from __future__ import annotations import datetime import warnings from os import PathLike from typing import ( TYPE_CHECKING, Any, Callable, Hashable, Iterable, Literal, Mapping, NoReturn, Sequence, cast, overload, ) import numpy as np import pandas as pd from xarray.coding.calendar_ops import convert_calendar, interp_calendar from xarray.coding.cftimeindex import CFTimeIndex from xarray.core import alignment, computation, dtypes, indexing, ops, utils from xarray.core._aggregations import DataArrayAggregations from xarray.core.accessor_dt import CombinedDatetimelikeAccessor from xarray.core.accessor_str import StringAccessor from xarray.core.alignment import ( _broadcast_helper, _get_broadcast_dims_map_common_coords, align, ) from xarray.core.arithmetic import DataArrayArithmetic from xarray.core.common import AbstractArray, DataWithCoords, get_chunksizes from xarray.core.computation import unify_chunks from xarray.core.coordinates import DataArrayCoordinates, assert_coordinate_consistent from xarray.core.dataset import Dataset from xarray.core.formatting import format_item from xarray.core.indexes import ( Index, Indexes, PandasMultiIndex, filter_indexes_from_coords, isel_indexes, ) from xarray.core.indexing import is_fancy_indexer, map_index_queries from xarray.core.merge import PANDAS_TYPES, MergeError, _create_indexes_from_coords from xarray.core.options import OPTIONS, _get_keep_attrs from xarray.core.utils import ( Default, HybridMappingProxy, ReprObject, _default, either_dict_or_kwargs, ) from xarray.core.variable import ( IndexVariable, Variable, as_compatible_data, as_variable, ) from xarray.plot.accessor import DataArrayPlotAccessor from xarray.plot.utils import _get_units_from_attrs if TYPE_CHECKING: from typing import TypeVar, Union from numpy.typing import ArrayLike try: from dask.delayed import Delayed except ImportError: Delayed = None # type: ignore try: from cdms2 import Variable as cdms2_Variable except ImportError: cdms2_Variable = None try: from iris.cube import Cube as iris_Cube except ImportError: iris_Cube = None from xarray.backends.api import T_NetcdfEngine, T_NetcdfTypes from xarray.core.groupby import DataArrayGroupBy from xarray.core.resample import DataArrayResample from xarray.core.rolling import DataArrayCoarsen, DataArrayRolling from xarray.core.types import ( CoarsenBoundaryOptions, DatetimeLike, DatetimeUnitOptions, Dims, ErrorOptions, ErrorOptionsWithWarn, InterpOptions, PadModeOptions, PadReflectOptions, QuantileMethods, QueryEngineOptions, QueryParserOptions, ReindexMethodOptions, SideOptions, T_DataArray, T_Xarray, ) from xarray.core.weighted import DataArrayWeighted T_XarrayOther = TypeVar("T_XarrayOther", bound=Union["DataArray", Dataset]) def _infer_coords_and_dims( shape, coords, dims ) -> tuple[dict[Hashable, Variable], tuple[Hashable, ...]]: """All the logic for creating a new DataArray""" if ( coords is not None and not utils.is_dict_like(coords) and len(coords) != len(shape) ): raise ValueError( f"coords is not dict-like, but it has {len(coords)} items, " f"which does not match the {len(shape)} dimensions of the " "data" ) if isinstance(dims, str): dims = (dims,) if dims is None: dims = [f"dim_{n}" for n in range(len(shape))] if coords is not None and len(coords) == len(shape): # try to infer dimensions from coords if utils.is_dict_like(coords): dims = list(coords.keys()) else: for n, (dim, coord) in enumerate(zip(dims, coords)): coord = as_variable(coord, name=dims[n]).to_index_variable() dims[n] = coord.name dims = tuple(dims) elif len(dims) != len(shape): raise ValueError( "different number of dimensions on data " f"and dims: {len(shape)} vs {len(dims)}" ) else: for d in dims: if not isinstance(d, str): raise TypeError(f"dimension {d} is not a string") new_coords: dict[Hashable, Variable] = {} if utils.is_dict_like(coords): for k, v in coords.items(): new_coords[k] = as_variable(v, name=k) elif coords is not None: for dim, coord in zip(dims, coords): var = as_variable(coord, name=dim) var.dims = (dim,) new_coords[dim] = var.to_index_variable() sizes = dict(zip(dims, shape)) for k, v in new_coords.items(): if any(d not in dims for d in v.dims): raise ValueError( f"coordinate {k} has dimensions {v.dims}, but these " "are not a subset of the DataArray " f"dimensions {dims}" ) for d, s in zip(v.dims, v.shape): if s != sizes[d]: raise ValueError( f"conflicting sizes for dimension {d!r}: " f"length {sizes[d]} on the data but length {s} on " f"coordinate {k!r}" ) if k in sizes and v.shape != (sizes[k],): raise ValueError( f"coordinate {k!r} is a DataArray dimension, but " f"it has shape {v.shape!r} rather than expected shape {sizes[k]!r} " "matching the dimension size" ) return new_coords, dims def _check_data_shape(data, coords, dims): if data is dtypes.NA: data = np.nan if coords is not None and utils.is_scalar(data, include_0d=False): if utils.is_dict_like(coords): if dims is None: return data else: data_shape = tuple( as_variable(coords[k], k).size if k in coords.keys() else 1 for k in dims ) else: data_shape = tuple(as_variable(coord, "foo").size for coord in coords) data = np.full(data_shape, data) return data class _LocIndexer: __slots__ = ("data_array",) def __init__(self, data_array: DataArray): self.data_array = data_array def __getitem__(self, key) -> DataArray: if not utils.is_dict_like(key): # expand the indexer so we can handle Ellipsis labels = indexing.expanded_indexer(key, self.data_array.ndim) key = dict(zip(self.data_array.dims, labels)) return self.data_array.sel(key) def __setitem__(self, key, value) -> None: if not utils.is_dict_like(key): # expand the indexer so we can handle Ellipsis labels = indexing.expanded_indexer(key, self.data_array.ndim) key = dict(zip(self.data_array.dims, labels)) dim_indexers = map_index_queries(self.data_array, key).dim_indexers self.data_array[dim_indexers] = value # Used as the key corresponding to a DataArray's variable when converting # arbitrary DataArray objects to datasets _THIS_ARRAY = ReprObject("") class DataArray( AbstractArray, DataWithCoords, DataArrayArithmetic, DataArrayAggregations, ): """N-dimensional array with labeled coordinates and dimensions. DataArray provides a wrapper around numpy ndarrays that uses labeled dimensions and coordinates to support metadata aware operations. The API is similar to that for the pandas Series or DataFrame, but DataArray objects can have any number of dimensions, and their contents have fixed data types. Additional features over raw numpy arrays: - Apply operations over dimensions by name: ``x.sum('time')``. - Select or assign values by integer location (like numpy): ``x[:10]`` or by label (like pandas): ``x.loc['2014-01-01']`` or ``x.sel(time='2014-01-01')``. - Mathematical operations (e.g., ``x - y``) vectorize across multiple dimensions (known in numpy as "broadcasting") based on dimension names, regardless of their original order. - Keep track of arbitrary metadata in the form of a Python dictionary: ``x.attrs`` - Convert to a pandas Series: ``x.to_series()``. Getting items from or doing mathematical operations with a DataArray always returns another DataArray. Parameters ---------- data : array_like Values for this array. Must be an ``numpy.ndarray``, ndarray like, or castable to an ``ndarray``. If a self-described xarray or pandas object, attempts are made to use this array's metadata to fill in other unspecified arguments. A view of the array's data is used instead of a copy if possible. coords : sequence or dict of array_like, optional Coordinates (tick labels) to use for indexing along each dimension. The following notations are accepted: - mapping {dimension name: array-like} - sequence of tuples that are valid arguments for ``xarray.Variable()`` - (dims, data) - (dims, data, attrs) - (dims, data, attrs, encoding) Additionally, it is possible to define a coord whose name does not match the dimension name, or a coord based on multiple dimensions, with one of the following notations: - mapping {coord name: DataArray} - mapping {coord name: Variable} - mapping {coord name: (dimension name, array-like)} - mapping {coord name: (tuple of dimension names, array-like)} dims : Hashable or sequence of Hashable, optional Name(s) of the data dimension(s). Must be either a Hashable (only for 1D data) or a sequence of Hashables with length equal to the number of dimensions. If this argument is omitted, dimension names are taken from ``coords`` (if possible) and otherwise default to ``['dim_0', ... 