1) – Minimum number of observations in window required to have a value (otherwise result is NA). The default is 1 (note this is different from Rolling, whose default is the size of the window).

Return type:

computation.rolling.DataArrayRolling

Examples

Create rolling seasonal average of monthly data e.g. DJF, JFM, …, SON:

>>> da = xr.DataArray(
...     np.linspace(0, 11, num=12),
...     coords=[
...         pd.date_range(
...             "1999-12-15",
...             periods=12,
...             freq=pd.DateOffset(months=1),
...         )
...     ],
...     dims="time",
... )

>>> da
<xarray.DataArray (time: 12)> Size: 96B
array([ 0.,  1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10., 11.])
Coordinates:
  * time     (time) datetime64[ns] 96B 1999-12-15 2000-01-15 ... 2000-11-15

>>> da.cumulative("time").sum()
<xarray.DataArray (time: 12)> Size: 96B
array([ 0.,  1.,  3.,  6., 10., 15., 21., 28., 36., 45., 55., 66.])
Coordinates:
  * time     (time) datetime64[ns] 96B 1999-12-15 2000-01-15 ... 2000-11-15

See also

DataArray.rolling, Dataset.cumulative, computation.rolling.DataArrayRolling

cumulative_integrate()

Integrate cumulatively along the given coordinate using the trapezoidal rule.

Note

This feature is limited to simple cartesian geometry, i.e. coord must be one dimensional.

The first entry of the cumulative integral is always 0, in order to keep the length of the dimension unchanged between input and output.

Parameters:

• coord (Hashable, or sequence of Hashable) – Coordinate(s) used for the integration.

• datetime_unit ({‘W’, ‘D’, ‘h’, ‘m’, ‘s’, ‘ms’, ‘us’, ‘ns’, ‘ps’, ‘fs’, ‘as’, None}, optional) – Specify the unit if a datetime coordinate is used.

Returns:

integrated

Return type:

DataArray

See also

Dataset.cumulative_integrate

scipy.integrate.cumulative_trapezoid

corresponding scipy function

Examples

>>> da = xr.DataArray(
...     np.arange(12).reshape(4, 3),
...     dims=["x", "y"],
...     coords={"x": [0, 0.1, 1.1, 1.2]},
... )
>>> da
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) float64 32B 0.0 0.1 1.1 1.2
Dimensions without coordinates: y
>>>
>>> da.cumulative_integrate("x")
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[0.  , 0.  , 0.  ],
       [0.15, 0.25, 0.35],
       [4.65, 5.75, 6.85],
       [5.4 , 6.6 , 7.8 ]])
Coordinates:
  * x        (x) float64 32B 0.0 0.1 1.1 1.2
Dimensions without coordinates: y

curvefit()

Curve fitting optimization for arbitrary functions.

Wraps scipy.optimize.curve_fit() with apply_ufunc().

Parameters:

• coords (Hashable, DataArray, or sequence of DataArray or Hashable) – Independent coordinate(s) over which to perform the curve fitting. Must share at least one dimension with the calling object. When fitting multi-dimensional functions, supply coords as a sequence in the same order as arguments in func. To fit along existing dimensions of the calling object, coords can also be specified as a str or sequence of strs.

• func (callable) – User specified function in the form f(x, *params) which returns a numpy array of length len(x). params are the fittable parameters which are optimized by scipy curve_fit. x can also be specified as a sequence containing multiple coordinates, e.g. f((x0, x1), *params).

• reduce_dims (str, Iterable of Hashable or None, optional) – Additional dimension(s) over which to aggregate while fitting. For example, calling ds.curvefit(coords=’time’, reduce_dims=[‘lat’, ‘lon’], …) will aggregate all lat and lon points and fit the specified function along the time dimension.

• skipna (bool, default: True) – Whether to skip missing values when fitting. Default is True.

• p0 (dict-like or None, optional) – Optional dictionary of parameter names to initial guesses passed to the curve_fit p0 arg. If the values are DataArrays, they will be appropriately broadcast to the coordinates of the array. If none or only some parameters are passed, the rest will be assigned initial values following the default scipy behavior.

• bounds (dict-like, optional) – Optional dictionary of parameter names to tuples of bounding values passed to the curve_fit bounds arg. If any of the bounds are DataArrays, they will be appropriately broadcast to the coordinates of the array. If none or only some parameters are passed, the rest will be unbounded following the default scipy behavior.

• param_names (sequence of Hashable or None, optional) – Sequence of names for the fittable parameters of func. If not supplied, this will be automatically determined by arguments of func. param_names should be manually supplied when fitting a function that takes a variable number of parameters.

• errors ({“raise”, “ignore”}, default: “raise”) – If ‘raise’, any errors from the scipy.optimize_curve_fit optimization will raise an exception. If ‘ignore’, the coefficients and covariances for the coordinates where the fitting failed will be NaN.

• **kwargs (optional) – Additional keyword arguments to passed to scipy curve_fit.

Returns:

curvefit_results – A single dataset which contains:

[var]_curvefit_coefficients

The coefficients of the best fit.

[var]_curvefit_covariance

The covariance matrix of the coefficient estimates.

Return type:

Dataset

Examples

Generate some exponentially decaying data, where the decay constant and amplitude are different for different values of the coordinate x:

>>> rng = np.random.default_rng(seed=0)
>>> def exp_decay(t, time_constant, amplitude):
...     return np.exp(-t / time_constant) * amplitude
...
>>> t = np.arange(11)
>>> da = xr.DataArray(
...     np.stack(
...         [
...             exp_decay(t, 1, 0.1),
...             exp_decay(t, 2, 0.2),
...             exp_decay(t, 3, 0.3),
...         ]
...     )
...     + rng.normal(size=(3, t.size)) * 0.01,
...     coords={"x": [0, 1, 2], "time": t},
... )
>>> da
<xarray.DataArray (x: 3, time: 11)> Size: 264B
array([[ 0.1012573 ,  0.0354669 ,  0.01993775,  0.00602771, -0.00352513,
         0.00428975,  0.01328788,  0.009562  , -0.00700381, -0.01264187,
        -0.0062282 ],
       [ 0.20041326,  0.09805582,  0.07138797,  0.03216692,  0.01974438,
         0.01097441,  0.00679441,  0.01015578,  0.01408826,  0.00093645,
         0.01501222],
       [ 0.29334805,  0.21847449,  0.16305984,  0.11130396,  0.07164415,
         0.04744543,  0.03602333,  0.03129354,  0.01074885,  0.01284436,
         0.00910995]])
Coordinates:
  * x        (x) int64 24B 0 1 2
  * time     (time) int64 88B 0 1 2 3 4 5 6 7 8 9 10

Fit the exponential decay function to the data along the time dimension:

>>> fit_result = da.curvefit("time", exp_decay)
>>> fit_result["curvefit_coefficients"].sel(
...     param="time_constant"
... )
<xarray.DataArray 'curvefit_coefficients' (x: 3)> Size: 24B
array([1.05692036, 1.73549638, 2.94215771])
Coordinates:
  * x        (x) int64 24B 0 1 2
    param    <U13 52B 'time_constant'
>>> fit_result["curvefit_coefficients"].sel(param="amplitude")
<xarray.DataArray 'curvefit_coefficients' (x: 3)> Size: 24B
array([0.1005489 , 0.19631423, 0.30003579])
Coordinates:
  * x        (x) int64 24B 0 1 2
    param    <U13 52B 'amplitude'

An initial guess can also be given with the p0 arg (although it does not make much of a difference in this simple example). To have a different guess for different coordinate points, the guess can be a DataArray. Here we use the same initial guess for the amplitude but different guesses for the time constant:

>>> fit_result = da.curvefit(
...     "time",
...     exp_decay,
...     p0={
...         "amplitude": 0.2,
...         "time_constant": xr.DataArray([1, 2, 3], coords=[da.x]),
...     },
... )
>>> fit_result["curvefit_coefficients"].sel(param="time_constant")
<xarray.DataArray 'curvefit_coefficients' (x: 3)> Size: 24B
array([1.0569213 , 1.73550052, 2.94215733])
Coordinates:
  * x        (x) int64 24B 0 1 2
    param    <U13 52B 'time_constant'
>>> fit_result["curvefit_coefficients"].sel(param="amplitude")
<xarray.DataArray 'curvefit_coefficients' (x: 3)> Size: 24B
array([0.10054889, 0.1963141 , 0.3000358 ])
Coordinates:
  * x        (x) int64 24B 0 1 2
    param    <U13 52B 'amplitude'

See also

DataArray.polyfit, scipy.optimize.curve_fit

xarray.DataArray.xlm.modelfit

External method from xarray-lmfit with more curve fitting functionality.

property data

The DataArray’s data as an array. The underlying array type (e.g. dask, sparse, pint) is preserved.

See also

DataArray.to_numpy, DataArray.as_numpy, DataArray.values

diff()

Calculate the n-th order discrete difference along given axis.

Parameters:

• dim (Hashable) – Dimension over which to calculate the finite difference.

• n (int, default: 1) – The number of times values are differenced.

• label ({“upper”, “lower”}, default: “upper”) – The new coordinate in dimension dim will have the values of either the minuend’s or subtrahend’s coordinate for values ‘upper’ and ‘lower’, respectively.

Returns:

difference – The n-th order finite difference of this object.

Return type:

DataArray

Notes

n matches numpy’s behavior and is different from pandas’ first argument named periods.

Examples

>>> arr = xr.DataArray([5, 5, 6, 6], [[1, 2, 3, 4]], ["x"])
>>> arr.diff("x")
<xarray.DataArray (x: 3)> Size: 24B
array([0, 1, 0])
Coordinates:
  * x        (x) int64 24B 2 3 4
>>> arr.diff("x", 2)
<xarray.DataArray (x: 2)> Size: 16B
array([ 1, -1])
Coordinates:
  * x        (x) int64 16B 3 4

See also

DataArray.differentiate

differentiate()

Differentiate the array with the second order accurate central differences.

Note

This feature is limited to simple cartesian geometry, i.e. coord must be one dimensional.

Parameters:

• coord (Hashable) – The coordinate to be used to compute the gradient.

• edge_order ({1, 2}, default: 1) – N-th order accurate differences at the boundaries.

• datetime_unit ({“W”, “D”, “h”, “m”, “s”, “ms”, “us”, “ns”, “ps”, “fs”, “as”, None}, optional) – Unit to compute gradient. Only valid for datetime coordinate. “Y” and “M” are not available as datetime_unit.

Returns:

differentiated

Return type:

DataArray

See also

numpy.gradient

corresponding numpy function

Examples

>>> da = xr.DataArray(
...     np.arange(12).reshape(4, 3),
...     dims=["x", "y"],
...     coords={"x": [0, 0.1, 1.1, 1.2]},
... )
>>> da
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) float64 32B 0.0 0.1 1.1 1.2
Dimensions without coordinates: y
>>>
>>> da.differentiate("x")
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[30.        , 30.        , 30.        ],
       [27.54545455, 27.54545455, 27.54545455],
       [27.54545455, 27.54545455, 27.54545455],
       [30.        , 30.        , 30.        ]])
Coordinates:
  * x        (x) float64 32B 0.0 0.1 1.1 1.2
Dimensions without coordinates: y

property dims

Tuple of dimension names associated with this array.

Note that the type of this property is inconsistent with Dataset.dims. See Dataset.sizes and DataArray.sizes for consistently named properties.

See also

DataArray.sizes, Dataset.dims

dot()

Perform dot product of two DataArrays along their shared dims.

Equivalent to taking taking tensordot over all shared dims.

Parameters:

• other (DataArray) – The other array with which the dot product is performed.

• dim (…, str, Iterable of Hashable or None, optional) – Which dimensions to sum over. Ellipsis (…) sums over all dimensions. If not specified, then all the common dimensions are summed over.

Returns:

result – Array resulting from the dot product over all shared dimensions.

Return type:

DataArray

See also

dot, numpy.tensordot

Examples

>>> da_vals = np.arange(6 * 5 * 4).reshape((6, 5, 4))
>>> da = xr.DataArray(da_vals, dims=["x", "y", "z"])
>>> dm_vals = np.arange(4)
>>> dm = xr.DataArray(dm_vals, dims=["z"])

>>> dm.dims
('z',)

>>> da.dims
('x', 'y', 'z')

>>> dot_result = da.dot(dm)
>>> dot_result.dims
('x', 'y')

drop()

Backward compatible method based on drop_vars and drop_sel

Using either drop_vars or drop_sel is encouraged

See also

DataArray.drop_vars, DataArray.drop_sel

drop_attrs()

Removes all attributes from the DataArray.

Parameters:

deep (bool, default True) – Removes attributes from coordinates.

Return type:

DataArray

drop_duplicates()

Returns a new DataArray with duplicate dimension values removed.

Parameters:

• dim (dimension label or labels) – Pass … to drop duplicates along all dimensions.

• keep ({“first”, “last”, False}, default: “first”) – Determines which duplicates (if any) to keep.

  • "first" : Drop duplicates except for the first occurrence.

  • "last" : Drop duplicates except for the last occurrence.

  • False : Drop all duplicates.

Return type:

DataArray

See also

Dataset.drop_duplicates

Examples

>>> da = xr.DataArray(
...     np.arange(25).reshape(5, 5),
...     dims=("x", "y"),
...     coords={"x": np.array([0, 0, 1, 2, 3]), "y": np.array([0, 1, 2, 3, 3])},
... )
>>> da
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Coordinates:
  * x        (x) int64 40B 0 0 1 2 3
  * y        (y) int64 40B 0 1 2 3 3

>>> da.drop_duplicates(dim="x")
<xarray.DataArray (x: 4, y: 5)> Size: 160B
array([[ 0,  1,  2,  3,  4],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Coordinates:
  * x        (x) int64 32B 0 1 2 3
  * y        (y) int64 40B 0 1 2 3 3

>>> da.drop_duplicates(dim="x", keep="last")
<xarray.DataArray (x: 4, y: 5)> Size: 160B
array([[ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Coordinates:
  * x        (x) int64 32B 0 1 2 3
  * y        (y) int64 40B 0 1 2 3 3

Drop all duplicate dimension values:

>>> da.drop_duplicates(dim=...)
<xarray.DataArray (x: 4, y: 4)> Size: 128B
array([[ 0,  1,  2,  3],
       [10, 11, 12, 13],
       [15, 16, 17, 18],
       [20, 21, 22, 23]])
Coordinates:
  * x        (x) int64 32B 0 1 2 3
  * y        (y) int64 32B 0 1 2 3

drop_encoding()

Return a new DataArray without encoding on the array or any attached coords.

drop_indexes()

Drop the indexes assigned to the given coordinates.

Parameters:

• coord_names (hashable or iterable of hashable) – Name(s) of the coordinate(s) for which to drop the index.

• errors ({“raise”, “ignore”}, default: “raise”) – If ‘raise’, raises a ValueError error if any of the coordinates passed have no index or are not in the dataset. If ‘ignore’, no error is raised.

Returns:

dropped – A new dataarray with dropped indexes.

Return type:

DataArray

drop_isel()

Drop index positions from this DataArray.

Parameters:

• indexers (mapping of Hashable to Any or None, default: None) – Index locations to drop

• **indexers_kwargs ({dim: position, …}, optional) – The keyword arguments form of dim and positions

Returns:

dropped

Return type:

DataArray

Raises:

IndexError –

Examples

>>> da = xr.DataArray(np.arange(25).reshape(5, 5), dims=("X", "Y"))
>>> da
<xarray.DataArray (X: 5, Y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Dimensions without coordinates: X, Y

>>> da.drop_isel(X=[0, 4], Y=2)
<xarray.DataArray (X: 3, Y: 4)> Size: 96B
array([[ 5,  6,  8,  9],
       [10, 11, 13, 14],
       [15, 16, 18, 19]])
Dimensions without coordinates: X, Y

>>> da.drop_isel({"X": 3, "Y": 3})
<xarray.DataArray (X: 4, Y: 4)> Size: 128B
array([[ 0,  1,  2,  4],
       [ 5,  6,  7,  9],
       [10, 11, 12, 14],
       [20, 21, 22, 24]])
Dimensions without coordinates: X, Y

drop_sel()

Drop index labels from this DataArray.

Parameters:

• labels (mapping of Hashable to Any) – Index labels to drop

• errors ({“raise”, “ignore”}, default: “raise”) – If ‘raise’, raises a ValueError error if any of the index labels passed are not in the dataset. If ‘ignore’, any given labels that are in the dataset are dropped and no error is raised.

• **labels_kwargs ({dim: label, …}, optional) – The keyword arguments form of dim and labels

Returns:

dropped

Return type:

DataArray

Examples

>>> da = xr.DataArray(
...     np.arange(25).reshape(5, 5),
...     coords={"x": np.arange(0, 9, 2), "y": np.arange(0, 13, 3)},
...     dims=("x", "y"),
... )
>>> da
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Coordinates:
  * x        (x) int64 40B 0 2 4 6 8
  * y        (y) int64 40B 0 3 6 9 12

>>> da.drop_sel(x=[0, 2], y=9)
<xarray.DataArray (x: 3, y: 4)> Size: 96B
array([[10, 11, 12, 14],
       [15, 16, 17, 19],
       [20, 21, 22, 24]])
Coordinates:
  * x        (x) int64 24B 4 6 8
  * y        (y) int64 32B 0 3 6 12

>>> da.drop_sel({"x": 6, "y": [0, 3]})
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 2,  3,  4],
       [ 7,  8,  9],
       [12, 13, 14],
       [22, 23, 24]])
Coordinates:
  * x        (x) int64 32B 0 2 4 8
  * y        (y) int64 24B 6 9 12

drop_vars()

Returns an array with dropped variables.

Parameters:

• names (Hashable or iterable of Hashable or Callable) – Name(s) of variables to drop. If a Callable, this object is passed as its only argument and its result is used.

• errors ({“raise”, “ignore”}, default: “raise”) – If ‘raise’, raises a ValueError error if any of the variable passed are not in the dataset. If ‘ignore’, any given names that are in the DataArray are dropped and no error is raised.

Returns:

dropped – New Dataset copied from self with variables removed.

Return type:

Dataset

Examples

>>> data = np.arange(12).reshape(4, 3)
>>> da = xr.DataArray(
...     data=data,
...     dims=["x", "y"],
...     coords={"x": [10, 20, 30, 40], "y": [70, 80, 90]},
... )
>>> da
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) int64 32B 10 20 30 40
  * y        (y) int64 24B 70 80 90

Removing a single variable:

>>> da.drop_vars("x")
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * y        (y) int64 24B 70 80 90
Dimensions without coordinates: x

Removing a list of variables:

>>> da.drop_vars(["x", "y"])
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Dimensions without coordinates: x, y

>>> da.drop_vars(lambda x: x.coords)
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Dimensions without coordinates: x, y

dropna()

Returns a new array with dropped labels for missing values along the provided dimension.

Parameters:

• dim (Hashable) – Dimension along which to drop missing values. Dropping along multiple dimensions simultaneously is not yet supported.

• how ({“any”, “all”}, default: “any”) –

  • any : if any NA values are present, drop that label

  • all : if all values are NA, drop that label

• thresh (int or None, default: None) – If supplied, require this many non-NA values.

Returns:

dropped

Return type:

DataArray

Examples

>>> temperature = [
...     [0, 4, 2, 9],
...     [np.nan, np.nan, np.nan, np.nan],
...     [np.nan, 4, 2, 0],
...     [3, 1, 0, 0],
... ]
>>> da = xr.DataArray(
...     data=temperature,
...     dims=["Y", "X"],
...     coords=dict(
...         lat=("Y", np.array([-20.0, -20.25, -20.50, -20.75])),
...         lon=("X", np.array([10.0, 10.25, 10.5, 10.75])),
...     ),
... )
>>> da
<xarray.DataArray (Y: 4, X: 4)> Size: 128B
array([[ 0.,  4.,  2.,  9.],
       [nan, nan, nan, nan],
       [nan,  4.,  2.,  0.],
       [ 3.,  1.,  0.,  0.]])
Coordinates:
    lat      (Y) float64 32B -20.0 -20.25 -20.5 -20.75
    lon      (X) float64 32B 10.0 10.25 10.5 10.75
Dimensions without coordinates: Y, X

>>> da.dropna(dim="Y", how="any")
<xarray.DataArray (Y: 2, X: 4)> Size: 64B
array([[0., 4., 2., 9.],
       [3., 1., 0., 0.]])
Coordinates:
    lat      (Y) float64 16B -20.0 -20.75
    lon      (X) float64 32B 10.0 10.25 10.5 10.75
Dimensions without coordinates: Y, X

Drop values only if all values along the dimension are NaN:

>>> da.dropna(dim="Y", how="all")
<xarray.DataArray (Y: 3, X: 4)> Size: 96B
array([[ 0.,  4.,  2.,  9.],
       [nan,  4.,  2.,  0.],
       [ 3.,  1.,  0.,  0.]])
Coordinates:
    lat      (Y) float64 24B -20.0 -20.5 -20.75
    lon      (X) float64 32B 10.0 10.25 10.5 10.75
Dimensions without coordinates: Y, X

dt

alias of CombinedDatetimelikeAccessor[DataArray]

property dtype

Data-type of the array’s elements.

See also

ndarray.dtype, numpy.dtype

property encoding

Dictionary of format-specific settings for how this array should be serialized.

equals()

True if two DataArrays have the same dimensions, coordinates and values; otherwise False.

DataArrays can still be equal (like pandas objects) if they have NaN values in the same locations.

This method is necessary because v1 == v2 for DataArray does element-wise comparisons (like numpy.ndarrays).

Parameters:

other (DataArray) – DataArray to compare to.

Returns:

equal – True if the two DataArrays are equal.

Return type:

bool

See also

DataArray.broadcast_equals, DataArray.identical

Examples

>>> a = xr.DataArray([1, 2, 3], dims="X")
>>> b = xr.DataArray([1, 2, 3], dims="X", attrs=dict(units="m"))
>>> c = xr.DataArray([1, 2, 3], dims="Y")
>>> d = xr.DataArray([3, 2, 1], dims="X")
>>> a
<xarray.DataArray (X: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: X
>>> b
<xarray.DataArray (X: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: X
:ivar units: m

>>> c
<xarray.DataArray (Y: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: Y
>>> d
<xarray.DataArray (X: 3)> Size: 24B
array([3, 2, 1])
Dimensions without coordinates: X

>>> a.equals(b)
True
>>> a.equals(c)
False
>>> a.equals(d)
False

expand_dims()

Return a new object with an additional axis (or axes) inserted at the corresponding position in the array shape. The new object is a view into the underlying array, not a copy.

If dim is already a scalar coordinate, it will be promoted to a 1D coordinate consisting of a single value.

The automatic creation of indexes to back new 1D coordinate variables controlled by the create_index_for_new_dim kwarg.

Parameters:

• dim (Hashable, sequence of Hashable, dict, or None, optional) – Dimensions to include on the new variable. If provided as str or sequence of str, then dimensions are inserted with length 1. If provided as a dict, then the keys are the new dimensions and the values are either integers (giving the length of the new dimensions) or sequence/ndarray (giving the coordinates of the new dimensions).

• axis (int, sequence of int, or None, default: None) – Axis position(s) where new axis is to be inserted (position(s) on the result array). If a sequence of integers is passed, multiple axes are inserted. In this case, dim arguments should be same length list. If axis=None is passed, all the axes will be inserted to the start of the result array.

• create_index_for_new_dim (bool, default: True) – Whether to create new PandasIndex objects when the object being expanded contains scalar variables with names in dim.

• **dim_kwargs (int or sequence or ndarray) – The keywords are arbitrary dimensions being inserted and the values are either the lengths of the new dims (if int is given), or their coordinates. Note, this is an alternative to passing a dict to the dim kwarg and will only be used if dim is None.

Returns:

expanded – This object, but with additional dimension(s).

Return type:

DataArray

See also

Dataset.expand_dims

Examples

>>> da = xr.DataArray(np.arange(5), dims=("x"))
>>> da
<xarray.DataArray (x: 5)> Size: 40B
array([0, 1, 2, 3, 4])
Dimensions without coordinates: x

Add new dimension of length 2:

>>> da.expand_dims(dim={"y": 2})
<xarray.DataArray (y: 2, x: 5)> Size: 80B
array([[0, 1, 2, 3, 4],
       [0, 1, 2, 3, 4]])
Dimensions without coordinates: y, x

>>> da.expand_dims(dim={"y": 2}, axis=1)
<xarray.DataArray (x: 5, y: 2)> Size: 80B
array([[0, 0],
       [1, 1],
       [2, 2],
       [3, 3],
       [4, 4]])
Dimensions without coordinates: x, y

Add a new dimension with coordinates from array:

>>> da.expand_dims(dim={"y": np.arange(5)}, axis=0)
<xarray.DataArray (y: 5, x: 5)> Size: 200B
array([[0, 1, 2, 3, 4],
       [0, 1, 2, 3, 4],
       [0, 1, 2, 3, 4],
       [0, 1, 2, 3, 4],
       [0, 1, 2, 3, 4]])
Coordinates:
  * y        (y) int64 40B 0 1 2 3 4
Dimensions without coordinates: x

ffill()

Fill NaN values by propagating values forward

Requires bottleneck.

Parameters:

• dim (Hashable) – Specifies the dimension along which to propagate values when filling.

• limit (int or None, default: None) – The maximum number of consecutive NaN values to forward fill. In other words, if there is a gap with more than this number of consecutive NaNs, it will only be partially filled. Must be greater than 0 or None for no limit. Must be None or greater than or equal to axis length if filling along chunked axes (dimensions).

Returns:

filled

Return type:

DataArray

Examples

>>> temperature = np.array(
...     [
...         [np.nan, 1, 3],
...         [0, np.nan, 5],
...         [5, np.nan, np.nan],
...         [3, np.nan, np.nan],
...         [0, 2, 0],
...     ]
... )
>>> da = xr.DataArray(
...     data=temperature,
...     dims=["Y", "X"],
...     coords=dict(
...         lat=("Y", np.array([-20.0, -20.25, -20.50, -20.75, -21.0])),
...         lon=("X", np.array([10.0, 10.25, 10.5])),
...     ),
... )
>>> da
<xarray.DataArray (Y: 5, X: 3)> Size: 120B
array([[nan,  1.,  3.],
       [ 0., nan,  5.],
       [ 5., nan, nan],
       [ 3., nan, nan],
       [ 0.,  2.,  0.]])
Coordinates:
    lat      (Y) float64 40B -20.0 -20.25 -20.5 -20.75 -21.0
    lon      (X) float64 24B 10.0 10.25 10.5
Dimensions without coordinates: Y, X

Fill all NaN values:

>>> da.ffill(dim="Y", limit=None)
<xarray.DataArray (Y: 5, X: 3)> Size: 120B
array([[nan,  1.,  3.],
       [ 0.,  1.,  5.],
       [ 5.,  1.,  5.],
       [ 3.,  1.,  5.],
       [ 0.,  2.,  0.]])
Coordinates:
    lat      (Y) float64 40B -20.0 -20.25 -20.5 -20.75 -21.0
    lon      (X) float64 24B 10.0 10.25 10.5
Dimensions without coordinates: Y, X

Fill only the first of consecutive NaN values:

>>> da.ffill(dim="Y", limit=1)
<xarray.DataArray (Y: 5, X: 3)> Size: 120B
array([[nan,  1.,  3.],
       [ 0.,  1.,  5.],
       [ 5., nan,  5.],
       [ 3., nan, nan],
       [ 0.,  2.,  0.]])
Coordinates:
    lat      (Y) float64 40B -20.0 -20.25 -20.5 -20.75 -21.0
    lon      (X) float64 24B 10.0 10.25 10.5
Dimensions without coordinates: Y, X

fillna()

Fill missing values in this object.

This operation follows the normal broadcasting and alignment rules that xarray uses for binary arithmetic, except the result is aligned to this object (join='left') instead of aligned to the intersection of index coordinates (join='inner').

Parameters:

value (scalar, ndarray or DataArray) – Used to fill all matching missing values in this array. If the argument is a DataArray, it is first aligned with (reindexed to) this array.

Returns:

filled

Return type:

DataArray

Examples

>>> da = xr.DataArray(
...     np.array([1, 4, np.nan, 0, 3, np.nan]),
...     dims="Z",
...     coords=dict(
...         Z=("Z", np.arange(6)),
...         height=("Z", np.array([0, 10, 20, 30, 40, 50])),
...     ),
... )
>>> da
<xarray.DataArray (Z: 6)> Size: 48B
array([ 1.,  4., nan,  0.,  3., nan])
Coordinates:
  * Z        (Z) int64 48B 0 1 2 3 4 5
    height   (Z) int64 48B 0 10 20 30 40 50

Fill all NaN values with 0:

>>> da.fillna(0)
<xarray.DataArray (Z: 6)> Size: 48B
array([1., 4., 0., 0., 3., 0.])
Coordinates:
  * Z        (Z) int64 48B 0 1 2 3 4 5
    height   (Z) int64 48B 0 10 20 30 40 50

Fill NaN values with corresponding values in array:

>>> da.fillna(np.array([2, 9, 4, 2, 8, 9]))
<xarray.DataArray (Z: 6)> Size: 48B
array([1., 4., 4., 0., 3., 9.])
Coordinates:
  * Z        (Z) int64 48B 0 1 2 3 4 5
    height   (Z) int64 48B 0 10 20 30 40 50

classmethod from_dict()

Convert a dictionary into an xarray.DataArray

Parameters:

d (dict) – Mapping with a minimum structure of {“dims”: […], “data”: […]}

Returns:

obj

Return type:

xarray.DataArray

See also

DataArray.to_dict, Dataset.from_dict

Examples

>>> d = {"dims": "t", "data": [1, 2, 3]}
>>> da = xr.DataArray.from_dict(d)
>>> da
<xarray.DataArray (t: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: t

>>> d = {
...     "coords": {
...         "t": {"dims": "t", "data": [0, 1, 2], "attrs": {"units": "s"}}
...     },
...     "attrs": {"title": "air temperature"},
...     "dims": "t",
...     "data": [10, 20, 30],
...     "name": "a",
... }
>>> da = xr.DataArray.from_dict(d)
>>> da
<xarray.DataArray 'a' (t: 3)> Size: 24B
array([10, 20, 30])
Coordinates:
  * t        (t) int64 24B 0 1 2
:ivar title: air temperature

classmethod from_iris()

Convert a iris.cube.Cube into an xarray.DataArray

classmethod from_series()

Convert a pandas.Series into an xarray.DataArray.

If the series’s index is a MultiIndex, it will be expanded into a tensor product of one-dimensional coordinates (filling in missing values with NaN). Thus this operation should be the inverse of the to_series method.

Parameters:

• series (Series) – Pandas Series object to convert.

• sparse (bool, default: False) – If sparse=True, creates a sparse array instead of a dense NumPy array. Requires the pydata/sparse package.

See also

DataArray.to_series, Dataset.from_dataframe

get_axis_num()

Return axis number(s) corresponding to dimension(s) in this array.

Parameters:

dim (str or iterable of str) – Dimension name(s) for which to lookup axes.

Returns:

Axis number or numbers corresponding to the given dimensions.

Return type:

int or tuple of int

get_index()

Get an index for a dimension, with fall-back to a default RangeIndex

groupby_bins()

Returns a DataArrayGroupBy object for performing grouped operations.

Rather than using all unique values of group, the values are discretized first by applying pandas.cut [1] to group.

Parameters:

• group (Hashable, DataArray or IndexVariable) – Array whose binned values should be used to group this array. If a Hashable, must be the name of a coordinate contained in this dataarray.

• bins (int or array-like) – If bins is an int, it defines the number of equal-width bins in the range of x. However, in this case, the range of x is extended by .1% on each side to include the min or max values of x. If bins is a sequence it defines the bin edges allowing for non-uniform bin width. No extension of the range of x is done in this case.

• right (bool, default: True) – Indicates whether the bins include the rightmost edge or not. If right == True (the default), then the bins [1,2,3,4] indicate (1,2], (2,3], (3,4].

• labels (array-like, False or None, default: None) – Used as labels for the resulting bins. Must be of the same length as the resulting bins. If False, string bin labels are assigned by pandas.cut.