'dim_n']``. name : str or None, optional Name of this array. attrs : dict_like or None, optional Attributes to assign to the new instance. By default, an empty attribute dictionary is initialized. Examples -------- Create data: >>> np.random.seed(0) >>> temperature = 15 + 8 * np.random.randn(2, 2, 3) >>> lon = [[-99.83, -99.32], [-99.79, -99.23]] >>> lat = [[42.25, 42.21], [42.63, 42.59]] >>> time = pd.date_range("2014-09-06", periods=3) >>> reference_time = pd.Timestamp("2014-09-05") Initialize a dataarray with multiple dimensions: >>> da = xr.DataArray( ... data=temperature, ... dims=["x", "y", "time"], ... coords=dict( ... lon=(["x", "y"], lon), ... lat=(["x", "y"], lat), ... time=time, ... reference_time=reference_time, ... ), ... attrs=dict( ... description="Ambient temperature.", ... units="degC", ... ), ... ) >>> da array([[[29.11241877, 18.20125767, 22.82990387], [32.92714559, 29.94046392, 7.18177696]], [[22.60070734, 13.78914233, 14.17424919], [18.28478802, 16.15234857, 26.63418806]]]) Coordinates: lon (x, y) float64 -99.83 -99.32 -99.79 -99.23 lat (x, y) float64 42.25 42.21 42.63 42.59 * time (time) datetime64[ns] 2014-09-06 2014-09-07 2014-09-08 reference_time datetime64[ns] 2014-09-05 Dimensions without coordinates: x, y Attributes: description: Ambient temperature. units: degC Find out where the coldest temperature was: >>> da.isel(da.argmin(...)) array(7.18177696) Coordinates: lon float64 -99.32 lat float64 42.21 time datetime64[ns] 2014-09-08 reference_time datetime64[ns] 2014-09-05 Attributes: description: Ambient temperature. units: degC """ _cache: dict[str, Any] _coords: dict[Any, Variable] _close: Callable[[], None] | None _indexes: dict[Hashable, Index] _name: Hashable | None _variable: Variable __slots__ = ( "_cache", "_coords", "_close", "_indexes", "_name", "_variable", "__weakref__", ) dt = utils.UncachedAccessor(CombinedDatetimelikeAccessor["DataArray"]) def __init__( self, data: Any = dtypes.NA, coords: Sequence[Sequence[Any] | pd.Index | DataArray] | Mapping[Any, Any] | None = None, dims: Hashable | Sequence[Hashable] | None = None, name: Hashable | None = None, attrs: Mapping | None = None, # internal parameters indexes: dict[Hashable, Index] | None = None, fastpath: bool = False, ) -> None: if fastpath: variable = data assert dims is None assert attrs is None assert indexes is not None else: # TODO: (benbovy - explicit indexes) remove # once it becomes part of the public interface if indexes is not None: raise ValueError("Providing explicit indexes is not supported yet") # try to fill in arguments from data if they weren't supplied if coords is None: if isinstance(data, DataArray): coords = data.coords elif isinstance(data, pd.Series): coords = [data.index] elif isinstance(data, pd.DataFrame): coords = [data.index, data.columns] elif isinstance(data, (pd.Index, IndexVariable)): coords = [data] if dims is None: dims = getattr(data, "dims", getattr(coords, "dims", None)) if name is None: name = getattr(data, "name", None) if attrs is None and not isinstance(data, PANDAS_TYPES): attrs = getattr(data, "attrs", None) data = _check_data_shape(data, coords, dims) data = as_compatible_data(data) coords, dims = _infer_coords_and_dims(data.shape, coords, dims) variable = Variable(dims, data, attrs, fastpath=True) indexes, coords = _create_indexes_from_coords(coords) # These fully describe a DataArray self._variable = variable assert isinstance(coords, dict) self._coords = coords self._name = name # TODO(shoyer): document this argument, once it becomes part of the # public interface. self._indexes = indexes self._close = None @classmethod def _construct_direct( cls: type[T_DataArray], variable: Variable, coords: dict[Any, Variable], name: Hashable, indexes: dict[Hashable, Index], ) -> T_DataArray: """Shortcut around __init__ for internal use when we want to skip costly validation """ obj = object.__new__(cls) obj._variable = variable obj._coords = coords obj._name = name obj._indexes = indexes obj._close = None return obj def _replace( self: T_DataArray, variable: Variable | None = None, coords=None, name: Hashable | None | Default = _default, indexes=None, ) -> T_DataArray: if variable is None: variable = self.variable if coords is None: coords = self._coords if indexes is None: indexes = self._indexes if name is _default: name = self.name return type(self)(variable, coords, name=name, indexes=indexes, fastpath=True) def _replace_maybe_drop_dims( self: T_DataArray, variable: Variable, name: Hashable | None | Default = _default, ) -> T_DataArray: if variable.dims == self.dims and variable.shape == self.shape: coords = self._coords.copy() indexes = self._indexes elif variable.dims == self.dims: # Shape has changed (e.g. from reduce(..., keepdims=True) new_sizes = dict(zip(self.dims, variable.shape)) coords = { k: v for k, v in self._coords.items() if v.shape == tuple(new_sizes[d] for d in v.dims) } indexes = filter_indexes_from_coords(self._indexes, set(coords)) else: allowed_dims = set(variable.dims) coords = { k: v for k, v in self._coords.items() if set(v.dims) <= allowed_dims } indexes = filter_indexes_from_coords(self._indexes, set(coords)) return self._replace(variable, coords, name, indexes=indexes) def _overwrite_indexes( self: T_DataArray, indexes: Mapping[Any, Index], coords: Mapping[Any, Variable] | None = None, drop_coords: list[Hashable] | None = None, rename_dims: Mapping[Any, Any] | None = None, ) -> T_DataArray: """Maybe replace indexes and their corresponding coordinates.""" if not indexes: return self if coords is None: coords = {} if drop_coords is None: drop_coords = [] new_variable = self.variable.copy() new_coords = self._coords.copy() new_indexes = dict(self._indexes) for name in indexes: new_coords[name] = coords[name] new_indexes[name] = indexes[name] for name in drop_coords: new_coords.pop(name) new_indexes.pop(name) if rename_dims: new_variable.dims = tuple(rename_dims.get(d, d) for d in new_variable.dims) return self._replace( variable=new_variable, coords=new_coords, indexes=new_indexes ) def _to_temp_dataset(self) -> Dataset: return self._to_dataset_whole(name=_THIS_ARRAY, shallow_copy=False) def _from_temp_dataset( self: T_DataArray, dataset: Dataset, name: Hashable | None | Default = _default ) -> T_DataArray: variable = dataset._variables.pop(_THIS_ARRAY) coords = dataset._variables indexes = dataset._indexes return self._replace(variable, coords, name, indexes=indexes) def _to_dataset_split(self, dim: Hashable) -> Dataset: """splits dataarray along dimension 'dim'""" def subset(dim, label): array = self.loc[{dim: label}] array.attrs = {} return as_variable(array) variables = {label: subset(dim, label) for label in self.get_index(dim)} variables.update({k: v for k, v in self._coords.items() if k != dim}) coord_names = set(self._coords) - {dim} indexes = filter_indexes_from_coords(self._indexes, coord_names) dataset = Dataset._construct_direct( variables, coord_names, indexes=indexes, attrs=self.attrs ) return dataset def _to_dataset_whole( self, name: Hashable = None, shallow_copy: bool = True ) -> Dataset: if name is None: name = self.name if name is None: raise ValueError( "unable to convert unnamed DataArray to a " "Dataset without providing an explicit name" ) if name in self.coords: raise ValueError( "cannot create a Dataset from a DataArray with " "the same name as one of its coordinates" ) # use private APIs for speed: this is called by _to_temp_dataset(), # which is used in the guts of a lot of operations (e.g., reindex) variables = self._coords.copy() variables[name] = self.variable if shallow_copy: for k in variables: variables[k] = variables[k].copy(deep=False) indexes = self._indexes coord_names = set(self._coords) return Dataset._construct_direct(variables, coord_names, indexes=indexes) def to_dataset( self, dim: Hashable = None, *, name: Hashable = None, promote_attrs: bool = False, ) -> 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 : Dataset """ if dim is not None and dim not in self.dims: raise TypeError( f"{dim} is not a dim. If supplying a ``name``, pass as a kwarg." ) if dim is not None: if name is not None: raise TypeError("cannot supply both dim and name arguments") result = self._to_dataset_split(dim) else: result = self._to_dataset_whole(name) if promote_attrs: result.attrs = dict(self.attrs) return result @property def name(self) -> Hashable | None: """The name of this array.""" return self._name @name.setter def name(self, value: Hashable | None) -> None: self._name = value @property def variable(self) -> Variable: """Low level interface to the Variable object for this DataArray.""" return self._variable @property def dtype(self) -> np.dtype: """ Data-type of the array’s elements. See Also -------- ndarray.dtype numpy.dtype """ return self.variable.dtype @property def shape(self) -> tuple[int, ...]: """ Tuple of array dimensions. See Also -------- numpy.ndarray.shape """ return self.variable.shape @property def size(self) -> int: """ 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 """ return self.variable.size @property def nbytes(self) -> int: """ 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``. """ return self.variable.nbytes @property def ndim(self) -> int: """ Number of array dimensions. See Also -------- numpy.ndarray.ndim """ return self.variable.ndim def __len__(self) -> int: return len(self.variable) @property def data(self) -> Any: """ 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 """ return self.variable.data @data.setter def data(self, value: Any) -> None: self.variable.data = value @property def values(self) -> np.ndarray: """ The array's