• precision (int, default: 3) – The precision at which to store and display the bins labels.

• include_lowest (bool, default: False) – Whether the first interval should be left-inclusive or not.

• squeeze (False) – This argument is deprecated.

• restore_coord_dims (bool, default: False) – If True, also restore the dimension order of multi-dimensional coordinates.

• duplicates ({“raise”, “drop”}, default: “raise”) – If bin edges are not unique, raise ValueError or drop non-uniques.

• eagerly_compute_group (bool, optional) – This argument is deprecated.

Returns:

grouped – A DataArrayGroupBy object patterned after pandas.GroupBy that can be iterated over in the form of (unique_value, grouped_array) pairs. The name of the group has the added suffix _bins in order to distinguish it from the original variable.

Return type:

DataArrayGroupBy

See also

GroupBy: Group and Bin Data

Users guide explanation of how to group and bin data.

DataArray.groupby, Dataset.groupby_bins, core.groupby.DataArrayGroupBy, pandas.DataFrame.groupby

References

[1]

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.cut.html

head()

Return a new DataArray whose data is given by the the first n values along the specified dimension(s). Default n = 5

See also

Dataset.head, DataArray.tail, DataArray.thin

Examples

>>> da = xr.DataArray(
...     np.arange(25).reshape(5, 5),
...     dims=("x", "y"),
... )
>>> da
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Dimensions without coordinates: x, y

>>> da.head(x=1)
<xarray.DataArray (x: 1, y: 5)> Size: 40B
array([[0, 1, 2, 3, 4]])
Dimensions without coordinates: x, y

>>> da.head({"x": 2, "y": 2})
<xarray.DataArray (x: 2, y: 2)> Size: 32B
array([[0, 1],
       [5, 6]])
Dimensions without coordinates: x, y

identical()

Like equals, but also checks the array name and attributes, and attributes on all coordinates.

Parameters:

other (DataArray) – DataArray to compare to.

Returns:

equal – True if the two DataArrays are identical.

Return type:

bool

See also

DataArray.broadcast_equals, DataArray.equals

Examples

>>> a = xr.DataArray([1, 2, 3], dims="X", attrs=dict(units="m"), name="Width")
>>> b = xr.DataArray([1, 2, 3], dims="X", attrs=dict(units="m"), name="Width")
>>> c = xr.DataArray([1, 2, 3], dims="X", attrs=dict(units="ft"), name="Width")
>>> a
<xarray.DataArray 'Width' (X: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: X
:ivar units: m

>>> b
<xarray.DataArray 'Width' (X: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: X
:ivar units: m

>>> c
<xarray.DataArray 'Width' (X: 3)> Size: 24B
array([1, 2, 3])
Dimensions without coordinates: X
:ivar units: ft

>>> a.equals(b)
True
>>> a.identical(b)
True

>>> a.equals(c)
True
>>> a.identical(c)
False

idxmax()

Return the coordinate label of the maximum value along a dimension.

Returns a new DataArray named after the dimension with the values of the coordinate labels along that dimension corresponding to maximum values along that dimension.

In comparison to argmax(), this returns the coordinate label while argmax() returns the index.

Parameters:

• dim (Hashable, optional) – Dimension over which to apply idxmax. This is optional for 1D arrays, but required for arrays with 2 or more dimensions.

• skipna (bool or None, default: None) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float, complex, and object dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (datetime64 or timedelta64).

• fill_value (Any, default: NaN) – Value to be filled in case all of the values along a dimension are null. By default this is NaN. The fill value and result are automatically converted to a compatible dtype if possible. Ignored if skipna is False.

• keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

Returns:

reduced – New DataArray object with idxmax applied to its data and the indicated dimension removed.

Return type:

DataArray

See also

Dataset.idxmax, DataArray.idxmin, DataArray.max, DataArray.argmax

Examples

>>> array = xr.DataArray(
...     [0, 2, 1, 0, -2], dims="x", coords={"x": ["a", "b", "c", "d", "e"]}
... )
>>> array.max()
<xarray.DataArray ()> Size: 8B
array(2)
>>> array.argmax(...)
{'x': <xarray.DataArray ()> Size: 8B
array(1)}
>>> array.idxmax()
<xarray.DataArray 'x' ()> Size: 4B
array('b', dtype='<U1')

>>> array = xr.DataArray(
...     [
...         [2.0, 1.0, 2.0, 0.0, -2.0],
...         [-4.0, np.nan, 2.0, np.nan, -2.0],
...         [np.nan, np.nan, 1.0, np.nan, np.nan],
...     ],
...     dims=["y", "x"],
...     coords={"y": [-1, 0, 1], "x": np.arange(5.0) ** 2},
... )
>>> array.max(dim="x")
<xarray.DataArray (y: 3)> Size: 24B
array([2., 2., 1.])
Coordinates:
  * y        (y) int64 24B -1 0 1
>>> array.argmax(dim="x")
<xarray.DataArray (y: 3)> Size: 24B
array([0, 2, 2])
Coordinates:
  * y        (y) int64 24B -1 0 1
>>> array.idxmax(dim="x")
<xarray.DataArray 'x' (y: 3)> Size: 24B
array([0., 4., 4.])
Coordinates:
  * y        (y) int64 24B -1 0 1

idxmin()

Return the coordinate label of the minimum value along a dimension.

Returns a new DataArray named after the dimension with the values of the coordinate labels along that dimension corresponding to minimum values along that dimension.

In comparison to argmin(), this returns the coordinate label while argmin() returns the index.

Parameters:

• dim (str, optional) – Dimension over which to apply idxmin. This is optional for 1D arrays, but required for arrays with 2 or more dimensions.

• skipna (bool or None, default: None) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float, complex, and object dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (datetime64 or timedelta64).

• fill_value (Any, default: NaN) – Value to be filled in case all of the values along a dimension are null. By default this is NaN. The fill value and result are automatically converted to a compatible dtype if possible. Ignored if skipna is False.

• keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

Returns:

reduced – New DataArray object with idxmin applied to its data and the indicated dimension removed.

Return type:

DataArray

See also

Dataset.idxmin, DataArray.idxmax, DataArray.min, DataArray.argmin

Examples

>>> array = xr.DataArray(
...     [0, 2, 1, 0, -2], dims="x", coords={"x": ["a", "b", "c", "d", "e"]}
... )
>>> array.min()
<xarray.DataArray ()> Size: 8B
array(-2)
>>> array.argmin(...)
{'x': <xarray.DataArray ()> Size: 8B
array(4)}
>>> array.idxmin()
<xarray.DataArray 'x' ()> Size: 4B
array('e', dtype='<U1')

>>> array = xr.DataArray(
...     [
...         [2.0, 1.0, 2.0, 0.0, -2.0],
...         [-4.0, np.nan, 2.0, np.nan, -2.0],
...         [np.nan, np.nan, 1.0, np.nan, np.nan],
...     ],
...     dims=["y", "x"],
...     coords={"y": [-1, 0, 1], "x": np.arange(5.0) ** 2},
... )
>>> array.min(dim="x")
<xarray.DataArray (y: 3)> Size: 24B
array([-2., -4.,  1.])
Coordinates:
  * y        (y) int64 24B -1 0 1
>>> array.argmin(dim="x")
<xarray.DataArray (y: 3)> Size: 24B
array([4, 0, 2])
Coordinates:
  * y        (y) int64 24B -1 0 1
>>> array.idxmin(dim="x")
<xarray.DataArray 'x' (y: 3)> Size: 24B
array([16.,  0.,  4.])
Coordinates:
  * y        (y) int64 24B -1 0 1

property imag

The imaginary part of the array.

See also

numpy.ndarray.imag

property indexes

Mapping of pandas.Index objects used for label based indexing.

Raises an error if this Dataset has indexes that cannot be coerced to pandas.Index objects.

See also

DataArray.xindexes

integrate()

Integrate along the given coordinate using the trapezoidal rule.

Note

This feature is limited to simple cartesian geometry, i.e. coord must be one dimensional.

Parameters:

• coord (Hashable, or sequence of Hashable) – Coordinate(s) used for the integration.

• datetime_unit ({‘W’, ‘D’, ‘h’, ‘m’, ‘s’, ‘ms’, ‘us’, ‘ns’, ‘ps’, ‘fs’, ‘as’, None}, optional) – Specify the unit if a datetime coordinate is used.

Returns:

integrated

Return type:

DataArray

See also

Dataset.integrate

numpy.trapz

corresponding numpy function

Examples

>>> da = xr.DataArray(
...     np.arange(12).reshape(4, 3),
...     dims=["x", "y"],
...     coords={"x": [0, 0.1, 1.1, 1.2]},
... )
>>> da
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) float64 32B 0.0 0.1 1.1 1.2
Dimensions without coordinates: y
>>>
>>> da.integrate("x")
<xarray.DataArray (y: 3)> Size: 24B
array([5.4, 6.6, 7.8])
Dimensions without coordinates: y

interp_calendar()

Interpolates the DataArray to another calendar based on decimal year measure.

Each timestamp in source and target are first converted to their decimal year equivalent then source is interpolated on the target coordinate. The decimal year of a timestamp is its year plus its sub-year component converted to the fraction of its year. For example “2000-03-01 12:00” is 2000.1653 in a standard calendar or 2000.16301 in a “noleap” calendar.

This method should only be used when the time (HH:MM:SS) information of time coordinate is not important.

Parameters:

• target (DataArray or DatetimeIndex or CFTimeIndex) – The target time coordinate of a valid dtype (np.datetime64 or cftime objects)

• dim (str) – The time coordinate name.

Returns:

The source interpolated on the decimal years of target,

Return type:

DataArray

interp_like()

Interpolate this object onto the coordinates of another object, filling out of range values with NaN.

Parameters:

• other (Dataset or DataArray) – Object with an ‘indexes’ attribute giving a mapping from dimension names to an 1d array-like, which provides coordinates upon which to index the variables in this dataset. Missing values are skipped.

• method ({ “linear”, “nearest”, “zero”, “slinear”, “quadratic”, “cubic”, “quintic”, “polynomial”, “pchip”, “barycentric”, “krogh”, “akima”, “makima” }) – Interpolation method to use (see descriptions above).

• assume_sorted (bool, default: False) – If False, values of coordinates that are interpolated over can be in any order and they are sorted first. If True, interpolated coordinates are assumed to be an array of monotonically increasing values.

• kwargs (dict, optional) – Additional keyword arguments passed to the interpolant.

Returns:

interpolated – Another dataarray by interpolating this dataarray’s data along the coordinates of the other object.

Return type:

DataArray

Notes

• scipy is required.

• If the dataarray has object-type coordinates, reindex is used for these

  coordinates instead of the interpolation.

• When interpolating along multiple dimensions with methods linear and nearest,

  the process attempts to decompose the interpolation into independent interpolations along one dimension at a time.

• The specific interpolation method and dimensionality determine which

  interpolant is used:

  1. Interpolation along one dimension of 1D data (`method=’linear’`)

     • Uses numpy.interp(), unless fill_value=’extrapolate’ is provided via kwargs.

  2. Interpolation along one dimension of N-dimensional data (N ≥ 1)

     • Methods {“linear”, “nearest”, “zero”, “slinear”, “quadratic”, “cubic”, “quintic”, “polynomial”}

       use scipy.interpolate.interp1d(), unless conditions permit the use of numpy.interp() (as in the case of method=’linear’ for 1D data).

     • If method=’polynomial’, the order keyword argument must also be provided.

  3. Special interpolants for interpolation along one dimension of N-dimensional data (N ≥ 1)

     • Depending on the method, the following interpolants from scipy.interpolate are used:

       • “pchip”: scipy.interpolate.PchipInterpolator

       • “barycentric”: scipy.interpolate.BarycentricInterpolator

       • “krogh”: scipy.interpolate.KroghInterpolator

       • “akima” or “makima”: scipy.interpolate.Akima1dInterpolator

         (makima is handled by passing the makima flag).

  4. Interpolation along multiple dimensions of multi-dimensional data

     • Uses scipy.interpolate.interpn() for methods {“linear”, “nearest”, “slinear”,

       “cubic”, “quintic”, “pchip”}.

See also

DataArray.interp(), DataArray.reindex_like(), scipy.interpolate

Examples

>>> data = np.arange(12).reshape(4, 3)
>>> da1 = xr.DataArray(
...     data=data,
...     dims=["x", "y"],
...     coords={"x": [10, 20, 30, 40], "y": [70, 80, 90]},
... )
>>> da1
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) int64 32B 10 20 30 40
  * y        (y) int64 24B 70 80 90
>>> da2 = xr.DataArray(
...     data=data,
...     dims=["x", "y"],
...     coords={"x": [10, 20, 29, 39], "y": [70, 80, 90]},
... )
>>> da2
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) int64 32B 10 20 29 39
  * y        (y) int64 24B 70 80 90

Interpolate the values in the coordinates of the other DataArray with respect to the source’s values:

>>> da2.interp_like(da1)
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[0. , 1. , 2. ],
       [3. , 4. , 5. ],
       [6.3, 7.3, 8.3],
       [nan, nan, nan]])
Coordinates:
  * x        (x) int64 32B 10 20 30 40
  * y        (y) int64 24B 70 80 90

Could also extrapolate missing values:

>>> da2.interp_like(da1, kwargs={"fill_value": "extrapolate"})
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0. ,  1. ,  2. ],
       [ 3. ,  4. ,  5. ],
       [ 6.3,  7.3,  8.3],
       [ 9.3, 10.3, 11.3]])
Coordinates:
  * x        (x) int64 32B 10 20 30 40
  * y        (y) int64 24B 70 80 90

interpolate_na()

Fill in NaNs by interpolating according to different methods.

Parameters:

• dim (Hashable or None, optional) – Specifies the dimension along which to interpolate.

• method ({“linear”, “nearest”, “zero”, “slinear”, “quadratic”, “cubic”, “polynomial”, “barycentric”, “krogh”, “pchip”, “spline”, “akima”}, default: “linear”) – String indicating which method to use for interpolation:

  • ‘linear’: linear interpolation. Additional keyword arguments are passed to numpy.interp()

  • ‘nearest’, ‘zero’, ‘slinear’, ‘quadratic’, ‘cubic’, ‘polynomial’: are passed to scipy.interpolate.interp1d(). If method='polynomial', the order keyword argument must also be provided.

  • ‘barycentric’, ‘krogh’, ‘pchip’, ‘spline’, ‘akima’: use their respective scipy.interpolate classes.

• use_coordinate (bool or str, default: True) – Specifies which index to use as the x values in the interpolation formulated as y = f(x). If False, values are treated as if equally-spaced along dim. If True, the IndexVariable dim is used. If use_coordinate is a string, it specifies the name of a coordinate variable to use as the index.

• limit (int or None, default: None) – Maximum number of consecutive NaNs to fill. Must be greater than 0 or None for no limit. This filling is done regardless of the size of the gap in the data. To only interpolate over gaps less than a given length, see max_gap.

• max_gap (int, float, str, pandas.Timedelta, numpy.timedelta64, datetime.timedelta, default: None) – Maximum size of gap, a continuous sequence of NaNs, that will be filled. Use None for no limit. When interpolating along a datetime64 dimension and use_coordinate=True, max_gap can be one of the following:

  • a string that is valid input for pandas.to_timedelta

  • a numpy.timedelta64 object

  • a pandas.Timedelta object

  • a datetime.timedelta object

  Otherwise, max_gap must be an int or a float. Use of max_gap with unlabeled dimensions has not been implemented yet. Gap length is defined as the difference between coordinate values at the first data point after a gap and the last value before a gap. For gaps at the beginning (end), gap length is defined as the difference between coordinate values at the first (last) valid data point and the first (last) NaN. For example, consider:

  <xarray.DataArray (x: 9)>
  array([nan, nan, nan,  1., nan, nan,  4., nan, nan])
  Coordinates:
    * x        (x) int64 0 1 2 3 4 5 6 7 8
  

  The gap lengths are 3-0 = 3; 6-3 = 3; and 8-6 = 2 respectively

• keep_attrs (bool or None, default: None) – If True, the dataarray’s attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (dict, optional) – parameters passed verbatim to the underlying interpolation function

Returns:

interpolated – Filled in DataArray.

Return type:

DataArray

See also

numpy.interp, scipy.interpolate

Examples

>>> da = xr.DataArray(
...     [np.nan, 2, 3, np.nan, 0], dims="x", coords={"x": [0, 1, 2, 3, 4]}
... )
>>> da
<xarray.DataArray (x: 5)> Size: 40B
array([nan,  2.,  3., nan,  0.])
Coordinates:
  * x        (x) int64 40B 0 1 2 3 4

>>> da.interpolate_na(dim="x", method="linear")
<xarray.DataArray (x: 5)> Size: 40B
array([nan, 2. , 3. , 1.5, 0. ])
Coordinates:
  * x        (x) int64 40B 0 1 2 3 4

>>> da.interpolate_na(dim="x", method="linear", fill_value="extrapolate")
<xarray.DataArray (x: 5)> Size: 40B
array([1. , 2. , 3. , 1.5, 0. ])
Coordinates:
  * x        (x) int64 40B 0 1 2 3 4

isin()

Tests each value in the array for whether it is in test elements.

Parameters:

test_elements (array_like) – The values against which to test each value of element. This argument is flattened if an array or array_like. See numpy notes for behavior with non-array-like parameters.

Returns:

isin – Has the same type and shape as this object, but with a bool dtype.

Return type:

DataArray or Dataset

Examples

>>> array = xr.DataArray([1, 2, 3], dims="x")
>>> array.isin([1, 3])
<xarray.DataArray (x: 3)> Size: 3B
array([ True, False,  True])
Dimensions without coordinates: x

See also

numpy.isin

isnull()

Test each value in the array for whether it is a missing value.

Parameters:

keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

Returns:

isnull – Same type and shape as object, but the dtype of the data is bool.

Return type:

DataArray or Dataset

See also

pandas.isnull

Examples

>>> array = xr.DataArray([1, np.nan, 3], dims="x")
>>> array
<xarray.DataArray (x: 3)> Size: 24B
array([ 1., nan,  3.])
Dimensions without coordinates: x
>>> array.isnull()
<xarray.DataArray (x: 3)> Size: 3B
array([False,  True, False])
Dimensions without coordinates: x

item(*args)

Copy an element of an array to a standard Python scalar and return it.

Parameters:

*args (Arguments (variable number and type)) –

• none: in this case, the method only works for arrays with one element (a.size == 1), which element is copied into a standard Python scalar object and returned.

• int_type: this argument is interpreted as a flat index into the array, specifying which element to copy and return.

• tuple of int_types: functions as does a single int_type argument, except that the argument is interpreted as an nd-index into the array.

Returns:

z – A copy of the specified element of the array as a suitable Python scalar

Return type:

Standard Python scalar object

Notes

When the data type of a is longdouble or clongdouble, item() returns a scalar array object because there is no available Python scalar that would not lose information. Void arrays return a buffer object for item(), unless fields are defined, in which case a tuple is returned.

item is very similar to a[args], except, instead of an array scalar, a standard Python scalar is returned. This can be useful for speeding up access to elements of the array and doing arithmetic on elements of the array using Python’s optimized math.

Examples

>>> import numpy as np
>>> np.random.seed(123)
>>> x = np.random.randint(9, size=(3, 3))
>>> x
array([[2, 2, 6],
       [1, 3, 6],
       [1, 0, 1]])
>>> x.item(3)
1
>>> x.item(7)
0
>>> x.item((0, 1))
2
>>> x.item((2, 2))
1

For an array with object dtype, elements are returned as-is.

>>> a = np.array([np.int64(1)], dtype=object)
>>> a.item() #return np.int64
np.int64(1)

load()

Trigger loading data into memory and return this dataarray.

Data will be computed and/or loaded from disk or a remote source.

Unlike .compute, the original dataarray is modified and returned.

Normally, it should not be necessary to call this method in user code, because all xarray functions should either work on deferred data or load data automatically. However, this method can be necessary when working with many file objects on disk.

Parameters:

**kwargs (dict) – Additional keyword arguments passed on to dask.compute.

Returns:

object – Same object but with lazy data and coordinates as in-memory arrays.

Return type:

DataArray

See also

dask.compute, DataArray.load_async, DataArray.compute, Dataset.load, Variable.load

async load_async()

Trigger and await asynchronous loading of data into memory and return this dataarray.

Data will be computed and/or loaded from disk or a remote source.

Unlike .compute, the original dataarray is modified and returned.

Only works when opening data lazily from IO storage backends which support lazy asynchronous loading. Otherwise will raise a NotImplementedError.

Note users are expected to limit concurrency themselves - xarray does not internally limit concurrency in any way.

Parameters:

**kwargs (dict) – Additional keyword arguments passed on to dask.compute.

Returns:

object – Same object but with lazy data and coordinates as in-memory arrays.

Return type:

Dataarray

See also

dask.compute, DataArray.compute, DataArray.load, Dataset.load_async, Variable.load_async

property loc

Attribute for location based indexing like pandas.

max()

Reduce this DataArray’s data by applying max along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply max. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating max on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with max applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.max, dask.array.max, Dataset.max

Aggregation

User guide on reduction or aggregation operations.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.max()
<xarray.DataArray ()> Size: 8B
array(3.)

Use skipna to control whether NaNs are ignored.

>>> da.max(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

mean()

Reduce this DataArray’s data by applying mean along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply mean. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating mean on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with mean applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.mean, dask.array.mean, Dataset.mean

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.mean()
<xarray.DataArray ()> Size: 8B
array(1.6)

Use skipna to control whether NaNs are ignored.

>>> da.mean(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

median()

Reduce this DataArray’s data by applying median along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply median. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating median on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with median applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.median, dask.array.median, Dataset.median

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.median()
<xarray.DataArray ()> Size: 8B
array(2.)

Use skipna to control whether NaNs are ignored.

>>> da.median(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

min()

Reduce this DataArray’s data by applying min along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply min. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating min on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with min applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.min, dask.array.min, Dataset.min

Aggregation

User guide on reduction or aggregation operations.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.min()
<xarray.DataArray ()> Size: 8B
array(0.)

Use skipna to control whether NaNs are ignored.

>>> da.min(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

property name

The name of this array.

property nbytes

Total bytes consumed by the elements of this DataArray’s data.

If the underlying data array does not include nbytes, estimates the bytes consumed based on the size and dtype.

property ndim

Number of array dimensions.

See also

numpy.ndarray.ndim

notnull()

Test each value in the array for whether it is not a missing value.

Parameters:

keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

Returns:

notnull – Same type and shape as object, but the dtype of the data is bool.

Return type:

DataArray or Dataset

See also

pandas.notnull

Examples

>>> array = xr.DataArray([1, np.nan, 3], dims="x")
>>> array
<xarray.DataArray (x: 3)> Size: 24B
array([ 1., nan,  3.])
Dimensions without coordinates: x
>>> array.notnull()
<xarray.DataArray (x: 3)> Size: 3B
array([ True, False,  True])
Dimensions without coordinates: x

pad()

Pad this array along one or more dimensions.

Warning

This function is experimental and its behaviour is likely to change especially regarding padding of dimension coordinates (or IndexVariables).

When using one of the modes (“edge”, “reflect”, “symmetric”, “wrap”), coordinates will be padded with the same mode, otherwise coordinates are padded using the “constant” mode with fill_value dtypes.NA.

Parameters:

• pad_width (mapping of Hashable to tuple of int) – Mapping with the form of {dim: (pad_before, pad_after)} describing the number of values padded along each dimension. {dim: pad} is a shortcut for pad_before = pad_after = pad

• mode ({“constant”, “edge”, “linear_ramp”, “maximum”, “mean”, “median”, “minimum”, “reflect”, “symmetric”, “wrap”}, default: “constant”) – How to pad the DataArray (taken from numpy docs):

  • “constant”: Pads with a constant value.

  • “edge”: Pads with the edge values of array.

  • “linear_ramp”: Pads with the linear ramp between end_value and the array edge value.

  • “maximum”: Pads with the maximum value of all or part of the vector along each axis.

  • “mean”: Pads with the mean value of all or part of the vector along each axis.

  • “median”: Pads with the median value of all or part of the vector along each axis.

  • “minimum”: Pads with the minimum value of all or part of the vector along each axis.

  • “reflect”: Pads with the reflection of the vector mirrored on the first and last values of the vector along each axis.

  • “symmetric”: Pads with the reflection of the vector mirrored along the edge of the array.

  • “wrap”: Pads with the wrap of the vector along the axis. The first values are used to pad the end and the end values are used to pad the beginning.

• stat_length (int, tuple or mapping of Hashable to tuple, default: None) – Used in ‘maximum’, ‘mean’, ‘median’, and ‘minimum’. Number of values at edge of each axis used to calculate the statistic value. {dim_1: (before_1, after_1), … dim_N: (before_N, after_N)} unique statistic lengths along each dimension. ((before, after),) yields same before and after statistic lengths for each dimension. (stat_length,) or int is a shortcut for before = after = statistic length for all axes. Default is None, to use the entire axis.

• constant_values (scalar, tuple or mapping of Hashable to tuple, default: 0) – Used in ‘constant’. The values to set the padded values for each axis. {dim_1: (before_1, after_1), ... dim_N: (before_N, after_N)} unique pad constants along each dimension. ((before, after),) yields same before and after constants for each dimension. (constant,) or constant is a shortcut for before = after = constant for all dimensions. Default is 0.

• end_values (scalar, tuple or mapping of Hashable to tuple, default: 0) – Used in ‘linear_ramp’. The values used for the ending value of the linear_ramp and that will form the edge of the padded array. {dim_1: (before_1, after_1), ... dim_N: (before_N, after_N)} unique end values along each dimension. ((before, after),) yields same before and after end values for each axis. (constant,) or constant is a shortcut for before = after = constant for all axes. Default is 0.

• reflect_type ({“even”, “odd”, None}, optional) – Used in “reflect”, and “symmetric”. The “even” style is the default with an unaltered reflection around the edge value. For the “odd” style, the extended part of the array is created by subtracting the reflected values from two times the edge value.

• keep_attrs (bool or None, optional) – If True, the attributes (attrs) will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **pad_width_kwargs – The keyword arguments form of pad_width. One of pad_width or pad_width_kwargs must be provided.

Returns:

padded – DataArray with the padded coordinates and data.

Return type:

DataArray

See also

DataArray.shift, DataArray.roll, DataArray.bfill, DataArray.ffill, numpy.pad, dask.array.pad

Notes

For mode="constant" and constant_values=None, integer types will be promoted to float and padded with np.nan.

Padding coordinates will drop their corresponding index (if any) and will reset default indexes for dimension coordinates.

Examples

>>> arr = xr.DataArray([5, 6, 7], coords=[("x", [0, 1, 2])])
>>> arr.pad(x=(1, 2), constant_values=0)
<xarray.DataArray (x: 6)> Size: 48B
array([0, 5, 6, 7, 0, 0])
Coordinates:
  * x        (x) float64 48B nan 0.0 1.0 2.0 nan nan

>>> da = xr.DataArray(
...     [[0, 1, 2, 3], [10, 11, 12, 13]],
...     dims=["x", "y"],
...     coords={"x": [0, 1], "y": [10, 20, 30, 40], "z": ("x", [100, 200])},
... )
>>> da.pad(x=1)
<xarray.DataArray (x: 4, y: 4)> Size: 128B
array([[nan, nan, nan, nan],
       [ 0.,  1.,  2.,  3.],
       [10., 11., 12., 13.],
       [nan, nan, nan, nan]])
Coordinates:
  * x        (x) float64 32B nan 0.0 1.0 nan
  * y        (y) int64 32B 10 20 30 40
    z        (x) float64 32B nan 100.0 200.0 nan

Careful, constant_values are coerced to the data type of the array which may lead to a loss of precision:

>>> da.pad(x=1, constant_values=1.23456789)
<xarray.DataArray (x: 4, y: 4)> Size: 128B
array([[ 1,  1,  1,  1],
       [ 0,  1,  2,  3],
       [10, 11, 12, 13],
       [ 1,  1,  1,  1]])
Coordinates:
  * x        (x) float64 32B nan 0.0 1.0 nan
  * y        (y) int64 32B 10 20 30 40
    z        (x) float64 32B nan 100.0 200.0 nan

persist()

Trigger computation in constituent dask arrays

This keeps them as dask arrays but encourages them to keep data in memory. This is particularly useful when on a distributed machine. When on a single machine consider using .compute() instead. Like compute (but unlike load), the original dataset is left unaltered.

Parameters:

**kwargs (dict) – Additional keyword arguments passed on to dask.persist.

Returns:

object – New object with all dask-backed data and coordinates as persisted dask arrays.

Return type:

DataArray

See also

dask.persist

pipe()

Apply func(self, *args, **kwargs)

This method replicates the pandas method of the same name.

Parameters:

• func (callable) – function to apply to this xarray object (Dataset/DataArray). args, and kwargs are passed into func. Alternatively a (callable, data_keyword) tuple where data_keyword is a string indicating the keyword of callable that expects the xarray object.

• *args – positional arguments passed into func.

• **kwargs – a dictionary of keyword arguments passed into func.

Returns:

object – the return type of func.

Return type:

Any

Notes

Use .pipe when chaining together functions that expect xarray or pandas objects, e.g., instead of writing

f(g(h(ds), arg1=a), arg2=b, arg3=c)

You can write

(ds.pipe(h).pipe(g, arg1=a).pipe(f, arg2=b, arg3=c))

If you have a function that takes the data as (say) the second argument, pass a tuple indicating which keyword expects the data. For example, suppose f takes its data as arg2:

(ds.pipe(h).pipe(g, arg1=a).pipe((f, "arg2"), arg1=a, arg3=c))

Examples

>>> x = xr.Dataset(
...     {
...         "temperature_c": (
...             ("lat", "lon"),
...             20 * np.random.rand(4).reshape(2, 2),
...         ),
...         "precipitation": (("lat", "lon"), np.random.rand(4).reshape(2, 2)),
...     },
...     coords={"lat": [10, 20], "lon": [150, 160]},
... )
>>> x
<xarray.Dataset> Size: 96B
Dimensions:        (lat: 2, lon: 2)
Coordinates:
  * lat            (lat) int64 16B 10 20
  * lon            (lon) int64 16B 150 160
Data variables:
    temperature_c  (lat, lon) float64 32B 10.98 14.3 12.06 10.9
    precipitation  (lat, lon) float64 32B 0.4237 0.6459 0.4376 0.8918

>>> def adder(data, arg):
...     return data + arg
...
>>> def div(data, arg):
...     return data / arg
...
>>> def sub_mult(data, sub_arg, mult_arg):
...     return (data * mult_arg) - sub_arg
...
>>> x.pipe(adder, 2)
<xarray.Dataset> Size: 96B
Dimensions:        (lat: 2, lon: 2)
Coordinates:
  * lat            (lat) int64 16B 10 20
  * lon            (lon) int64 16B 150 160
Data variables:
    temperature_c  (lat, lon) float64 32B 12.98 16.3 14.06 12.9
    precipitation  (lat, lon) float64 32B 2.424 2.646 2.438 2.892

>>> x.pipe(adder, arg=2)
<xarray.Dataset> Size: 96B
Dimensions:        (lat: 2, lon: 2)
Coordinates:
  * lat            (lat) int64 16B 10 20
  * lon            (lon) int64 16B 150 160
Data variables:
    temperature_c  (lat, lon) float64 32B 12.98 16.3 14.06 12.9
    precipitation  (lat, lon) float64 32B 2.424 2.646 2.438 2.892

>>> (
...     x.pipe(adder, arg=2)
...     .pipe(div, arg=2)
...     .pipe(sub_mult, sub_arg=2, mult_arg=2)
... )
<xarray.Dataset> Size: 96B
Dimensions:        (lat: 2, lon: 2)
Coordinates:
  * lat            (lat) int64 16B 10 20
  * lon            (lon) int64 16B 150 160
Data variables:
    temperature_c  (lat, lon) float64 32B 10.98 14.3 12.06 10.9
    precipitation  (lat, lon) float64 32B 0.4237 0.6459 0.4376 0.8918

See also

pandas.DataFrame.pipe

plot

alias of DataArrayPlotAccessor

polyfit()

Least squares polynomial fit.