data as a numpy.ndarray. If the array's data is not a numpy.ndarray this will attempt to convert it naively using np.array(), which will raise an error if the array type does not support coercion like this (e.g. cupy). """ return self.variable.values @values.setter def values(self, value: Any) -> None: self.variable.values = value def to_numpy(self) -> np.ndarray: """ 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 """ return self.variable.to_numpy() def as_numpy(self: T_DataArray) -> T_DataArray: """ 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 """ coords = {k: v.as_numpy() for k, v in self._coords.items()} return self._replace(self.variable.as_numpy(), coords, indexes=self._indexes) @property def _in_memory(self) -> bool: return self.variable._in_memory def _to_index(self) -> pd.Index: return self.variable._to_index() def to_index(self) -> pd.Index: """Convert this variable to a pandas.Index. Only possible for 1D arrays. """ return self.variable.to_index() @property def dims(self) -> tuple[Hashable, ...]: """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 """ return self.variable.dims @dims.setter def dims(self, value: Any) -> NoReturn: raise AttributeError( "you cannot assign dims on a DataArray. Use " ".rename() or .swap_dims() instead." ) def _item_key_to_dict(self, key: Any) -> Mapping[Hashable, Any]: if utils.is_dict_like(key): return key key = indexing.expanded_indexer(key, self.ndim) return dict(zip(self.dims, key)) def _getitem_coord(self: T_DataArray, key: Any) -> T_DataArray: from xarray.core.dataset import _get_virtual_variable try: var = self._coords[key] except KeyError: dim_sizes = dict(zip(self.dims, self.shape)) _, key, var = _get_virtual_variable(self._coords, key, dim_sizes) return self._replace_maybe_drop_dims(var, name=key) def __getitem__(self: T_DataArray, key: Any) -> T_DataArray: if isinstance(key, str): return self._getitem_coord(key) else: # xarray-style array indexing return self.isel(indexers=self._item_key_to_dict(key)) def __setitem__(self, key: Any, value: Any) -> None: if isinstance(key, str): self.coords[key] = value else: # Coordinates in key, value and self[key] should be consistent. # TODO Coordinate consistency in key is checked here, but it # causes unnecessary indexing. It should be optimized. obj = self[key] if isinstance(value, DataArray): assert_coordinate_consistent(value, obj.coords.variables) # DataArray key -> Variable key key = { k: v.variable if isinstance(v, DataArray) else v for k, v in self._item_key_to_dict(key).items() } self.variable[key] = value def __delitem__(self, key: Any) -> None: del self.coords[key] @property def _attr_sources(self) -> Iterable[Mapping[Hashable, Any]]: """Places to look-up items for attribute-style access""" yield from self._item_sources yield self.attrs @property def _item_sources(self) -> Iterable[Mapping[Hashable, Any]]: """Places to look-up items for key-completion""" yield HybridMappingProxy(keys=self._coords, mapping=self.coords) # virtual coordinates # uses empty dict -- everything here can already be found in self.coords. yield HybridMappingProxy(keys=self.dims, mapping={}) def __contains__(self, key: Any) -> bool: return key in self.data @property def loc(self) -> _LocIndexer: """Attribute for location based indexing like pandas.""" return _LocIndexer(self) @property def attrs(self) -> dict[Any, Any]: """Dictionary storing arbitrary metadata with this array.""" return self.variable.attrs @attrs.setter def attrs(self, value: Mapping[Any, Any]) -> None: self.variable.attrs = dict(value) @property def encoding(self) -> dict[Any, Any]: """Dictionary of format-specific settings for how this array should be serialized.""" return self.variable.encoding @encoding.setter def encoding(self, value: Mapping[Any, Any]) -> None: self.variable.encoding = dict(value) @property def indexes(self) -> 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 """ return self.xindexes.to_pandas_indexes() @property def xindexes(self) -> Indexes: """Mapping of xarray Index objects used for label based indexing.""" return Indexes(self._indexes, {k: self._coords[k] for k in self._indexes}) @property def coords(self) -> DataArrayCoordinates: """Dictionary-like container of coordinate arrays.""" return DataArrayCoordinates(self) @overload def reset_coords( self: T_DataArray, names: Dims = None, drop: Literal[False] = False, ) -> Dataset: ... @overload def reset_coords( self: T_DataArray, names: Dims = None, *, drop: Literal[True], ) -> T_DataArray: ... def reset_coords( self: T_DataArray, names: Dims = None, drop: bool = False, ) -> T_DataArray | Dataset: """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. Returns ------- 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 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 10 11 12 13 14 lat (y) int64 20 21 22 23 24 Pressure (x, y) int64 50 51 52 53 54 55 56 57 ... 67 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") Dimensions: (x: 5, y: 5) Coordinates: lon (x) int64 10 11 12 13 14 lat (y) int64 20 21 22 23 24 Dimensions without coordinates: x, y Data variables: Pressure (x, y) int64 50 51 52 53 54 55 56 57 ... 68 69 70 71 72 73 74 Temperature (x, y) int64 0 1 2 3 4 5 6 7 8 9 ... 16 17 18 19 20 21 22 23 24 Return DataArray without targeted coordinate: >>> da.reset_coords(names="Pressure", drop=True) 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 10 11 12 13 14 lat (y) int64 20 21 22 23 24 Dimensions without coordinates: x, y """ if names is None: names = set(self.coords) - set(self._indexes) dataset = self.coords.to_dataset().reset_coords(names, drop) if drop: return self._replace(coords=dataset._variables) if self.name is None: raise ValueError( "cannot reset_coords with drop=False on an unnamed DataArrray" ) dataset[self.name] = self.variable return dataset def __dask_tokenize__(self): from dask.base import normalize_token return normalize_token((type(self), self._variable, self._coords, self._name)) def __dask_graph__(self): return self._to_temp_dataset().__dask_graph__() def __dask_keys__(self): return self._to_temp_dataset().__dask_keys__() def __dask_layers__(self): return self._to_temp_dataset().__dask_layers__() @property def __dask_optimize__(self): return self._to_temp_dataset().__dask_optimize__ @property def __dask_scheduler__(self): return self._to_temp_dataset().__dask_scheduler__ def __dask_postcompute__(self): func, args = self._to_temp_dataset().__dask_postcompute__() return self._dask_finalize, (self.name, func) + args def __dask_postpersist__(self): func, args = self._to_temp_dataset().__dask_postpersist__() return self._dask_finalize, (self.name, func) + args @staticmethod def _dask_finalize(results, name, func, *args, **kwargs) -> DataArray: ds = func(results, *args, **kwargs) variable = ds._variables.pop(_THIS_ARRAY) coords = ds._variables indexes = ds._indexes return DataArray(variable, coords, name=name, indexes=indexes, fastpath=True) def load(self: T_DataArray, **kwargs) -> T_DataArray: """Manually trigger loading of this array's data from disk or a remote source into memory and return this array. 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``. See Also -------- dask.compute """ ds = self._to_temp_dataset().load(**kwargs) new = self._from_temp_dataset(ds) self._variable = new._variable self._coords = new._coords return self def compute(self: T_DataArray, **kwargs) -> T_DataArray: """Manually trigger loading of this array's data from disk or a remote source into memory and return a new array. The original 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``. See Also -------- dask.compute """ new = self.copy(deep=False) return new.load(**kwargs) def persist(self: T_DataArray, **kwargs) -> T_DataArray: """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. Parameters ---------- **kwargs : dict Additional keyword arguments passed on to ``dask.persist``. See Also -------- dask.persist """ ds = self._to_temp_dataset().persist(**kwargs) return self._from_temp_dataset(ds) def copy(self: T_DataArray, deep: bool = True, data: Any = None) -> T_DataArray: """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 : DataArray New object with dimensions, attributes, coordinates, name, encoding, and optionally data copied from original. Examples -------- Shallow versus deep copy >>> array = xr.DataArray([1, 2, 3], dims="x", coords={"x": ["a", "b", "c"]}) >>> array.copy() array([1, 2, 3]) Coordinates: * x (x) >> array_0 = array.copy(deep=False) >>> array_0[0] = 7 >>> array_0 array([7, 2, 3]) Coordinates: * x (x) >> array array([7, 2, 3]) Coordinates: * x (x) >> array.copy(data=[0.1, 0.2, 0.3]) array([0.1, 0.2, 0.3]) Coordinates: * x (x) >> array array([7, 2, 3]) Coordinates: * x (x) T_DataArray: variable = self.variable._copy(deep=deep, data=data, memo=memo) indexes, index_vars = self.xindexes.copy_indexes(deep=deep) coords = {} for k, v in self._coords.items(): if k in index_vars: coords[k] = index_vars[k] else: coords[k] = v._copy(deep=deep, memo=memo) return self._replace(variable, coords, indexes=indexes) def __copy__(self: T_DataArray) -> T_DataArray: return self._copy(deep=False) def __deepcopy__( self: T_DataArray, memo: dict[int, Any] | None = None ) -> T_DataArray: return self._copy(deep=True, memo=memo) # mutable objects should not be Hashable # https://github.com/python/mypy/issues/4266 __hash__ = None # type: ignore[assignment] @property def chunks(self) -> tuple[tuple[int, ...], ...] | None: """ 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 """ return self.variable.chunks @property def chunksizes(self) -> Mapping[Any, tuple[int, ...]]: """ Mapping from dimension names to block lengths for this dataarray's data, or None if the underlying data is not a dask array. 