This replicates the behaviour of numpy.polyfit but differs by skipping invalid values when skipna = True.

Parameters:

• dim (Hashable) – Coordinate along which to fit the polynomials.

• deg (int) – Degree of the fitting polynomial.

• skipna (bool or None, optional) – If True, removes all invalid values before fitting each 1D slices of the array. Default is True if data is stored in a dask.array or if there is any invalid values, False otherwise.

• rcond (float or None, optional) – Relative condition number to the fit.

• w (Hashable, array-like or None, optional) – Weights to apply to the y-coordinate of the sample points. Can be an array-like object or the name of a coordinate in the dataset.

• full (bool, default: False) – Whether to return the residuals, matrix rank and singular values in addition to the coefficients.

• cov (bool or “unscaled”, default: False) – Whether to return to the covariance matrix in addition to the coefficients. The matrix is not scaled if cov=’unscaled’.

Returns:

polyfit_results – A single dataset which contains:

polyfit_coefficients

The coefficients of the best fit.

polyfit_residuals

The residuals of the least-square computation (only included if full=True). When the matrix rank is deficient, np.nan is returned.

[dim]_matrix_rank

The effective rank of the scaled Vandermonde coefficient matrix (only included if full=True)

[dim]_singular_value

The singular values of the scaled Vandermonde coefficient matrix (only included if full=True)

polyfit_covariance

The covariance matrix of the polynomial coefficient estimates (only included if full=False and cov=True)

Return type:

Dataset

See also

numpy.polyfit, numpy.polyval, xarray.polyval, DataArray.curvefit

prod()

Reduce this DataArray’s data by applying prod along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply prod. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• min_count (int or None, optional) – The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA. Only used if skipna is set to True or defaults to True for the array’s dtype. Changed in version 0.17.0: if specified on an integer array and skipna=True, the result will be a float array.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating prod on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with prod applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.prod, dask.array.prod, Dataset.prod

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.prod()
<xarray.DataArray ()> Size: 8B
array(0.)

Use skipna to control whether NaNs are ignored.

>>> da.prod(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

Specify min_count for finer control over when NaNs are ignored.

>>> da.prod(skipna=True, min_count=2)
<xarray.DataArray ()> Size: 8B
array(0.)

quantile()

Compute the qth quantile of the data along the specified dimension.

Returns the qth quantiles(s) of the array elements.

Parameters:

• q (float or array-like of float) – Quantile to compute, which must be between 0 and 1 inclusive.

• dim (str or Iterable of Hashable, optional) – Dimension(s) over which to apply quantile.

• method (str, default: “linear”) – This optional parameter specifies the interpolation method to use when the desired quantile lies between two data points. The options sorted by their R type as summarized in the H&F paper [1] are:

  1. “inverted_cdf”

  2. “averaged_inverted_cdf”

  3. “closest_observation”

  4. “interpolated_inverted_cdf”

  5. “hazen”

  6. “weibull”

  7. “linear” (default)

  8. “median_unbiased”

  9. “normal_unbiased”

  The first three methods are discontiuous. The following discontinuous variations of the default “linear” (7.) option are also available:

  • “lower”

  • “higher”

  • “midpoint”

  • “nearest”

  See numpy.quantile() or [1] for details. The “method” argument was previously called “interpolation”, renamed in accordance with numpy version 1.22.0.

• keep_attrs (bool or None, optional) – If True, the dataset’s attributes (attrs) will be copied from the original object to the new one. If False (default), the new object will be returned without attributes.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

Returns:

quantiles – If q is a single quantile, then the result is a scalar. If multiple percentiles are given, first axis of the result corresponds to the quantile and a quantile dimension is added to the return array. The other dimensions are the dimensions that remain after the reduction of the array.

Return type:

DataArray

See also

numpy.nanquantile, numpy.quantile, pandas.Series.quantile, Dataset.quantile

Examples

>>> da = xr.DataArray(
...     data=[[0.7, 4.2, 9.4, 1.5], [6.5, 7.3, 2.6, 1.9]],
...     coords={"x": [7, 9], "y": [1, 1.5, 2, 2.5]},
...     dims=("x", "y"),
... )
>>> da.quantile(0)  # or da.quantile(0, dim=...)
<xarray.DataArray ()> Size: 8B
array(0.7)
Coordinates:
    quantile  float64 8B 0.0
>>> da.quantile(0, dim="x")
<xarray.DataArray (y: 4)> Size: 32B
array([0.7, 4.2, 2.6, 1.5])
Coordinates:
  * y         (y) float64 32B 1.0 1.5 2.0 2.5
    quantile  float64 8B 0.0
>>> da.quantile([0, 0.5, 1])
<xarray.DataArray (quantile: 3)> Size: 24B
array([0.7, 3.4, 9.4])
Coordinates:
  * quantile  (quantile) float64 24B 0.0 0.5 1.0
>>> da.quantile([0, 0.5, 1], dim="x")
<xarray.DataArray (quantile: 3, y: 4)> Size: 96B
array([[0.7 , 4.2 , 2.6 , 1.5 ],
       [3.6 , 5.75, 6.  , 1.7 ],
       [6.5 , 7.3 , 9.4 , 1.9 ]])
Coordinates:
  * y         (y) float64 32B 1.0 1.5 2.0 2.5
  * quantile  (quantile) float64 24B 0.0 0.5 1.0

References

[1] (1,2)

R. J. Hyndman and Y. Fan, “Sample quantiles in statistical packages,” The American Statistician, 50(4), pp. 361-365, 1996

query()

Return a new data array indexed along the specified dimension(s), where the indexers are given as strings containing Python expressions to be evaluated against the values in the array.

Parameters:

• queries (dict-like or None, optional) – A dict-like with keys matching dimensions and values given by strings containing Python expressions to be evaluated against the data variables in the dataset. The expressions will be evaluated using the pandas eval() function, and can contain any valid Python expressions but cannot contain any Python statements.

• parser ({“pandas”, “python”}, default: “pandas”) – The parser to use to construct the syntax tree from the expression. The default of ‘pandas’ parses code slightly different than standard Python. Alternatively, you can parse an expression using the ‘python’ parser to retain strict Python semantics.

• engine ({“python”, “numexpr”, None}, default: None) – The engine used to evaluate the expression. Supported engines are:

  • None: tries to use numexpr, falls back to python

  • “numexpr”: evaluates expressions using numexpr

  • “python”: performs operations as if you had eval’d in top level python

• missing_dims ({“raise”, “warn”, “ignore”}, default: “raise”) – What to do if dimensions that should be selected from are not present in the DataArray:

  • “raise”: raise an exception

  • “warn”: raise a warning, and ignore the missing dimensions

  • “ignore”: ignore the missing dimensions

• **queries_kwargs ({dim: query, …}, optional) – The keyword arguments form of queries. One of queries or queries_kwargs must be provided.

Returns:

obj – A new DataArray with the same contents as this dataset, indexed by the results of the appropriate queries.

Return type:

DataArray

See also

DataArray.isel, Dataset.query, pandas.eval

Examples

>>> da = xr.DataArray(np.arange(0, 5, 1), dims="x", name="a")
>>> da
<xarray.DataArray 'a' (x: 5)> Size: 40B
array([0, 1, 2, 3, 4])
Dimensions without coordinates: x
>>> da.query(x="a > 2")
<xarray.DataArray 'a' (x: 2)> Size: 16B
array([3, 4])
Dimensions without coordinates: x

rank()

Ranks the data.

Equal values are assigned a rank that is the average of the ranks that would have been otherwise assigned to all of the values within that set. Ranks begin at 1, not 0. If pct, computes percentage ranks.

NaNs in the input array are returned as NaNs.

The bottleneck library is required.

Parameters:

• dim (Hashable) – Dimension over which to compute rank.

• pct (bool, default: False) – If True, compute percentage ranks, otherwise compute integer ranks.

• keep_attrs (bool or None, optional) – If True, the dataset’s attributes (attrs) will be copied from the original object to the new one. If False (default), the new object will be returned without attributes.

Returns:

ranked – DataArray with the same coordinates and dtype ‘float64’.

Return type:

DataArray

Examples

>>> arr = xr.DataArray([5, 6, 7], dims="x")
>>> arr.rank("x")
<xarray.DataArray (x: 3)> Size: 24B
array([1., 2., 3.])
Dimensions without coordinates: x

property real

The real part of the array.

See also

numpy.ndarray.real

reduce()

Reduce this array by applying func along some dimension(s).

Parameters:

• func (callable) – Function which can be called in the form f(x, axis=axis, **kwargs) to return the result of reducing an np.ndarray over an integer valued axis.

• dim (”…”, str, Iterable of Hashable or None, optional) – Dimension(s) over which to apply func. By default func is applied over all dimensions.

• axis (int or sequence of int, optional) – Axis(es) over which to repeatedly apply func. Only one of the ‘dim’ and ‘axis’ arguments can be supplied. If neither are supplied, then the reduction is calculated over the flattened array (by calling f(x) without an axis argument).

• keep_attrs (bool or None, optional) – If True, the variable’s attributes (attrs) will be copied from the original object to the new one. If False (default), the new object will be returned without attributes.

• keepdims (bool, default: False) – If True, the dimensions which are reduced are left in the result as dimensions of size one. Coordinates that use these dimensions are removed.

• **kwargs (dict) – Additional keyword arguments passed on to func.

Returns:

reduced – DataArray with this object’s array replaced with an array with summarized data and the indicated dimension(s) removed.

Return type:

DataArray

reindex()

Conform this object onto the indexes of another object, filling in missing values with fill_value. The default fill value is NaN.

Parameters:

• indexers (dict, optional) – Dictionary with keys given by dimension names and values given by arrays of coordinates tick labels. Any mismatched coordinate values will be filled in with NaN, and any mismatched dimension names will simply be ignored. One of indexers or indexers_kwargs must be provided.

• copy (bool, optional) – If copy=True, data in the return value is always copied. If copy=False and reindexing is unnecessary, or can be performed with only slice operations, then the output may share memory with the input. In either case, a new xarray object is always returned.

• method ({None, ‘nearest’, ‘pad’/’ffill’, ‘backfill’/’bfill’}, optional) – Method to use for filling index values in indexers not found on this data array:

  • None (default): don’t fill gaps

  • pad / ffill: propagate last valid index value forward

  • backfill / bfill: propagate next valid index value backward

  • nearest: use nearest valid index value

• tolerance (float | Iterable[float] | str | None, default: None) – Maximum distance between original and new labels for inexact matches. The values of the index at the matching locations must satisfy the equation abs(index[indexer] - target) <= tolerance. Tolerance may be a scalar value, which applies the same tolerance to all values, or list-like, which applies variable tolerance per element. List-like must be the same size as the index and its dtype must exactly match the index’s type.

• fill_value (scalar or dict-like, optional) – Value to use for newly missing values. If a dict-like, maps variable names (including coordinates) to fill values. Use this data array’s name to refer to the data array’s values.

• **indexers_kwargs ({dim: indexer, …}, optional) – The keyword arguments form of indexers. One of indexers or indexers_kwargs must be provided.

Returns:

reindexed – Another dataset array, with this array’s data but replaced coordinates.

Return type:

DataArray

Examples

Reverse latitude:

>>> da = xr.DataArray(
...     np.arange(4),
...     coords=[np.array([90, 89, 88, 87])],
...     dims="lat",
... )
>>> da
<xarray.DataArray (lat: 4)> Size: 32B
array([0, 1, 2, 3])
Coordinates:
  * lat      (lat) int64 32B 90 89 88 87
>>> da.reindex(lat=da.lat[::-1])
<xarray.DataArray (lat: 4)> Size: 32B
array([3, 2, 1, 0])
Coordinates:
  * lat      (lat) int64 32B 87 88 89 90

See also

DataArray.reindex_like, align

reindex_like()

Conform this object onto the indexes of another object, for indexes which the objects share. Missing values are filled with fill_value. The default fill value is NaN.

Parameters:

• other (Dataset or DataArray) – Object with an ‘indexes’ attribute giving a mapping from dimension names to pandas.Index objects, which provides coordinates upon which to index the variables in this dataset. The indexes on this other object need not be the same as the indexes on this dataset. Any mismatched index values will be filled in with NaN, and any mismatched dimension names will simply be ignored.

• method ({None, “nearest”, “pad”, “ffill”, “backfill”, “bfill”}, optional) – Method to use for filling index values from other not found on this data array:

  • None (default): don’t fill gaps

  • pad / ffill: propagate last valid index value forward

  • backfill / bfill: propagate next valid index value backward

  • nearest: use nearest valid index value

• tolerance (float | Iterable[float] | str | None, default: None) – Maximum distance between original and new labels for inexact matches. The values of the index at the matching locations must satisfy the equation abs(index[indexer] - target) <= tolerance. Tolerance may be a scalar value, which applies the same tolerance to all values, or list-like, which applies variable tolerance per element. List-like must be the same size as the index and its dtype must exactly match the index’s type.

• copy (bool, default: True) – If copy=True, data in the return value is always copied. If copy=False and reindexing is unnecessary, or can be performed with only slice operations, then the output may share memory with the input. In either case, a new xarray object is always returned.

• fill_value (scalar or dict-like, optional) – Value to use for newly missing values. If a dict-like, maps variable names (including coordinates) to fill values. Use this data array’s name to refer to the data array’s values.

Returns:

reindexed – Another dataset array, with this array’s data but coordinates from the other object.

Return type:

DataArray

Examples

>>> data = np.arange(12).reshape(4, 3)
>>> da1 = xr.DataArray(
...     data=data,
...     dims=["x", "y"],
...     coords={"x": [10, 20, 30, 40], "y": [70, 80, 90]},
... )
>>> da1
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) int64 32B 10 20 30 40
  * y        (y) int64 24B 70 80 90
>>> da2 = xr.DataArray(
...     data=data,
...     dims=["x", "y"],
...     coords={"x": [40, 30, 20, 10], "y": [90, 80, 70]},
... )
>>> da2
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 0,  1,  2],
       [ 3,  4,  5],
       [ 6,  7,  8],
       [ 9, 10, 11]])
Coordinates:
  * x        (x) int64 32B 40 30 20 10
  * y        (y) int64 24B 90 80 70

Reindexing with both DataArrays having the same coordinates set, but in different order:

>>> da1.reindex_like(da2)
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[11, 10,  9],
       [ 8,  7,  6],
       [ 5,  4,  3],
       [ 2,  1,  0]])
Coordinates:
  * x        (x) int64 32B 40 30 20 10
  * y        (y) int64 24B 90 80 70

Reindexing with the other array having additional coordinates:

>>> da3 = xr.DataArray(
...     data=data,
...     dims=["x", "y"],
...     coords={"x": [20, 10, 29, 39], "y": [70, 80, 90]},
... )
>>> da1.reindex_like(da3)
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 3.,  4.,  5.],
       [ 0.,  1.,  2.],
       [nan, nan, nan],
       [nan, nan, nan]])
Coordinates:
  * x        (x) int64 32B 20 10 29 39
  * y        (y) int64 24B 70 80 90

Filling missing values with the previous valid index with respect to the coordinates’ value:

>>> da1.reindex_like(da3, method="ffill")
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[3, 4, 5],
       [0, 1, 2],
       [3, 4, 5],
       [6, 7, 8]])
Coordinates:
  * x        (x) int64 32B 20 10 29 39
  * y        (y) int64 24B 70 80 90

Filling missing values while tolerating specified error for inexact matches:

>>> da1.reindex_like(da3, method="ffill", tolerance=5)
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 3.,  4.,  5.],
       [ 0.,  1.,  2.],
       [nan, nan, nan],
       [nan, nan, nan]])
Coordinates:
  * x        (x) int64 32B 20 10 29 39
  * y        (y) int64 24B 70 80 90

Filling missing values with manually specified values:

>>> da1.reindex_like(da3, fill_value=19)
<xarray.DataArray (x: 4, y: 3)> Size: 96B
array([[ 3,  4,  5],
       [ 0,  1,  2],
       [19, 19, 19],
       [19, 19, 19]])
Coordinates:
  * x        (x) int64 32B 20 10 29 39
  * y        (y) int64 24B 70 80 90

Note that unlike broadcast_like, reindex_like doesn’t create new dimensions:

>>> da1.sel(x=20)
<xarray.DataArray (y: 3)> Size: 24B
array([3, 4, 5])
Coordinates:
    x        int64 8B 20
  * y        (y) int64 24B 70 80 90

…so b in not added here:

>>> da1.sel(x=20).reindex_like(da1)
<xarray.DataArray (y: 3)> Size: 24B
array([3, 4, 5])
Coordinates:
    x        int64 8B 20
  * y        (y) int64 24B 70 80 90

See also

DataArray.reindex, DataArray.broadcast_like, align

rename()

Returns a new DataArray with renamed coordinates, dimensions or a new name.