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 """ all_variables = [self.variable] + [c.variable for c in self.coords.values()] return get_chunksizes(all_variables) def chunk( self: T_DataArray, chunks: ( int | Literal["auto"] | tuple[int, ...] | tuple[tuple[int, ...], ...] | Mapping[Any, None | int | tuple[int, ...]] ) = {}, # {} even though it's technically unsafe, is being used intentionally here (#4667) name_prefix: str = "xarray-", token: str | None = None, lock: bool = False, inline_array: bool = False, **chunks_kwargs: Any, ) -> T_DataArray: """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. Parameters ---------- chunks : int, "auto", tuple of int or mapping of Hashable to int, optional Chunk sizes along each dimension, e.g., ``5``, ``"auto"``, ``(5, 5)`` or ``{"x": 5, "y": 5}``. name_prefix : str, optional Prefix for the name of the new dask array. token : str, optional Token uniquely identifying this array. lock : optional Passed on to :py:func:`dask.array.from_array`, if the array is not already as dask array. inline_array: optional Passed on to :py:func:`dask.array.from_array`, if the array is not already as dask array. **chunks_kwargs : {dim: chunks, ...}, optional The keyword arguments form of ``chunks``. One of chunks or chunks_kwargs must be provided. Returns ------- chunked : xarray.DataArray See Also -------- DataArray.chunks DataArray.chunksizes xarray.unify_chunks dask.array.from_array """ if chunks is None: warnings.warn( "None value for 'chunks' is deprecated. " "It will raise an error in the future. Use instead '{}'", category=FutureWarning, ) chunks = {} if isinstance(chunks, (float, str, int)): # ignoring type; unclear why it won't accept a Literal into the value. chunks = dict.fromkeys(self.dims, chunks) # type: ignore elif isinstance(chunks, (tuple, list)): chunks = dict(zip(self.dims, chunks)) else: chunks = either_dict_or_kwargs(chunks, chunks_kwargs, "chunk") ds = self._to_temp_dataset().chunk( chunks, name_prefix=name_prefix, token=token, lock=lock, inline_array=inline_array, ) return self._from_temp_dataset(ds) def isel( self: T_DataArray, indexers: Mapping[Any, Any] | None = None, drop: bool = False, missing_dims: ErrorOptionsWithWarn = "raise", **indexers_kwargs: Any, ) -> T_DataArray: """Return a new DataArray whose data is given by selecting indexes along the specified dimension(s). Parameters ---------- indexers : dict, optional A dict with keys matching dimensions and values given by integers, slice objects or arrays. indexer can be a integer, slice, array-like or DataArray. If DataArrays are passed as indexers, xarray-style indexing will be carried out. See :ref:`indexing` for the details. One of indexers or indexers_kwargs must be provided. drop : bool, default: False If ``drop=True``, drop coordinates variables indexed by integers instead of making them scalar. 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 **indexers_kwargs : {dim: indexer, ...}, optional The keyword arguments form of ``indexers``. Returns ------- indexed : xarray.DataArray See Also -------- Dataset.isel DataArray.sel Examples -------- >>> da = xr.DataArray(np.arange(25).reshape(5, 5), dims=("x", "y")) >>> da 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 >>> tgt_x = xr.DataArray(np.arange(0, 5), dims="points") >>> tgt_y = xr.DataArray(np.arange(0, 5), dims="points") >>> da = da.isel(x=tgt_x, y=tgt_y) >>> da array([ 0, 6, 12, 18, 24]) Dimensions without coordinates: points """ indexers = either_dict_or_kwargs(indexers, indexers_kwargs, "isel") if any(is_fancy_indexer(idx) for idx in indexers.values()): ds = self._to_temp_dataset()._isel_fancy( indexers, drop=drop, missing_dims=missing_dims ) return self._from_temp_dataset(ds) # Much faster algorithm for when all indexers are ints, slices, one-dimensional # lists, or zero or one-dimensional np.ndarray's variable = self._variable.isel(indexers, missing_dims=missing_dims) indexes, index_variables = isel_indexes(self.xindexes, indexers) coords = {} for coord_name, coord_value in self._coords.items(): if coord_name in index_variables: coord_value = index_variables[coord_name] else: coord_indexers = { k: v for k, v in indexers.items() if k in coord_value.dims } if coord_indexers: coord_value = coord_value.isel(coord_indexers) if drop and coord_value.ndim == 0: continue coords[coord_name] = coord_value return self._replace(variable=variable, coords=coords, indexes=indexes) def sel( self: T_DataArray, indexers: Mapping[Any, Any] | None = None, method: str | None = None, tolerance=None, drop: bool = False, **indexers_kwargs: Any, ) -> T_DataArray: """Return a new DataArray whose data is given by selecting index labels along the specified dimension(s). In contrast to `DataArray.isel`, indexers for this method should use labels instead of integers. Under the hood, this method is powered by using pandas's powerful Index objects. This makes label based indexing essentially just as fast as using integer indexing. It also means this method uses pandas's (well documented) logic for indexing. This means you can use string shortcuts for datetime indexes (e.g., '2000-01' to select all values in January 2000). It also means that slices are treated as inclusive of both the start and stop values, unlike normal Python indexing. .. warning:: Do not try to assign values when using any of the indexing methods ``isel`` or ``sel``:: da = xr.DataArray([0, 1, 2, 3], dims=['x']) # DO NOT do this da.isel(x=[0, 1, 2])[1] = -1 Assigning values with the chained indexing using ``.sel`` or ``.isel`` fails silently. Parameters ---------- indexers : dict, optional A dict with keys matching dimensions and values given by scalars, slices or arrays of tick labels. For dimensions with multi-index, the indexer may also be a dict-like object with keys matching index level names. If DataArrays are passed as indexers, xarray-style indexing will be carried out. See :ref:`indexing` for the details. One of indexers or indexers_kwargs must be provided. method : {None, "nearest", "pad", "ffill", "backfill", "bfill"}, optional Method to use for inexact matches: - None (default): only exact matches - pad / ffill: propagate last valid index value forward - backfill / bfill: propagate next valid index value backward - nearest: use nearest valid index value tolerance : optional 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``. drop : bool, optional If ``drop=True``, drop coordinates variables in `indexers` instead of making them scalar. **indexers_kwargs : {dim: indexer, ...}, optional The keyword arguments form of ``indexers``. One of indexers or indexers_kwargs must be provided. Returns ------- obj : DataArray A new DataArray with the same contents as this DataArray, except the data and each dimension is indexed by the appropriate indexers. If indexer DataArrays have coordinates that do not conflict with this object, then these coordinates will be attached. In general, each array's data will be a view of the array's data in this DataArray, unless vectorized indexing was triggered by using an array indexer, in which case the data will be a copy. See Also -------- Dataset.sel DataArray.isel Examples -------- >>> da = xr.DataArray( ... np.arange(25).reshape(5, 5), ... coords={"x": np.arange(5), "y": np.arange(5)}, ... dims=("x", "y"), ... ) >>> da 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 0 1 2 3 4 * y (y) int64 0 1 2 3 4 >>> tgt_x = xr.DataArray(np.linspace(0, 4, num=5), dims="points") >>> tgt_y = xr.DataArray(np.linspace(0, 4, num=5), dims="points") >>> da = da.sel(x=tgt_x, y=tgt_y, method="nearest") >>> da array([ 0, 6, 12, 18, 24]) Coordinates: x (points) int64 0 1 2 3 4 y (points) int64 0 1 2 3 4 Dimensions without coordinates: points """ ds = self._to_temp_dataset().sel( indexers=indexers, drop=drop, method=method, tolerance=tolerance, **indexers_kwargs, ) return self._from_temp_dataset(ds) def head( self: T_DataArray, indexers: Mapping[Any, int] | int | None = None, **indexers_kwargs: Any, ) -> T_DataArray: """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 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) array([[0, 1, 2, 3, 4]]) Dimensions without coordinates: x, y >>> da.head({"x": 2, "y": 2}) array([[0, 1], [5, 6]]) Dimensions without coordinates: x, y """ ds = self._to_temp_dataset().head(indexers, **indexers_kwargs) return self._from_temp_dataset(ds) def