Parameters:

• new_name_or_name_dict (str or dict-like, optional) – If the argument is dict-like, it used as a mapping from old names to new names for coordinates or dimensions. Otherwise, use the argument as the new name for this array.

• **names (Hashable, optional) – The keyword arguments form of a mapping from old names to new names for coordinates or dimensions. One of new_name_or_name_dict or names must be provided.

Returns:

renamed – Renamed array or array with renamed coordinates.

Return type:

DataArray

See also

Dataset.rename, DataArray.swap_dims

reorder_levels()

Rearrange index levels using input order.

Parameters:

• dim_order dict-like of Hashable to int or Hashable (optional) – Mapping from names matching dimensions and values given by lists representing new level orders. Every given dimension must have a multi-index.

• **dim_order_kwargs (optional) – The keyword arguments form of dim_order. One of dim_order or dim_order_kwargs must be provided.

Returns:

obj – Another dataarray, with this dataarray’s data but replaced coordinates.

Return type:

DataArray

resample()

Returns a Resample object for performing resampling operations.

Handles both downsampling and upsampling. The resampled dimension must be a datetime-like coordinate. If any intervals contain no values from the original object, they will be given the value NaN.

Parameters:

• indexer (Mapping of Hashable to str, datetime.timedelta, pd.Timedelta, pd.DateOffset, or Resampler, optional) – Mapping from the dimension name to resample frequency [1]. The dimension must be datetime-like.

• skipna (bool, optional) – Whether to skip missing values when aggregating in downsampling.

• closed ({“left”, “right”}, optional) – Side of each interval to treat as closed.

• label ({“left”, “right”}, optional) – Side of each interval to use for labeling.

• origin ({‘epoch’, ‘start’, ‘start_day’, ‘end’, ‘end_day’}, pd.Timestamp, datetime.datetime, np.datetime64, or cftime.datetime, default ‘start_day’) – The datetime on which to adjust the grouping. The timezone of origin must match the timezone of the index.

  If a datetime is not used, these values are also supported: - ‘epoch’: origin is 1970-01-01 - ‘start’: origin is the first value of the timeseries - ‘start_day’: origin is the first day at midnight of the timeseries - ‘end’: origin is the last value of the timeseries - ‘end_day’: origin is the ceiling midnight of the last day

• offset (pd.Timedelta, datetime.timedelta, or str, default is None) – An offset timedelta added to the origin.

• restore_coord_dims (bool, optional) – If True, also restore the dimension order of multi-dimensional coordinates.

• **indexer_kwargs (str, datetime.timedelta, pd.Timedelta, pd.DateOffset, or Resampler) – The keyword arguments form of indexer. One of indexer or indexer_kwargs must be provided.

Returns:

resampled – This object resampled.

Return type:

core.resample.DataArrayResample

Examples

Downsample monthly time-series data to seasonal data:

>>> da = xr.DataArray(
...     np.linspace(0, 11, num=12),
...     coords=[
...         pd.date_range(
...             "1999-12-15",
...             periods=12,
...             freq=pd.DateOffset(months=1),
...         )
...     ],
...     dims="time",
... )
>>> da
<xarray.DataArray (time: 12)> Size: 96B
array([ 0.,  1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10., 11.])
Coordinates:
  * time     (time) datetime64[ns] 96B 1999-12-15 2000-01-15 ... 2000-11-15
>>> da.resample(time="QS-DEC").mean()
<xarray.DataArray (time: 4)> Size: 32B
array([ 1.,  4.,  7., 10.])
Coordinates:
  * time     (time) datetime64[ns] 32B 1999-12-01 2000-03-01 ... 2000-09-01

Upsample monthly time-series data to daily data:

>>> da.resample(time="1D").interpolate("linear")  # +doctest: ELLIPSIS
<xarray.DataArray (time: 337)> Size: 3kB
array([ 0.        ,  0.03225806,  0.06451613,  0.09677419,  0.12903226,
        0.16129032,  0.19354839,  0.22580645,  0.25806452,  0.29032258,
        0.32258065,  0.35483871,  0.38709677,  0.41935484,  0.4516129 ,
        0.48387097,  0.51612903,  0.5483871 ,  0.58064516,  0.61290323,
        0.64516129,  0.67741935,  0.70967742,  0.74193548,  0.77419355,
        0.80645161,  0.83870968,  0.87096774,  0.90322581,  0.93548387,
        0.96774194,  1.        ,  ...,
        9.        ,  9.03333333,  9.06666667,  9.1       ,  9.13333333,
        9.16666667,  9.2       ,  9.23333333,  9.26666667,  9.3       ,
        9.33333333,  9.36666667,  9.4       ,  9.43333333,  9.46666667,
        9.5       ,  9.53333333,  9.56666667,  9.6       ,  9.63333333,
        9.66666667,  9.7       ,  9.73333333,  9.76666667,  9.8       ,
        9.83333333,  9.86666667,  9.9       ,  9.93333333,  9.96666667,
       10.        , 10.03225806, 10.06451613, 10.09677419, 10.12903226,
       10.16129032, 10.19354839, 10.22580645, 10.25806452, 10.29032258,
       10.32258065, 10.35483871, 10.38709677, 10.41935484, 10.4516129 ,
       10.48387097, 10.51612903, 10.5483871 , 10.58064516, 10.61290323,
       10.64516129, 10.67741935, 10.70967742, 10.74193548, 10.77419355,
       10.80645161, 10.83870968, 10.87096774, 10.90322581, 10.93548387,
       10.96774194, 11.        ])
Coordinates:
  * time     (time) datetime64[ns] 3kB 1999-12-15 1999-12-16 ... 2000-11-15

Limit scope of upsampling method

>>> da.resample(time="1D").nearest(tolerance="1D")
<xarray.DataArray (time: 337)> Size: 3kB
array([ 0.,  0., nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan,  1.,  1.,  1., nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan,  2.,  2.,  2., nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,  3.,
        3.,  3., nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan,  4.,  4.,  4., nan, nan, nan, nan, nan, ...,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, 10., 10., 10., nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, nan,
       nan, nan, nan, nan, nan, nan, nan, nan, nan, nan, 11., 11.])
Coordinates:
  * time     (time) datetime64[ns] 3kB 1999-12-15 1999-12-16 ... 2000-11-15

See also

Dataset.resample, pandas.Series.resample, pandas.DataFrame.resample

References

[1]

https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html#offset-aliases

reset_coords()

Given names of coordinates, reset them to become variables.

Parameters:

• names (str, Iterable of Hashable or None, optional) – Name(s) of non-index coordinates in this dataset to reset into variables. By default, all non-index coordinates are reset.

• drop (bool, default: False) – If True, remove coordinates instead of converting them into variables.

Return type:

Dataset, or DataArray if drop == True

Examples

>>> temperature = np.arange(25).reshape(5, 5)
>>> pressure = np.arange(50, 75).reshape(5, 5)
>>> da = xr.DataArray(
...     data=temperature,
...     dims=["x", "y"],
...     coords=dict(
...         lon=("x", np.arange(10, 15)),
...         lat=("y", np.arange(20, 25)),
...         Pressure=(["x", "y"], pressure),
...     ),
...     name="Temperature",
... )
>>> da
<xarray.DataArray 'Temperature' (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Coordinates:
    lon       (x) int64 40B 10 11 12 13 14
    lat       (y) int64 40B 20 21 22 23 24
    Pressure  (x, y) int64 200B 50 51 52 53 54 55 56 57 ... 68 69 70 71 72 73 74
Dimensions without coordinates: x, y

Return Dataset with target coordinate as a data variable rather than a coordinate variable:

>>> da.reset_coords(names="Pressure")
<xarray.Dataset> Size: 480B
Dimensions:      (x: 5, y: 5)
Coordinates:
    lon          (x) int64 40B 10 11 12 13 14
    lat          (y) int64 40B 20 21 22 23 24
Dimensions without coordinates: x, y
Data variables:
    Pressure     (x, y) int64 200B 50 51 52 53 54 55 56 ... 68 69 70 71 72 73 74
    Temperature  (x, y) int64 200B 0 1 2 3 4 5 6 7 8 ... 17 18 19 20 21 22 23 24

Return DataArray without targeted coordinate:

>>> da.reset_coords(names="Pressure", drop=True)
<xarray.DataArray 'Temperature' (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Coordinates:
    lon      (x) int64 40B 10 11 12 13 14
    lat      (y) int64 40B 20 21 22 23 24
Dimensions without coordinates: x, y

reset_encoding()

reset_index()

Reset the specified index(es) or multi-index level(s).

This legacy method is specific to pandas (multi-)indexes and 1-dimensional “dimension” coordinates. See the more generic drop_indexes() and set_xindex() method to respectively drop and set pandas or custom indexes for arbitrary coordinates.

Parameters:

• dims_or_levels (Hashable or sequence of Hashable) – Name(s) of the dimension(s) and/or multi-index level(s) that will be reset.

• drop (bool, default: False) – If True, remove the specified indexes and/or multi-index levels instead of extracting them as new coordinates (default: False).

Returns:

obj – Another dataarray, with this dataarray’s data but replaced coordinates.

Return type:

DataArray

See also

DataArray.set_index, DataArray.set_xindex, DataArray.drop_indexes

roll()

Roll this array by an offset along one or more dimensions.

Unlike shift, roll treats the given dimensions as periodic, so will not create any missing values to be filled.

Unlike shift, roll may rotate all variables, including coordinates if specified. The direction of rotation is consistent with numpy.roll().

Parameters:

• shifts (mapping of Hashable to int, optional) – Integer offset to rotate each of the given dimensions. Positive offsets roll to the right; negative offsets roll to the left.

• roll_coords (bool, default: False) – Indicates whether to roll the coordinates by the offset too.

• **shifts_kwargs ({dim: offset, …}, optional) – The keyword arguments form of shifts. One of shifts or shifts_kwargs must be provided.

Returns:

rolled – DataArray with the same attributes but rolled data and coordinates.

Return type:

DataArray

See also

shift

Examples

>>> arr = xr.DataArray([5, 6, 7], dims="x")
>>> arr.roll(x=1)
<xarray.DataArray (x: 3)> Size: 24B
array([7, 5, 6])
Dimensions without coordinates: x

rolling()

Rolling window object for DataArrays.

Parameters:

• dim (dict, optional) – Mapping from the dimension name to create the rolling iterator along (e.g. time) to its moving window size.

• min_periods (int or None, default: None) – Minimum number of observations in window required to have a value (otherwise result is NA). The default, None, is equivalent to setting min_periods equal to the size of the window.

• center (bool or Mapping to int, default: False) – Set the labels at the center of the window. The default, False, sets the labels at the right edge of the window.

• **window_kwargs (optional) – The keyword arguments form of dim. One of dim or window_kwargs must be provided.

Return type:

computation.rolling.DataArrayRolling

Examples

Create rolling seasonal average of monthly data e.g. DJF, JFM, …, SON:

>>> da = xr.DataArray(
...     np.linspace(0, 11, num=12),
...     coords=[
...         pd.date_range(
...             "1999-12-15",
...             periods=12,
...             freq=pd.DateOffset(months=1),
...         )
...     ],
...     dims="time",
... )
>>> da
<xarray.DataArray (time: 12)> Size: 96B
array([ 0.,  1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10., 11.])
Coordinates:
  * time     (time) datetime64[ns] 96B 1999-12-15 2000-01-15 ... 2000-11-15
>>> da.rolling(time=3, center=True).mean()
<xarray.DataArray (time: 12)> Size: 96B
array([nan,  1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10., nan])
Coordinates:
  * time     (time) datetime64[ns] 96B 1999-12-15 2000-01-15 ... 2000-11-15

Remove the NaNs using dropna():

>>> da.rolling(time=3, center=True).mean().dropna("time")
<xarray.DataArray (time: 10)> Size: 80B
array([ 1.,  2.,  3.,  4.,  5.,  6.,  7.,  8.,  9., 10.])
Coordinates:
  * time     (time) datetime64[ns] 80B 2000-01-15 2000-02-15 ... 2000-10-15

See also

DataArray.cumulative, Dataset.rolling, computation.rolling.DataArrayRolling

rolling_exp()

Exponentially-weighted moving window. Similar to EWM in pandas

Requires the optional Numbagg dependency.

Parameters:

• window (mapping of hashable to int, optional) – A mapping from the name of the dimension to create the rolling exponential window along (e.g. time) to the size of the moving window.

• window_type ({“span”, “com”, “halflife”, “alpha”}, default: “span”) – The format of the previously supplied window. Each is a simple numerical transformation of the others. Described in detail: https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.ewm.html

• **window_kwargs (optional) – The keyword arguments form of window. One of window or window_kwargs must be provided.

See also

core.rolling_exp.RollingExp

round()

searchsorted(v, side='left', sorter=None)

Find indices where elements of v should be inserted in a to maintain order.

For full documentation, see numpy.searchsorted

See also

numpy.searchsorted

equivalent function

set_close()

Register the function that releases any resources linked to this object.

This method controls how xarray cleans up resources associated with this object when the .close() method is called. It is mostly intended for backend developers and it is rarely needed by regular end-users.

Parameters:

close (callable) – The function that when called like close() releases any resources linked to this object.

set_index()

Set DataArray (multi-)indexes using one or more existing coordinates.

This legacy method is limited to pandas (multi-)indexes and 1-dimensional “dimension” coordinates. See set_xindex() for setting a pandas or a custom Xarray-compatible index from one or more arbitrary coordinates.

Parameters:

• indexes ({dim: index, …}) – Mapping from names matching dimensions and values given by (lists of) the names of existing coordinates or variables to set as new (multi-)index.

• append (bool, default: False) – If True, append the supplied index(es) to the existing index(es). Otherwise replace the existing index(es).

• **indexes_kwargs (optional) – The keyword arguments form of indexes. One of indexes or indexes_kwargs must be provided.

Returns:

obj – Another DataArray, with this data but replaced coordinates.

Return type:

DataArray

Examples

>>> arr = xr.DataArray(
...     data=np.ones((2, 3)),
...     dims=["x", "y"],
...     coords={"x": range(2), "y": range(3), "a": ("x", [3, 4])},
... )
>>> arr
<xarray.DataArray (x: 2, y: 3)> Size: 48B
array([[1., 1., 1.],
       [1., 1., 1.]])
Coordinates:
  * x        (x) int64 16B 0 1
  * y        (y) int64 24B 0 1 2
    a        (x) int64 16B 3 4
>>> arr.set_index(x="a")
<xarray.DataArray (x: 2, y: 3)> Size: 48B
array([[1., 1., 1.],
       [1., 1., 1.]])
Coordinates:
  * x        (x) int64 16B 3 4
  * y        (y) int64 24B 0 1 2

See also

DataArray.reset_index, DataArray.set_xindex

set_xindex()

Set a new, Xarray-compatible index from one or more existing coordinate(s).

Parameters:

• coord_names (str or list) – Name(s) of the coordinate(s) used to build the index. If several names are given, their order matters.

• index_cls (subclass of Index) – The type of index to create. By default, try setting a pandas (multi-)index from the supplied coordinates.

• **options – Options passed to the index constructor.

Returns:

obj – Another dataarray, with this dataarray’s data and with a new index.

Return type:

DataArray

property shape

Tuple of array dimensions.

See also

numpy.ndarray.shape

shift()

Shift this DataArray by an offset along one or more dimensions.

Only the data is moved; coordinates stay in place. This is consistent with the behavior of shift in pandas.

Values shifted from beyond array bounds will appear at one end of each dimension, which are filled according to fill_value. For periodic offsets instead see roll.