tail( self: T_DataArray, indexers: Mapping[Any, int] | int | None = None, **indexers_kwargs: Any, ) -> T_DataArray: """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 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) array([[ 4], [ 9], [14], [19], [24]]) Dimensions without coordinates: x, y >>> da.tail({"x": 2, "y": 2}) array([[18, 19], [23, 24]]) Dimensions without coordinates: x, y """ ds = self._to_temp_dataset().tail(indexers, **indexers_kwargs) return self._from_temp_dataset(ds) def thin( self: T_DataArray, indexers: Mapping[Any, int] | int | None = None, **indexers_kwargs: Any, ) -> T_DataArray: """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 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 0 1 * y (y) int64 0 1 2 3 4 5 6 7 8 9 10 11 12 >>> >>> x.thin(3) array([[ 0, 3, 6, 9, 12]]) Coordinates: * x (x) int64 0 * y (y) int64 0 3 6 9 12 >>> x.thin({"x": 2, "y": 5}) array([[ 0, 5, 10]]) Coordinates: * x (x) int64 0 * y (y) int64 0 5 10 See Also -------- Dataset.thin DataArray.head DataArray.tail """ ds = self._to_temp_dataset().thin(indexers, **indexers_kwargs) return self._from_temp_dataset(ds) def broadcast_like( self: T_DataArray, other: DataArray | Dataset, exclude: Iterable[Hashable] | None = None, ) -> T_DataArray: """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 : DataArray The caller broadcasted against ``other``. 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 array([[ 1.76405235, 0.40015721, 0.97873798], [ 2.2408932 , 1.86755799, -0.97727788]]) Coordinates: * x (x) >> arr2 array([[ 0.95008842, -0.15135721], [-0.10321885, 0.4105985 ], [ 0.14404357, 1.45427351]]) Coordinates: * x (x) >> arr1.broadcast_like(arr2) array([[ 1.76405235, 0.40015721, 0.97873798], [ 2.2408932 , 1.86755799, -0.97727788], [ nan, nan, nan]]) Coordinates: * x (x) T_DataArray: """Callback called from ``Aligner`` to create a new reindexed DataArray.""" if isinstance(fill_value, dict): fill_value = fill_value.copy() sentinel = object() value = fill_value.pop(self.name, sentinel) if value is not sentinel: fill_value[_THIS_ARRAY] = value ds = self._to_temp_dataset() reindexed = ds._reindex_callback( aligner, dim_pos_indexers, variables, indexes, fill_value, exclude_dims, exclude_vars, ) return self._from_temp_dataset(reindexed) def reindex_like( self: T_DataArray, other: DataArray | Dataset, method: ReindexMethodOptions = None, tolerance: int | float | Iterable[int | float] | None = None, copy: bool = True, fill_value=dtypes.NA, ) -> T_DataArray: """Conform this object onto the indexes of another object, filling in missing values 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 mis-matched index values will be filled in with NaN, and any mis-matched 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 : optional 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 : DataArray Another dataset array, with this array's data but coordinates from the other object. 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 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) int64 10 20 30 40 * y (y) int64 70 80 90 >>> da2 = xr.DataArray( ... data=data, ... dims=["x", "y"], ... coords={"x": [40, 30, 20, 10], "y": [90, 80, 70]}, ... ) >>> da2 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) int64 40 30 20 10 * y (y) int64 90 80 70 Reindexing with both DataArrays having the same coordinates set, but in different order: >>> da1.reindex_like(da2) array([[11, 10, 9], [ 8, 7, 6], [ 5, 4, 3], [ 2, 1, 0]]) Coordinates: * x (x) int64 40 30 20 10 * y (y) int64 90 80 70 Reindexing with the other array having coordinates which the source array doesn't have: >>> 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]}, ... ) >>> da2 = xr.DataArray( ... data=data, ... dims=["x", "y"], ... coords={"x": [20, 10, 29, 39], "y": [70, 80, 90]}, ... ) >>> da1.reindex_like(da2) array([[ 3., 4., 5.], [ 0., 1., 2.], [nan, nan, nan], [nan, nan, nan]]) Coordinates: * x (x) int64 20 10 29 39 * y (y) int64 70 80 90 Filling missing values with the previous valid index with respect to the coordinates' value: >>> da1.reindex_like(da2, method="ffill") array([[3, 4, 5], [0, 1, 2], [3, 4, 5], [6, 7, 8]]) Coordinates: * x (x) int64 20 10 29 39 * y (y) int64 70 80 90 Filling missing values while tolerating specified error for inexact matches: >>> da1.reindex_like(da2, method="ffill", tolerance=5) array([[ 3., 4., 5.], [ 0., 1., 2.], [nan, nan, nan], [nan, nan, nan]]) Coordinates: * x (x) int64 20 10 29 39 * y (y) int64 70 80 90 Filling missing values with manually specified values: >>> da1.reindex_like(da2, fill_value=19) array([[ 3, 4, 5], [ 0, 1, 2], [19, 19, 19], [19, 19, 19]]) Coordinates: * x (x) int64 20 10 29 39 * y (y) int64 70 80 90 See Also -------- DataArray.reindex align """ return alignment.reindex_like( self, other=other, method=method, tolerance=tolerance, copy=copy, fill_value=fill_value, ) def reindex( self: T_DataArray, indexers: Mapping[Any, Any] | None = None, method: ReindexMethodOptions = None, tolerance: float | Iterable[float] | None = None, copy: bool = True, fill_value=dtypes.NA, **indexers_kwargs: Any, ) -> T_DataArray: """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 mis-matched coordinate values will be filled in with NaN, and any mis-matched 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] | 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 : DataArray Another dataset array, with this array's data but replaced coordinates. Examples -------- Reverse latitude: >>> da = xr.DataArray( ... np.arange(4), ... coords=[np.array([90, 89, 88, 87])], ... dims="lat", ... ) >>> da array([0, 1, 2, 3]) Coordinates: * lat (lat) int64 90 89 88 87 >>> da.reindex(lat=da.lat[::-1]) array([3, 2, 1, 0]) Coordinates: * lat (lat) int64 87 88 89 90 See Also -------- DataArray.reindex_like align """ indexers = utils.either_dict_or_kwargs(indexers, indexers_kwargs, "reindex") return alignment.reindex( self, indexers=indexers, method=method, tolerance=tolerance, copy=copy, fill_value=fill_value, ) def interp( self: T_DataArray, coords: Mapping[Any, Any] | None = None, method: InterpOptions = "linear", assume_sorted: bool = False, kwargs: Mapping[str, Any] | None = None, **coords_kwargs: Any, ) -> T_DataArray: """Interpolate a DataArray onto new coordinates Performs univariate or multivariate interpolation of a DataArray onto new coordinates using scipy's interpolation routines. If interpolating along an existing dimension, :py:class:`scipy.interpolate.interp1d` is called. When interpolating along multiple existing dimensions, an attempt is made to decompose the interpolation into multiple 1-dimensional interpolations. If this is possible, :py:class:`scipy.interpolate.interp1d` is called. Otherwise, :py:func:`scipy.interpolate.interpn` is called. Parameters ---------- coords : dict, optional Mapping from dimension names to the new coordinates. New coordinate can be a scalar, array-like or DataArray. If DataArrays are passed as new coordinates, their dimensions are used for the broadcasting. Missing values are skipped. method : {"linear", "nearest", "zero", "slinear", "quadratic", "cubic", "polynomial"}, default: "linear" The method used to interpolate. The method should be supported by the scipy interpolator: - ``interp1d``: {"linear", "nearest", "zero", "slinear", "quadratic", "cubic", "polynomial"} - ``interpn``: {"linear", "nearest"} If ``"polynomial"`` is passed, the ``order`` keyword argument must also be provided. assume_sorted : bool, default: False If False, values of x can be in any order and they are sorted first. If True, x has to be an array of monotonically increasing values. kwargs : dict-like or None, default: None Additional keyword arguments passed to scipy's interpolator. Valid options and their behavior depend whether ``interp1d`` or ``interpn`` is used. **coords_kwargs : {dim: coordinate, ...}, optional The keyword arguments form of ``coords``. One of coords or coords_kwargs must be provided. Returns ------- interpolated : DataArray New dataarray on the new coordinates. Notes ----- scipy is required. See Also -------- scipy.interpolate.interp1d scipy.interpolate.interpn Examples -------- >>> da = xr.DataArray( ... data=[[1, 4, 2, 9], [2, 7, 6, np.nan], [6, np.nan, 5, 8]], ... dims=("x", "y"), ... coords={"x": [0, 1, 2], "y": [10, 12, 14, 16]}, ... ) >>> da array([[ 1., 4., 2., 9.], [ 2., 7., 6., nan], [ 6., nan, 5., 8.]]) Coordinates: * x (x) int64 0 1 2 * y (y) int64 10 12 14 16 1D linear interpolation (the default): >>> da.interp(x=[0, 0.75, 1.25, 1.75]) array([[1. , 4. , 2. , nan], [1.75, 6.25, 5. , nan], [3. , nan, 5.75, nan], [5. , nan, 5.25, nan]]) Coordinates: * y (y) int64 10 12 14 16 * x (x) float64 0.0 0.75 1.25 1.75 1D nearest interpolation: >>> da.interp(x=[0, 0.75, 1.25, 1.75], method="nearest") array([[ 1., 4., 2., 9.], [ 2., 7., 6., nan], [ 2., 7., 6., nan], [ 6., nan, 5., 8.]]) Coordinates: * y (y) int64 10 12 14 16 * x (x) float64 0.0 0.75 1.25 1.75 1D linear extrapolation: >>> da.interp( ... x=[1, 1.5, 2.5, 3.5], ... method="linear", ... kwargs={"fill_value": "extrapolate"}, ... ) array([[ 2. , 7. , 6. , nan], [ 4. , nan, 5.5, nan], [ 8. , nan, 4.5, nan], [12. , nan, 3.5, nan]]) Coordinates: * y (y) int64 10 12 14 16 * x (x) float64 1.0 1.5 2.5 3.5 2D linear interpolation: >>> da.interp(x=[0, 0.75, 1.25, 1.75], y=[11, 13, 15], method="linear") array([[2.5 , 3. , nan], [4. , 5.625, nan], [ nan, nan, nan], [ nan, nan, nan]]) Coordinates: * x (x) float64 0.0 0.75 1.25 1.75 * y (y) int64 11 13 15 """ if self.dtype.kind not in "uifc": raise TypeError( "interp only works for a numeric type array. " "Given {}.".format(self.dtype) ) ds = self._to_temp_dataset().interp( coords, method=method, kwargs=kwargs, assume_sorted=assume_sorted, **coords_kwargs, ) return self._from_temp_dataset(ds) def interp_like( self: T_DataArray, other: DataArray | Dataset, method: InterpOptions = "linear", assume_sorted: bool = False, kwargs: Mapping[str, Any] | None = None, ) -> T_DataArray: """Interpolate this object onto the coordinates of another object, filling out of range values with NaN. If interpolating along a single existing dimension, :py:class:`scipy.interpolate.interp1d` is called. When interpolating along multiple existing dimensions, an attempt is made to decompose the interpolation into multiple 1-dimensional interpolations. If this is possible, :py:class:`scipy.interpolate.interp1d` is called. Otherwise, :py:func:`scipy.interpolate.interpn` is called. 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", "polynomial"}, default: "linear" The method used to interpolate. The method should be supported by the scipy interpolator: - {"linear", "nearest", "zero", "slinear", "quadratic", "cubic", "polynomial"} when ``interp1d`` is called. - {"linear", "nearest"} when ``interpn`` is called. If ``"polynomial"`` is passed, the ``order`` keyword argument must also be provided. 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 passed to scipy's interpolator. Returns ------- interpolated : DataArray Another dataarray by interpolating this dataarray's data along the coordinates of the other object. 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 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) int64 10 20 30 40 * y (y) int64 70 80 90 >>> da2 = xr.DataArray( ... data=data, ... dims=["x", "y"], ... coords={"x": [10, 20, 29, 39], "y": [70, 80, 90]}, ... ) >>> da2 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) int64 10 20 29 39 * y (y) int64 70 80 90 Interpolate the values in the coordinates of the other DataArray with respect to the source's values: >>> da2.interp_like(da1) array([[0. , 1. , 2. ], [3. , 4. , 5. ], [6.3, 7.3, 8.3], [nan, nan, nan]]) Coordinates: * x (x) int64 10 20 30 40 * y (y) int64 70 80 90 Could also extrapolate missing values: >>> da2.interp_like(da1, kwargs={"fill_value": "extrapolate"}) array([[ 0. , 1. , 2. ], [ 3. , 4. , 5. ], [ 6.3, 7.3, 8.3], [ 9.3, 10.3, 11.3]]) Coordinates: * x (x) int64 10 20 30 40 * y (y) int64 70 80 90 Notes ----- scipy is required. If the dataarray has object-type coordinates, reindex is used for these coordinates instead of the interpolation. See Also -------- DataArray.interp DataArray.reindex_like """ if self.dtype.kind not in "uifc": raise TypeError( "interp only works for a numeric type array. " "Given {}.".format(self.dtype) ) ds = self._to_temp_dataset().interp_like( other, method=method, kwargs=kwargs, assume_sorted=assume_sorted ) return self._from_temp_dataset(ds) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def rename( self, new_name_or_name_dict: Hashable | Mapping[Any, Hashable] | None = None, **names: Hashable, ) -> DataArray: """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 : DataArray Renamed array or array with renamed coordinates. See Also -------- Dataset.rename DataArray.swap_dims """ if new_name_or_name_dict is None and not names: # change name to None? return self._replace(name=None) if utils.is_dict_like(new_name_or_name_dict) or new_name_or_name_dict is None: # change dims/coords name_dict = either_dict_or_kwargs(new_name_or_name_dict, names, "rename") dataset = self._to_temp_dataset()._rename(name_dict) return self._from_temp_dataset(dataset) if utils.hashable(new_name_or_name_dict) and names: # change name + dims/coords dataset = self._to_temp_dataset()._rename(names) dataarray = self._from_temp_dataset(dataset) return dataarray._replace(name=new_name_or_name_dict) # only change name return self._replace(name=new_name_or_name_dict) def swap_dims( self: T_DataArray, dims_dict: Mapping[Any, Hashable] | None = None, **dims_kwargs, ) -> T_DataArray: """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 DataArray with swapped dimensions. Examples -------- >>> arr = xr.DataArray( ... data=[0, 1], ... dims="x", ... coords={"x": ["a", "b"], "y": ("x", [0, 1])}, ... ) >>> arr array([0, 1]) Coordinates: * x (x) >> arr.swap_dims({"x": "y"}) array([0, 1]) Coordinates: x (y) >> arr.swap_dims({"x": "z"}) array([0, 1]) Coordinates: x (z) DataArray: """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. 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. **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 : DataArray This object, but with additional dimension(s). See Also -------- Dataset.expand_dims Examples -------- >>> da = xr.DataArray(np.arange(5), dims=("x")) >>> da array([0, 1, 2, 3, 4]) Dimensions without coordinates: x Add new dimension of length 2: >>> da.expand_dims(dim={"y": 2}) array([[0, 1, 2, 3, 4], [0, 1, 2, 3, 4]]) Dimensions without coordinates: y, x >>> da.expand_dims(dim={"y": 2}, axis=1) 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) 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 0 1 2 3 4 Dimensions without coordinates: x """ if isinstance(dim, int): raise TypeError("dim should be Hashable or sequence/mapping of Hashables") elif isinstance(dim, Sequence) and not isinstance(dim, str): if len(dim) != len(set(dim)): raise ValueError("dims should not contain duplicate values.") dim = dict.fromkeys(dim, 1) elif dim is not None and not isinstance(dim, Mapping): dim = {cast(Hashable, dim): 1} dim = either_dict_or_kwargs(dim, dim_kwargs, "expand_dims") ds = self._to_temp_dataset().expand_dims(dim, axis) return self._from_temp_dataset(ds) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def set_index( self, indexes: Mapping[Any, Hashable | Sequence[Hashable]] | None = None, append: bool = False, **indexes_kwargs: Hashable | Sequence[Hashable], ) -> DataArray: """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 :py:meth:`~DataArray.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 : DataArray Another DataArray, with this data but replaced coordinates. Examples -------- >>> arr = xr.DataArray( ... data=np.ones((2, 3)), ... dims=["x", "y"], ... coords={"x": range(2), "y": range(3), "a": ("x", [3, 4])}, ... ) >>> arr array([[1., 1., 1.], [1., 1., 1.]]) Coordinates: * x (x) int64 0 1 * y (y) int64 0 1 2 a (x) int64 3 4 >>> arr.set_index(x="a") array([[1., 1., 1.], [1., 1., 1.]]) Coordinates: * x (x) int64 3 4 * y (y) int64 0 1 2 See Also -------- DataArray.reset_index DataArray.set_xindex """ ds = self._to_temp_dataset().set_index(indexes, append=append, **indexes_kwargs) return self._from_temp_dataset(ds) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def reset_index( self, dims_or_levels: Hashable | Sequence[Hashable], drop: bool = False, ) -> DataArray: """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 :py:meth:`~DataArray.drop_indexes` and :py:meth:`~DataArray.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 : DataArray Another dataarray, with this dataarray's data but replaced coordinates. See Also -------- DataArray.set_index DataArray.set_xindex DataArray.drop_indexes """ ds = self._to_temp_dataset().reset_index(dims_or_levels, drop=drop) return self._from_temp_dataset(ds) def set_xindex( self: T_DataArray, coord_names: str | Sequence[Hashable], index_cls: type[Index] | None = None, **options, ) -> T_DataArray: """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 :class:`~xarray.indexes.