Parameters:

• shifts (mapping of Hashable to int or None, optional) – Integer offset to shift along each of the given dimensions. Positive offsets shift to the right; negative offsets shift to the left.

• fill_value (scalar, optional) – Value to use for newly missing values

• **shifts_kwargs – The keyword arguments form of shifts. One of shifts or shifts_kwargs must be provided.

Returns:

shifted – DataArray with the same coordinates and attributes but shifted data.

Return type:

DataArray

See also

roll

Examples

>>> arr = xr.DataArray([5, 6, 7], dims="x")
>>> arr.shift(x=1)
<xarray.DataArray (x: 3)> Size: 24B
array([nan,  5.,  6.])
Dimensions without coordinates: x

property size

Number of elements in the array.

Equal to np.prod(a.shape), i.e., the product of the array’s dimensions.

See also

numpy.ndarray.size

property sizes

Ordered mapping from dimension names to lengths.

Immutable.

See also

Dataset.sizes

sortby()

Sort object by labels or values (along an axis).

Sorts the dataarray, either along specified dimensions, or according to values of 1-D dataarrays that share dimension with calling object.

If the input variables are dataarrays, then the dataarrays are aligned (via left-join) to the calling object prior to sorting by cell values. NaNs are sorted to the end, following Numpy convention.

If multiple sorts along the same dimension is given, numpy’s lexsort is performed along that dimension: https://numpy.org/doc/stable/reference/generated/numpy.lexsort.html and the FIRST key in the sequence is used as the primary sort key, followed by the 2nd key, etc.

Parameters:

• variables (Hashable, DataArray, sequence of Hashable or DataArray, or Callable) – 1D DataArray objects or name(s) of 1D variable(s) in coords whose values are used to sort this array. If a callable, the callable is passed this object, and the result is used as the value for cond.

• ascending (bool, default: True) – Whether to sort by ascending or descending order.

Returns:

sorted – A new dataarray where all the specified dims are sorted by dim labels.

Return type:

DataArray

See also

Dataset.sortby, numpy.sort, pandas.sort_values, pandas.sort_index

Examples

>>> da = xr.DataArray(
...     np.arange(5, 0, -1),
...     coords=[pd.date_range("1/1/2000", periods=5)],
...     dims="time",
... )
>>> da
<xarray.DataArray (time: 5)> Size: 40B
array([5, 4, 3, 2, 1])
Coordinates:
  * time     (time) datetime64[ns] 40B 2000-01-01 2000-01-02 ... 2000-01-05

>>> da.sortby(da)
<xarray.DataArray (time: 5)> Size: 40B
array([1, 2, 3, 4, 5])
Coordinates:
  * time     (time) datetime64[ns] 40B 2000-01-05 2000-01-04 ... 2000-01-01

>>> da.sortby(lambda x: x)
<xarray.DataArray (time: 5)> Size: 40B
array([1, 2, 3, 4, 5])
Coordinates:
  * time     (time) datetime64[ns] 40B 2000-01-05 2000-01-04 ... 2000-01-01

squeeze()

Return a new object with squeezed data.

Parameters:

• dim (None or Hashable or iterable of Hashable, optional) – Selects a subset of the length one dimensions. If a dimension is selected with length greater than one, an error is raised. If None, all length one dimensions are squeezed.

• drop (bool, default: False) – If drop=True, drop squeezed coordinates instead of making them scalar.

• axis (None or int or iterable of int, optional) – Like dim, but positional.

Returns:

squeezed – This object, but with with all or a subset of the dimensions of length 1 removed.

Return type:

same type as caller

See also

numpy.squeeze

stack()

Stack any number of existing dimensions into a single new dimension.

New dimensions will be added at the end, and the corresponding coordinate variables will be combined into a MultiIndex.

Parameters:

• dim (mapping of Hashable to sequence of Hashable) – Mapping of the form new_name=(dim1, dim2, …). Names of new dimensions, and the existing dimensions that they replace. An ellipsis (…) will be replaced by all unlisted dimensions. Passing a list containing an ellipsis (stacked_dim=[…]) will stack over all dimensions.

• create_index (bool or None, default: True) – If True, create a multi-index for each of the stacked dimensions. If False, don’t create any index. If None, create a multi-index only if exactly one single (1-d) coordinate index is found for every dimension to stack.

• index_cls (class, optional) – Can be used to pass a custom multi-index type. Must be an Xarray index that implements .stack(). By default, a pandas multi-index wrapper is used.

• **dim_kwargs – The keyword arguments form of dim. One of dim or dim_kwargs must be provided.

Returns:

stacked – DataArray with stacked data.

Return type:

DataArray

Examples

>>> arr = xr.DataArray(
...     np.arange(6).reshape(2, 3),
...     coords=[("x", ["a", "b"]), ("y", [0, 1, 2])],
... )
>>> arr
<xarray.DataArray (x: 2, y: 3)> Size: 48B
array([[0, 1, 2],
       [3, 4, 5]])
Coordinates:
  * x        (x) <U1 8B 'a' 'b'
  * y        (y) int64 24B 0 1 2
>>> stacked = arr.stack(z=("x", "y"))
>>> stacked.indexes["z"]
MultiIndex([('a', 0),
            ('a', 1),
            ('a', 2),
            ('b', 0),
            ('b', 1),
            ('b', 2)],
           name='z')

See also

DataArray.unstack

std()

Reduce this DataArray’s data by applying std along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply std. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• ddof (int, default: 0) – “Delta Degrees of Freedom”: the divisor used in the calculation is N - ddof, where N represents the number of elements.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating std on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with std applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.std, dask.array.std, Dataset.std

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.std()
<xarray.DataArray ()> Size: 8B
array(1.0198039)

Use skipna to control whether NaNs are ignored.

>>> da.std(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

Specify ddof=1 for an unbiased estimate.

>>> da.std(skipna=True, ddof=1)
<xarray.DataArray ()> Size: 8B
array(1.14017543)

str

alias of StringAccessor[DataArray]

sum()

Reduce this DataArray’s data by applying sum along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply sum. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• min_count (int or None, optional) – The required number of valid values to perform the operation. If fewer than min_count non-NA values are present the result will be NA. Only used if skipna is set to True or defaults to True for the array’s dtype. Changed in version 0.17.0: if specified on an integer array and skipna=True, the result will be a float array.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating sum on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with sum applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.sum, dask.array.sum, Dataset.sum

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.sum()
<xarray.DataArray ()> Size: 8B
array(8.)

Use skipna to control whether NaNs are ignored.

>>> da.sum(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

Specify min_count for finer control over when NaNs are ignored.

>>> da.sum(skipna=True, min_count=2)
<xarray.DataArray ()> Size: 8B
array(8.)

swap_dims()

Returns a new DataArray with swapped dimensions.

Parameters:

• dims_dict (dict-like) – Dictionary whose keys are current dimension names and whose values are new names.

• **dims_kwargs ({existing_dim: new_dim, …}, optional) – The keyword arguments form of dims_dict. One of dims_dict or dims_kwargs must be provided.

Returns:

swapped – DataArray with swapped dimensions.

Return type:

DataArray

Examples

>>> arr = xr.DataArray(
...     data=[0, 1],
...     dims="x",
...     coords={"x": ["a", "b"], "y": ("x", [0, 1])},
... )
>>> arr
<xarray.DataArray (x: 2)> Size: 16B
array([0, 1])
Coordinates:
  * x        (x) <U1 8B 'a' 'b'
    y        (x) int64 16B 0 1

>>> arr.swap_dims({"x": "y"})
<xarray.DataArray (y: 2)> Size: 16B
array([0, 1])
Coordinates:
    x        (y) <U1 8B 'a' 'b'
  * y        (y) int64 16B 0 1

>>> arr.swap_dims({"x": "z"})
<xarray.DataArray (z: 2)> Size: 16B
array([0, 1])
Coordinates:
    x        (z) <U1 8B 'a' 'b'
    y        (z) int64 16B 0 1
Dimensions without coordinates: z

See also

DataArray.rename, Dataset.swap_dims

tail()

Return a new DataArray whose data is given by the the last n values along the specified dimension(s). Default n = 5

See also

Dataset.tail, DataArray.head, DataArray.thin

Examples

>>> da = xr.DataArray(
...     np.arange(25).reshape(5, 5),
...     dims=("x", "y"),
... )
>>> da
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Dimensions without coordinates: x, y

>>> da.tail(y=1)
<xarray.DataArray (x: 5, y: 1)> Size: 40B
array([[ 4],
       [ 9],
       [14],
       [19],
       [24]])
Dimensions without coordinates: x, y

>>> da.tail({"x": 2, "y": 2})
<xarray.DataArray (x: 2, y: 2)> Size: 32B
array([[18, 19],
       [23, 24]])
Dimensions without coordinates: x, y

thin()

Return a new DataArray whose data is given by each n value along the specified dimension(s).

Examples

>>> x_arr = np.arange(0, 26)
>>> x_arr
array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
       17, 18, 19, 20, 21, 22, 23, 24, 25])
>>> x = xr.DataArray(
...     np.reshape(x_arr, (2, 13)),
...     dims=("x", "y"),
...     coords={"x": [0, 1], "y": np.arange(0, 13)},
... )
>>> x
<xarray.DataArray (x: 2, y: 13)> Size: 208B
array([[ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12],
       [13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25]])
Coordinates:
  * x        (x) int64 16B 0 1
  * y        (y) int64 104B 0 1 2 3 4 5 6 7 8 9 10 11 12

>>>
>>> x.thin(3)
<xarray.DataArray (x: 1, y: 5)> Size: 40B
array([[ 0,  3,  6,  9, 12]])
Coordinates:
  * x        (x) int64 8B 0
  * y        (y) int64 40B 0 3 6 9 12
>>> x.thin({"x": 2, "y": 5})
<xarray.DataArray (x: 1, y: 3)> Size: 24B
array([[ 0,  5, 10]])
Coordinates:
  * x        (x) int64 8B 0
  * y        (y) int64 24B 0 5 10

See also

Dataset.thin, DataArray.head, DataArray.tail

to_dask_dataframe()

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.

Return type:

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

to_dataframe()

Convert this array and its coordinates into a tidy pandas.DataFrame.

The DataFrame is indexed by the Cartesian product of index coordinates (in the form of a pandas.MultiIndex). Other coordinates are included as columns in the DataFrame.

For 1D and 2D DataArrays, see also DataArray.to_pandas() which doesn’t rely on a MultiIndex to build the DataFrame.

Parameters:

• name (Hashable or None, optional) – Name to give to this array (required if unnamed).

• 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 dataframe.

  If provided, must include all dimensions of this DataArray. By default, dimensions are sorted according to the DataArray dimensions order.

Returns:

result – DataArray as a pandas DataFrame.

Return type:

DataFrame

See also

DataArray.to_pandas, DataArray.to_series

to_dataset()

Convert a DataArray to a Dataset.

Parameters:

• dim (Hashable, optional) – Name of the dimension on this array along which to split this array into separate variables. If not provided, this array is converted into a Dataset of one variable.

• name (Hashable, optional) – Name to substitute for this array’s name. Only valid if dim is not provided.

• promote_attrs (bool, default: False) – Set to True to shallow copy attrs of DataArray to returned Dataset.

Returns:

dataset

Return type:

Dataset

to_dict()

Convert this xarray.DataArray into a dictionary following xarray naming conventions.

Converts all variables and attributes to native Python objects. Useful for converting to json. To avoid datetime incompatibility use decode_times=False kwarg in xarray.open_dataset.

Parameters:

• data (bool or {“list”, “array”}, default: “list”) – Whether to include the actual data in the dictionary. When set to False, returns just the schema. If set to “array”, returns data as underlying array type. If set to “list” (or True for backwards compatibility), returns data in lists of Python data types. Note that for obtaining the “list” output efficiently, use da.compute().to_dict(data=”list”).

• encoding (bool, default: False) – Whether to include the Dataset’s encoding in the dictionary.

Returns:

dict

Return type:

dict

See also

DataArray.from_dict, Dataset.to_dict

to_index()

Convert this variable to a pandas.Index. Only possible for 1D arrays.

to_iris()

Convert this array into a iris.cube.Cube

to_masked_array()

Convert this array into a numpy.ma.MaskedArray

Parameters:

copy (bool, default: True) – If True make a copy of the array in the result. If False, a MaskedArray view of DataArray.values is returned.

Returns:

result – Masked where invalid values (nan or inf) occur.

Return type:

MaskedArray

to_netcdf()

Write DataArray contents to a netCDF file.

Parameters:

• path (str, path-like, file-like or None, optional) – Path to which to save this datatree, or a file-like object to write it to (which must support read and write and be seekable) or None (default) to return in-memory bytes as a memoryview.

• mode ({“w”, “a”}, default: “w”) – Write (‘w’) or append (‘a’) mode. If mode=’w’, any existing file at this location will be overwritten. If mode=’a’, existing variables will be overwritten.

• format ({“NETCDF4”, “NETCDF4_CLASSIC”, “NETCDF3_64BIT”, “NETCDF3_CLASSIC”}, optional) – File format for the resulting netCDF file:

  • NETCDF4: Data is stored in an HDF5 file, using netCDF4 API features.

  • NETCDF4_CLASSIC: Data is stored in an HDF5 file, using only netCDF 3 compatible API features.

  • NETCDF3_64BIT: 64-bit offset version of the netCDF 3 file format, which fully supports 2+ GB files, but is only compatible with clients linked against netCDF version 3.6.0 or later.

  • NETCDF3_CLASSIC: The classic netCDF 3 file format. It does not handle 2+ GB files very well.

  All formats are supported by the netCDF4-python library. scipy.io.netcdf only supports the last two formats.

  The default format is NETCDF4 if you are saving a file to disk and have the netCDF4-python library available. Otherwise, xarray falls back to using scipy to write netCDF files and defaults to the NETCDF3_64BIT format (scipy does not support netCDF4).

• group (str, optional) – Path to the netCDF4 group in the given file to open (only works for format=’NETCDF4’). The group(s) will be created if necessary.

• engine ({“netcdf4”, “scipy”, “h5netcdf”}, optional) – Engine to use when writing netCDF files. If not provided, the default engine is chosen based on available dependencies, with a preference for ‘netcdf4’ if writing to a file on disk.

• encoding (dict, optional) – Nested dictionary with variable names as keys and dictionaries of variable specific encodings as values, e.g., {"my_variable": {"dtype": "int16", "scale_factor": 0.1, "zlib": True}, ...}

  The h5netcdf engine supports both the NetCDF4-style compression encoding parameters {"zlib": True, "complevel": 9} and the h5py ones {"compression": "gzip", "compression_opts": 9}. This allows using any compression plugin installed in the HDF5 library, e.g. LZF.

• unlimited_dims (iterable of Hashable, optional) – Dimension(s) that should be serialized as unlimited dimensions. By default, no dimensions are treated as unlimited dimensions. Note that unlimited_dims may also be set via dataset.encoding["unlimited_dims"].

• compute (bool, default: True) – If true compute immediately, otherwise return a dask.delayed.Delayed object that can be computed later.

• invalid_netcdf (bool, default: False) – Only valid along with engine="h5netcdf". If True, allow writing hdf5 files which are invalid netcdf as described in h5netcdf/h5netcdf.

Returns:

• * memoryview if path is None

• * dask.delayed.Delayed if compute is False

• * None otherwise

Notes

Only xarray.Dataset objects can be written to netCDF files, so the xarray.DataArray is converted to a xarray.Dataset object containing a single variable. If the DataArray has no name, or if the name is the same as a coordinate name, then it is given the name "__xarray_dataarray_variable__".

[netCDF4 backend only] netCDF4 enums are decoded into the dataarray dtype metadata.

See also

Dataset.to_netcdf

to_numpy()

Coerces wrapped data to numpy and returns a numpy.ndarray.

See also

DataArray.as_numpy

Same but returns the surrounding DataArray instead.

Dataset.as_numpy, DataArray.values, DataArray.data

to_pandas()

Convert this array into a pandas object with the same shape.

The type of the returned object depends on the number of DataArray dimensions:

• 0D -> xarray.DataArray

• 1D -> pandas.Series

• 2D -> pandas.DataFrame

Only works for arrays with 2 or fewer dimensions.

The DataArray constructor performs the inverse transformation.