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 : DataArray Another dataarray, with this dataarray's data and with a new index. """ ds = self._to_temp_dataset().set_xindex(coord_names, index_cls, **options) return self._from_temp_dataset(ds) def reorder_levels( self: T_DataArray, dim_order: Mapping[Any, Sequence[int | Hashable]] | None = None, **dim_order_kwargs: Sequence[int | Hashable], ) -> T_DataArray: """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 : DataArray Another dataarray, with this dataarray's data but replaced coordinates. """ ds = self._to_temp_dataset().reorder_levels(dim_order, **dim_order_kwargs) return self._from_temp_dataset(ds) def stack( self: T_DataArray, dimensions: Mapping[Any, Sequence[Hashable]] | None = None, create_index: bool | None = True, index_cls: type[Index] = PandasMultiIndex, **dimensions_kwargs: Sequence[Hashable], ) -> T_DataArray: """ 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 ---------- dimensions : 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. **dimensions_kwargs The keyword arguments form of ``dimensions``. One of dimensions or dimensions_kwargs must be provided. Returns ------- stacked : DataArray DataArray with stacked data. Examples -------- >>> arr = xr.DataArray( ... np.arange(6).reshape(2, 3), ... coords=[("x", ["a", "b"]), ("y", [0, 1, 2])], ... ) >>> arr array([[0, 1, 2], [3, 4, 5]]) Coordinates: * x (x) >> 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 """ ds = self._to_temp_dataset().stack( dimensions, create_index=create_index, index_cls=index_cls, **dimensions_kwargs, ) return self._from_temp_dataset(ds) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def unstack( self, dim: Dims = None, fill_value: Any = dtypes.NA, sparse: bool = False, ) -> DataArray: """ 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 : DataArray Array with unstacked data. Examples -------- >>> arr = xr.DataArray( ... np.arange(6).reshape(2, 3), ... coords=[("x", ["a", "b"]), ("y", [0, 1, 2])], ... ) >>> arr array([[0, 1, 2], [3, 4, 5]]) Coordinates: * x (x) >> 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 """ ds = self._to_temp_dataset().unstack(dim, fill_value, sparse) return self._from_temp_dataset(ds) def to_unstacked_dataset(self, dim: Hashable, level: int | Hashable = 0) -> 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: 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 Dimensions: (x: 2, y: 3) Coordinates: * x (x) >> stacked = data.to_stacked_array("z", ["x"]) >>> stacked.indexes["z"] MultiIndex([('a', 0.0), ('a', 1.0), ('a', 2.0), ('b', nan)], name='z') >>> roundtripped = stacked.to_unstacked_dataset(dim="z") >>> data.identical(roundtripped) True See Also -------- Dataset.to_stacked_array """ idx = self._indexes[dim].to_pandas_index() if not isinstance(idx, pd.MultiIndex): raise ValueError(f"'{dim}' is not a stacked coordinate") level_number = idx._get_level_number(level) variables = idx.levels[level_number] variable_dim = idx.names[level_number] # pull variables out of datarray data_dict = {} for k in variables: data_dict[k] = self.sel({variable_dim: k}, drop=True).squeeze(drop=True) # unstacked dataset return Dataset(data_dict) def transpose( self: T_DataArray, *dims: Hashable, transpose_coords: bool = True, missing_dims: ErrorOptionsWithWarn = "raise", ) -> T_DataArray: """Return a new DataArray object with transposed dimensions. Parameters ---------- *dims : 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 : DataArray The returned DataArray's array is transposed. 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 """ if dims: dims = tuple(utils.infix_dims(dims, self.dims, missing_dims)) variable = self.variable.transpose(*dims) if transpose_coords: coords: dict[Hashable, Variable] = {} for name, coord in self.coords.items(): coord_dims = tuple(dim for dim in dims if dim in coord.dims) coords[name] = coord.variable.transpose(*coord_dims) return self._replace(variable, coords) else: return self._replace(variable) @property def T(self: T_DataArray) -> T_DataArray: return self.transpose() # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def drop_vars( self, names: Hashable | Iterable[Hashable], *, errors: ErrorOptions = "raise", ) -> DataArray: """Returns an array with dropped variables. Parameters ---------- names : Hashable or iterable of Hashable Name(s) of variables to drop. 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 : Dataset New Dataset copied from `self` with variables removed. 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 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) int64 10 20 30 40 * y (y) int64 70 80 90 Removing a single variable: >>> da.drop_vars("x") array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * y (y) int64 70 80 90 Dimensions without coordinates: x Removing a list of variables: >>> da.drop_vars(["x", "y"]) array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Dimensions without coordinates: x, y """ ds = self._to_temp_dataset().drop_vars(names, errors=errors) return self._from_temp_dataset(ds) def drop_indexes( self: T_DataArray, coord_names: Hashable | Iterable[Hashable], *, errors: ErrorOptions = "raise", ) -> T_DataArray: """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 : DataArray A new dataarray with dropped indexes. """ ds = self._to_temp_dataset().drop_indexes(coord_names, errors=errors) return self._from_temp_dataset(ds) def drop( self: T_DataArray, labels: Mapping[Any, Any] | None = None, dim: Hashable | None = None, *, errors: ErrorOptions = "raise", **labels_kwargs, ) -> T_DataArray: """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 """ ds = self._to_temp_dataset().drop(labels, dim, errors=errors, **labels_kwargs) return self._from_temp_dataset(ds) def drop_sel( self: T_DataArray, labels: Mapping[Any, Any] | None = None, *, errors: ErrorOptions = "raise", **labels_kwargs, ) -> T_DataArray: """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 : 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 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 0 2 4 6 8 * y (y) int64 0 3 6 9 12 >>> da.drop_sel(x=[0, 2], y=9) array([[10, 11, 12, 14], [15, 16, 17, 19], [20, 21, 22, 24]]) Coordinates: * x (x) int64 4 6 8 * y (y) int64 0 3 6 12 >>> da.drop_sel({"x": 6, "y": [0, 3]}) array([[ 2, 3, 4], [ 7, 8, 9], [12, 13, 14], [22, 23, 24]]) Coordinates: * x (x) int64 0 2 4 8 * y (y) int64 6 9 12 """ if labels_kwargs or isinstance(labels, dict): labels = either_dict_or_kwargs(labels, labels_kwargs, "drop") ds = self._to_temp_dataset().drop_sel(labels, errors=errors) return self._from_temp_dataset(ds) def drop_isel( self: T_DataArray, indexers: Mapping[Any, Any] | None = None, **indexers_kwargs ) -> T_DataArray: """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 : DataArray Raises ------ IndexError Examples -------- >>> da = xr.DataArray(np.arange(25).reshape(5, 5), dims=("X", "Y")) >>> da 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) 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}) array([[ 0, 1, 2, 4], [ 5, 6, 7, 9], [10, 11, 12, 14], [20, 21, 22, 24]]) Dimensions without coordinates: X, Y """ dataset = self._to_temp_dataset() dataset = dataset.drop_isel(indexers=indexers, **indexers_kwargs) return self._from_temp_dataset(dataset) def dropna( self: T_DataArray, dim: Hashable, how: Literal["any", "all"] = "any", thresh: int | None = None, ) -> T_DataArray: """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 : 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 array([[ 0., 4., 2., 9.], [nan, nan, nan, nan], [nan, 4., 2., 0.], [ 3., 1., 0., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 lon (X) float64 10.0 10.25 10.5 10.75 Dimensions without coordinates: Y, X >>> da.dropna(dim="Y", how="any") array([[0., 4., 2., 9.], [3., 1., 0., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.75 lon (X) float64 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") array([[ 0., 4., 2., 9.], [nan, 4., 2., 0.], [ 3., 1., 0., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.5 -20.75 lon (X) float64 10.0 10.25 10.5 10.75 Dimensions without coordinates: Y, X """ ds = self._to_temp_dataset().dropna(dim, how=how, thresh=thresh) return self._from_temp_dataset(ds) def fillna(self: T_DataArray, value: Any) -> T_DataArray: """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 : 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 array([ 1., 4., nan, 0., 3., nan]) Coordinates: * Z (Z) int64 0 1 2 3 4 5 height (Z) int64 0 10 20 30 40 50 Fill all NaN values with 0: >>> da.fillna(0) array([1., 4., 0., 0., 3., 0.]) Coordinates: * Z (Z) int64 0 1 2 3 4 5 height (Z) int64 0 10 20 30 40 50 Fill NaN values with corresponding values in array: >>> da.fillna(np.array([2, 9, 4, 2, 8, 9])) array([1., 4., 4., 0., 3., 9.]) Coordinates: * Z (Z) int64 0 1 2 3 4 5 height (Z) int64 0 10 20 30 40 50 """ if utils.is_dict_like(value): raise TypeError( "cannot provide fill value as a dictionary with " "fillna on a DataArray" ) out = ops.fillna(self, value) return out def interpolate_na( self: T_DataArray, dim: Hashable | None = None, method: InterpOptions = "linear", limit: int | None = None, use_coordinate: bool | str = True, max_gap: ( None | int | float | str | pd.Timedelta | np.timedelta64 | datetime.timedelta ) = None, keep_attrs: bool | None = None, **kwargs: Any, ) -> T_DataArray: """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", "krog", "pchip", "spline", "akima"}, default: "linear" String indicating which method to use for interpolation: - 'linear': linear interpolation. Additional keyword arguments are passed to :py:func:`numpy.interp` - 'nearest', 'zero', 'slinear', 'quadratic', 'cubic', 'polynomial': are passed to :py:func:`scipy.interpolate.interp1d`. If ``method='polynomial'``, the ``order`` keyword argument must also be provided. - 'barycentric', 'krog', 'pchip', 'spline', 'akima': use their respective :py:class:`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 eqaully-spaced along ``dim``. If True, the IndexVariable `dim` is used. If ``use_coordinate`` is a string, it specifies the name of a coordinate