Returns:

result – DataArray, pandas Series or pandas DataFrame.

Return type:

DataArray | Series | DataFrame

to_series()

Convert this array into a pandas.Series.

The Series is indexed by the Cartesian product of index coordinates (in the form of a pandas.MultiIndex).

Returns:

result – DataArray as a pandas Series.

Return type:

Series

See also

DataArray.to_pandas, DataArray.to_dataframe

to_unstacked_dataset()

Unstack DataArray expanding to Dataset along a given level of a stacked coordinate.

This is the inverse operation of Dataset.to_stacked_array.

Parameters:

• dim (Hashable) – Name of existing dimension to unstack

• level (int or Hashable, default: 0) – The MultiIndex level to expand to a dataset along. Can either be the integer index of the level or its name.

Returns:

unstacked

Return type:

Dataset

Examples

>>> arr = xr.DataArray(
...     np.arange(6).reshape(2, 3),
...     coords=[("x", ["a", "b"]), ("y", [0, 1, 2])],
... )
>>> data = xr.Dataset({"a": arr, "b": arr.isel(y=0)})
>>> data
<xarray.Dataset> Size: 96B
Dimensions:  (x: 2, y: 3)
Coordinates:
  * x        (x) <U1 8B 'a' 'b'
  * y        (y) int64 24B 0 1 2
Data variables:
    a        (x, y) int64 48B 0 1 2 3 4 5
    b        (x) int64 16B 0 3
>>> stacked = data.to_stacked_array("z", ["x"])
>>> stacked.indexes["z"]
MultiIndex([('a',   0),
            ('a',   1),
            ('a',   2),
            ('b', nan)],
           name='z')
>>> roundtripped = stacked.to_unstacked_dataset(dim="z")
>>> data.identical(roundtripped)
True

See also

Dataset.to_stacked_array

to_zarr()

Write DataArray contents to a Zarr store

Zarr chunks are determined in the following way:

• From the chunks attribute in each variable’s encoding (can be set via DataArray.chunk).

• If the variable is a Dask array, from the dask chunks

• If neither Dask chunks nor encoding chunks are present, chunks will be determined automatically by Zarr

• If both Dask chunks and encoding chunks are present, encoding chunks will be used, provided that there is a many-to-one relationship between encoding chunks and dask chunks (i.e. Dask chunks are bigger than and evenly divide encoding chunks); otherwise raise a ValueError. This restriction ensures that no synchronization / locks are required when writing. To disable this restriction, use safe_chunks=False.

Parameters:

• store (MutableMapping, str or path-like, optional) – Store or path to directory in local or remote file system.

• chunk_store (MutableMapping, str or path-like, optional) – Store or path to directory in local or remote file system only for Zarr array chunks. Requires zarr-python v2.4.0 or later.

• mode ({“w”, “w-”, “a”, “a-”, r+”, None}, optional) – Persistence mode: “w” means create (overwrite if exists); “w-” means create (fail if exists); “a” means override all existing variables including dimension coordinates (create if does not exist); “a-” means only append those variables that have append_dim. “r+” means modify existing array values only (raise an error if any metadata or shapes would change). The default mode is “a” if append_dim is set. Otherwise, it is “r+” if region is set and w- otherwise.

• synchronizer (object, optional) – Zarr array synchronizer.

• group (str, optional) – Group path. (a.k.a. path in zarr terminology.)

• encoding (dict, optional) – Nested dictionary with variable names as keys and dictionaries of variable specific encodings as values, e.g., {"my_variable": {"dtype": "int16", "scale_factor": 0.1,}, ...}

• compute (bool, default: True) – If True write array data immediately, otherwise return a dask.delayed.Delayed object that can be computed to write array data later. Metadata is always updated eagerly.

• consolidated (bool, optional) – If True, apply zarr’s consolidate_metadata function to the store after writing metadata and read existing stores with consolidated metadata; if False, do not. The default (consolidated=None) means write consolidated metadata and attempt to read consolidated metadata for existing stores (falling back to non-consolidated).

  When the experimental zarr_version=3, consolidated must be either be None or False.

• append_dim (hashable, optional) – If set, the dimension along which the data will be appended. All other dimensions on overridden variables must remain the same size.

• region (dict, optional) – Optional mapping from dimension names to integer slices along dataarray dimensions to indicate the region of existing zarr array(s) in which to write this datarray’s data. For example, {'x': slice(0, 1000), 'y': slice(10000, 11000)} would indicate that values should be written to the region 0:1000 along x and 10000:11000 along y.

  Two restrictions apply to the use of region:

  • If region is set, _all_ variables in a dataarray must have at least one dimension in common with the region. Other variables should be written in a separate call to to_zarr().

  • Dimensions cannot be included in both region and append_dim at the same time. To create empty arrays to fill in with region, use a separate call to to_zarr() with compute=False. See “Modifying existing Zarr stores” in the reference documentation for full details.

  Users are expected to ensure that the specified region aligns with Zarr chunk boundaries, and that dask chunks are also aligned. Xarray makes limited checks that these multiple chunk boundaries line up. It is possible to write incomplete chunks and corrupt the data with this option if you are not careful.

• safe_chunks (bool, default: True) – If True, only allow writes to when there is a many-to-one relationship between Zarr chunks (specified in encoding) and Dask chunks. Set False to override this restriction; however, data may become corrupted if Zarr arrays are written in parallel. This option may be useful in combination with compute=False to initialize a Zarr store from an existing DataArray with arbitrary chunk structure. In addition to the many-to-one relationship validation, it also detects partial chunks writes when using the region parameter, these partial chunks are considered unsafe in the mode “r+” but safe in the mode “a”. Note: Even with these validations it can still be unsafe to write two or more chunked arrays in the same location in parallel if they are not writing in independent regions, for those cases it is better to use a synchronizer.

• align_chunks (bool, default False) – If True, rechunks the Dask array to align with Zarr chunks before writing. This ensures each Dask chunk maps to one or more contiguous Zarr chunks, which avoids race conditions. Internally, the process sets safe_chunks=False and tries to preserve the original Dask chunking as much as possible. Note: While this alignment avoids write conflicts stemming from chunk boundary misalignment, it does not protect against race conditions if multiple uncoordinated processes write to the same Zarr array concurrently.

• storage_options (dict, optional) – Any additional parameters for the storage backend (ignored for local paths).

• zarr_version (int or None, optional) – .. deprecated:: 2024.9.1 Use zarr_format instead.

• zarr_format (int or None, optional) – The desired zarr format to target (currently 2 or 3). The default of None will attempt to determine the zarr version from store when possible, otherwise defaulting to the default version used by the zarr-python library installed.

• write_empty_chunks (bool or None, optional) – If True, all chunks will be stored regardless of their contents. If False, each chunk is compared to the array’s fill value prior to storing. If a chunk is uniformly equal to the fill value, then that chunk is not be stored, and the store entry for that chunk’s key is deleted. This setting enables sparser storage, as only chunks with non-fill-value data are stored, at the expense of overhead associated with checking the data of each chunk. If None (default) fall back to specification(s) in encoding or Zarr defaults. A ValueError will be raised if the value of this (if not None) differs with encoding.

• chunkmanager_store_kwargs (dict, optional) – Additional keyword arguments passed on to the ChunkManager.store method used to store chunked arrays. For example for a dask array additional kwargs will be passed eventually to dask.array.store(). Experimental API that should not be relied upon.

Returns:

• * dask.delayed.Delayed if compute is False

• * ZarrStore otherwise

References

https://zarr.readthedocs.io/

Notes

Zarr chunking behavior:

If chunks are found in the encoding argument or attribute corresponding to any DataArray, those chunks are used. If a DataArray is a dask array, it is written with those chunks. If not other chunks are found, Zarr uses its own heuristics to choose automatic chunk sizes.

encoding:

The encoding attribute (if exists) of the DataArray(s) will be used. Override any existing encodings by providing the encoding kwarg.

fill_value handling:

There exists a subtlety in interpreting zarr’s fill_value property. For zarr v2 format arrays, fill_value is always interpreted as an invalid value similar to the _FillValue attribute in CF/netCDF. For Zarr v3 format arrays, only an explicit _FillValue attribute will be used to mask the data if requested using mask_and_scale=True. See this Github issue for more.

See also

Dataset.to_zarr

Zarr

The I/O user guide, with more details and examples.

transpose()

Return a new DataArray object with transposed dimensions.

Parameters:

• *dim (Hashable, optional) – By default, reverse the dimensions. Otherwise, reorder the dimensions to this order.

• transpose_coords (bool, default: True) – If True, also transpose the coordinates of this DataArray.

• missing_dims ({“raise”, “warn”, “ignore”}, default: “raise”) – What to do if dimensions that should be selected from are not present in the DataArray: - “raise”: raise an exception - “warn”: raise a warning, and ignore the missing dimensions - “ignore”: ignore the missing dimensions

Returns:

transposed – The returned DataArray’s array is transposed.

Return type:

DataArray

Notes

This operation returns a view of this array’s data. It is lazy for dask-backed DataArrays but not for numpy-backed DataArrays – the data will be fully loaded.

See also

numpy.transpose, Dataset.transpose

unify_chunks()

Unify chunk size along all chunked dimensions of this DataArray.

Return type:

DataArray with consistent chunk sizes for all dask-array variables

See also

dask.array.core.unify_chunks

unstack()

Unstack existing dimensions corresponding to MultiIndexes into multiple new dimensions.

New dimensions will be added at the end.

Parameters:

• dim (str, Iterable of Hashable or None, optional) – Dimension(s) over which to unstack. By default unstacks all MultiIndexes.

• fill_value (scalar or dict-like, default: nan) – Value to be filled. If a dict-like, maps variable names to fill values. Use the data array’s name to refer to its name. If not provided or if the dict-like does not contain all variables, the dtype’s NA value will be used.

• sparse (bool, default: False) – Use sparse-array if True

Returns:

unstacked – Array with unstacked data.

Return type:

DataArray

Examples

>>> arr = xr.DataArray(
...     np.arange(6).reshape(2, 3),
...     coords=[("x", ["a", "b"]), ("y", [0, 1, 2])],
... )
>>> arr
<xarray.DataArray (x: 2, y: 3)> Size: 48B
array([[0, 1, 2],
       [3, 4, 5]])
Coordinates:
  * x        (x) <U1 8B 'a' 'b'
  * y        (y) int64 24B 0 1 2
>>> stacked = arr.stack(z=("x", "y"))
>>> stacked.indexes["z"]
MultiIndex([('a', 0),
            ('a', 1),
            ('a', 2),
            ('b', 0),
            ('b', 1),
            ('b', 2)],
           name='z')
>>> roundtripped = stacked.unstack()
>>> arr.identical(roundtripped)
True

See also

DataArray.stack

property values

The array’s data converted to numpy.ndarray.

This will attempt to convert the array naively using np.array(), which will raise an error if the array type does not support coercion like this (e.g. cupy).

Note that this array is not copied; operations on it follow numpy’s rules of what generates a view vs. a copy, and changes to this array may be reflected in the DataArray as well.

var()

Reduce this DataArray’s data by applying var along some dimension(s).

Parameters:

• dim (str, Iterable of Hashable, “…” or None, default: None) – Name of dimension[s] along which to apply var. For e.g. dim="x" or dim=["x", "y"]. If “…” or None, will reduce over all dimensions.

• skipna (bool or None, optional) – If True, skip missing values (as marked by NaN). By default, only skips missing values for float dtypes; other dtypes either do not have a sentinel missing value (int) or skipna=True has not been implemented (object, datetime64 or timedelta64).

• ddof (int, default: 0) – “Delta Degrees of Freedom”: the divisor used in the calculation is N - ddof, where N represents the number of elements.

• keep_attrs (bool or None, optional) – If True, attrs will be copied from the original object to the new one. If False, the new object will be returned without attributes.

• **kwargs (Any) – Additional keyword arguments passed on to the appropriate array function for calculating var on this object’s data. These could include dask-specific kwargs like split_every.

Returns:

reduced – New DataArray with var applied to its data and the indicated dimension(s) removed

Return type:

DataArray

See also

numpy.var, dask.array.var, Dataset.var

Aggregation

User guide on reduction or aggregation operations.

Notes

Non-numeric variables will be removed prior to reducing.

Examples

>>> da = xr.DataArray(
...     np.array([1, 2, 3, 0, 2, np.nan]),
...     dims="time",
...     coords=dict(
...         time=("time", pd.date_range("2001-01-01", freq="ME", periods=6)),
...         labels=("time", np.array(["a", "b", "c", "c", "b", "a"])),
...     ),
... )
>>> da
<xarray.DataArray (time: 6)> Size: 48B
array([ 1.,  2.,  3.,  0.,  2., nan])
Coordinates:
  * time     (time) datetime64[ns] 48B 2001-01-31 2001-02-28 ... 2001-06-30
    labels   (time) <U1 24B 'a' 'b' 'c' 'c' 'b' 'a'

>>> da.var()
<xarray.DataArray ()> Size: 8B
array(1.04)

Use skipna to control whether NaNs are ignored.

>>> da.var(skipna=False)
<xarray.DataArray ()> Size: 8B
array(nan)

Specify ddof=1 for an unbiased estimate.

>>> da.var(skipna=True, ddof=1)
<xarray.DataArray ()> Size: 8B
array(1.3)

property variable

Low level interface to the Variable object for this DataArray.

where()

Filter elements from this object according to a condition.

Returns elements from ‘DataArray’, where ‘cond’ is True, otherwise fill in ‘other’.

This operation follows the normal broadcasting and alignment rules that xarray uses for binary arithmetic.

Parameters:

• cond (DataArray, Dataset, or callable) – Locations at which to preserve this object’s values. dtype must be bool. If a callable, the callable is passed this object, and the result is used as the value for cond.

• other (scalar, DataArray, Dataset, or callable, optional) – Value to use for locations in this object where cond is False. By default, these locations are filled with NA. If a callable, it must expect this object as its only parameter.

• drop (bool, default: False) – If True, coordinate labels that only correspond to False values of the condition are dropped from the result.

Returns:

Same xarray type as caller, with dtype float64.

Return type:

DataArray or Dataset

Examples

>>> a = xr.DataArray(np.arange(25).reshape(5, 5), dims=("x", "y"))
>>> a
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8,  9],
       [10, 11, 12, 13, 14],
       [15, 16, 17, 18, 19],
       [20, 21, 22, 23, 24]])
Dimensions without coordinates: x, y

>>> a.where(a.x + a.y < 4)
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0.,  1.,  2.,  3., nan],
       [ 5.,  6.,  7., nan, nan],
       [10., 11., nan, nan, nan],
       [15., nan, nan, nan, nan],
       [nan, nan, nan, nan, nan]])
Dimensions without coordinates: x, y

>>> a.where(a.x + a.y < 5, -1)
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[ 0,  1,  2,  3,  4],
       [ 5,  6,  7,  8, -1],
       [10, 11, 12, -1, -1],
       [15, 16, -1, -1, -1],
       [20, -1, -1, -1, -1]])
Dimensions without coordinates: x, y

>>> a.where(a.x + a.y < 4, drop=True)
<xarray.DataArray (x: 4, y: 4)> Size: 128B
array([[ 0.,  1.,  2.,  3.],
       [ 5.,  6.,  7., nan],
       [10., 11., nan, nan],
       [15., nan, nan, nan]])
Dimensions without coordinates: x, y

>>> a.where(lambda x: x.x + x.y < 4, lambda x: -x)
<xarray.DataArray (x: 5, y: 5)> Size: 200B
array([[  0,   1,   2,   3,  -4],
       [  5,   6,   7,  -8,  -9],
       [ 10,  11, -12, -13, -14],
       [ 15, -16, -17, -18, -19],
       [-20, -21, -22, -23, -24]])
Dimensions without coordinates: x, y

>>> a.where(a.x + a.y < 4, drop=True)
<xarray.DataArray (x: 4, y: 4)> Size: 128B
array([[ 0.,  1.,  2.,  3.],
       [ 5.,  6.,  7., nan],
       [10., 11., nan, nan],
       [15., nan, nan, nan]])
Dimensions without coordinates: x, y

See also

numpy.where

corresponding numpy function

where

equivalent function

property xindexes

Mapping of Index objects used for label based indexing.