variariable 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 :py:class:`numpy.timedelta64` object - a :py:class:`pandas.Timedelta` object - a :py:class:`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:: 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: DataArray Filled in 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 array([nan, 2., 3., nan, 0.]) Coordinates: * x (x) int64 0 1 2 3 4 >>> da.interpolate_na(dim="x", method="linear") array([nan, 2. , 3. , 1.5, 0. ]) Coordinates: * x (x) int64 0 1 2 3 4 >>> da.interpolate_na(dim="x", method="linear", fill_value="extrapolate") array([1. , 2. , 3. , 1.5, 0. ]) Coordinates: * x (x) int64 0 1 2 3 4 """ from xarray.core.missing import interp_na return interp_na( self, dim=dim, method=method, limit=limit, use_coordinate=use_coordinate, max_gap=max_gap, keep_attrs=keep_attrs, **kwargs, ) def ffill( self: T_DataArray, dim: Hashable, limit: int | None = None ) -> T_DataArray: """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 : 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 array([[nan, 1., 3.], [ 0., nan, 5.], [ 5., nan, nan], [ 3., nan, nan], [ 0., 2., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 -21.0 lon (X) float64 10.0 10.25 10.5 Dimensions without coordinates: Y, X Fill all NaN values: >>> da.ffill(dim="Y", limit=None) array([[nan, 1., 3.], [ 0., 1., 5.], [ 5., 1., 5.], [ 3., 1., 5.], [ 0., 2., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 -21.0 lon (X) float64 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) array([[nan, 1., 3.], [ 0., 1., 5.], [ 5., nan, 5.], [ 3., nan, nan], [ 0., 2., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 -21.0 lon (X) float64 10.0 10.25 10.5 Dimensions without coordinates: Y, X """ from xarray.core.missing import ffill return ffill(self, dim, limit=limit) def bfill( self: T_DataArray, dim: Hashable, limit: int | None = None ) -> T_DataArray: """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 : 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 array([[ 0., 1., 3.], [ 0., nan, 5.], [ 5., nan, nan], [ 3., nan, nan], [nan, 2., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 -21.0 lon (X) float64 10.0 10.25 10.5 Dimensions without coordinates: Y, X Fill all NaN values: >>> da.bfill(dim="Y", limit=None) array([[ 0., 1., 3.], [ 0., 2., 5.], [ 5., 2., 0.], [ 3., 2., 0.], [nan, 2., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 -21.0 lon (X) float64 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) array([[ 0., 1., 3.], [ 0., nan, 5.], [ 5., nan, nan], [ 3., 2., 0.], [nan, 2., 0.]]) Coordinates: lat (Y) float64 -20.0 -20.25 -20.5 -20.75 -21.0 lon (X) float64 10.0 10.25 10.5 Dimensions without coordinates: Y, X """ from xarray.core.missing import bfill return bfill(self, dim, limit=limit) def combine_first(self: T_DataArray, other: T_DataArray) -> T_DataArray: """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. Returns ------- DataArray """ return ops.fillna(self, other, join="outer") def reduce( self: T_DataArray, func: Callable[..., Any], dim: Dims = None, *, axis: int | Sequence[int] | None = None, keep_attrs: bool | None = None, keepdims: bool = False, **kwargs: Any, ) -> T_DataArray: """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 DataArray with this object's array replaced with an array with summarized data and the indicated dimension(s) removed. """ var = self.variable.reduce(func, dim, axis, keep_attrs, keepdims, **kwargs) return self._replace_maybe_drop_dims(var) def to_pandas(self) -> DataArray | pd.Series | pd.DataFrame: """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 | Series | DataFrame DataArray, pandas Series or pandas DataFrame. """ # TODO: consolidate the info about pandas constructors and the # attributes that correspond to their indexes into a separate module? constructors = {0: lambda x: x, 1: pd.Series, 2: pd.DataFrame} try: constructor = constructors[self.ndim] except KeyError: raise ValueError( f"Cannot convert arrays with {self.ndim} dimensions into " "pandas objects. Requires 2 or fewer dimensions." ) indexes = [self.get_index(dim) for dim in self.dims] return constructor(self.values, *indexes) def to_dataframe( self, name: Hashable | None = None, dim_order: Sequence[Hashable] | None = None ) -> pd.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 :py:class:`pandas.MultiIndex`). Other coordinates are included as columns in the DataFrame. For 1D and 2D DataArrays, see also :py:func:`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: DataFrame DataArray as a pandas DataFrame. See also -------- DataArray.to_pandas DataArray.to_series """ if name is None: name = self.name if name is None: raise ValueError( "cannot convert an unnamed DataArray to a " "DataFrame: use the ``name`` parameter" ) if self.ndim == 0: raise ValueError("cannot convert a scalar to a DataFrame") # By using a unique name, we can convert a DataArray into a DataFrame # even if it shares a name with one of its coordinates. # I would normally use unique_name = object() but that results in a # dataframe with columns in the wrong order, for reasons I have not # been able to debug (possibly a pandas bug?). unique_name = "__unique_name_identifier_z98xfz98xugfg73ho__" ds = self._to_dataset_whole(name=unique_name) if dim_order is None: ordered_dims = dict(zip(self.dims, self.shape)) else: ordered_dims = ds._normalize_dim_order(dim_order=dim_order) df = ds._to_dataframe(ordered_dims) df.columns = [name if c == unique_name else c for c in df.columns] return df def to_series(self) -> pd.Series: """Convert this array into a pandas.Series. The Series is indexed by the Cartesian product of index coordinates (in the form of a :py:class:`pandas.MultiIndex`). Returns ------- result : Series DataArray as a pandas Series. See also -------- DataArray.to_pandas DataArray.to_dataframe """ index = self.coords.to_index() return pd.Series(self.values.reshape(-1), index=index, name=self.name) def to_masked_array(self, copy: bool = True) -> np.ma.MaskedArray: """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 : MaskedArray Masked where invalid values (nan or inf) occur. """ values = self.to_numpy() # only compute lazy arrays once isnull = pd.isnull(values) return np.ma.MaskedArray(data=values, mask=isnull, copy=copy) # path=None writes to bytes @overload def to_netcdf( self, path: None = None, mode: Literal["w", "a"] = "w", format: T_NetcdfTypes | None = None, group: str | None = None, engine: T_NetcdfEngine | None = None, encoding: Mapping[Hashable, Mapping[str, Any]] | None = None, unlimited_dims: Iterable[Hashable] | None = None, compute: bool = True, invalid_netcdf: bool = False, ) -> bytes: ... # default return None @overload def to_netcdf( self, path: str | PathLike, mode: Literal["w", "a"] = "w", format: T_NetcdfTypes | None = None, group: str | None = None, engine: T_NetcdfEngine | None = None, encoding: Mapping[Hashable, Mapping[str, Any]] | None = None, unlimited_dims: Iterable[Hashable] | None = None, compute: Literal[True] = True, invalid_netcdf: bool = False, ) -> None: ... # compute=False returns dask.Delayed @overload def to_netcdf( self, path: str | PathLike, mode: Literal["w", "a"] = "w", format: T_NetcdfTypes | None = None, group: str | None = None, engine: T_NetcdfEngine | None = None, encoding: Mapping[Hashable, Mapping[str, Any]] | None = None, unlimited_dims: Iterable[Hashable] | None = None, *, compute: Literal[False], invalid_netcdf: bool = False, ) -> Delayed: ... def to_netcdf( self, path: str | PathLike | None = None, mode: Literal["w", "a"] = "w", format: T_NetcdfTypes | None = None, group: str | None = None, engine: T_NetcdfEngine | None = None, encoding: Mapping[Hashable, Mapping[str, Any]] | None = None, unlimited_dims: Iterable[Hashable] | None = None, compute: bool = True, invalid_netcdf: bool = False, ) -> bytes | Delayed | None: """Write dataset contents to a netCDF file. Parameters ---------- path : str, path-like or None, optional Path to which to save this dataset. File-like objects are only supported by the scipy engine. If no path is provided, this function returns the resulting netCDF file as bytes; in this case, we need to use scipy, which does not support netCDF version 4 (the default format becomes NETCDF3_64BIT). 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 https://github.com/h5netcdf/h5netcdf. Returns ------- store: bytes or Delayed or None * ``bytes`` 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__"``. See Also -------- Dataset.to_netcdf """ from xarray.backends.api import DATAARRAY_NAME, DATAARRAY_VARIABLE, to_netcdf if self.name is None: # If no name is set then use a generic xarray name dataset = self.to_dataset(name=DATAARRAY_VARIABLE) elif self.name in self.coords or self.name in self.dims: # The name is the same as one of the coords names, which netCDF # doesn't support, so rename it but keep track of the old name dataset = self.to_dataset(name=DATAARRAY_VARIABLE) dataset.attrs[DATAARRAY_NAME] = self.name else: # No problems with the name - so we're fine! dataset = self.to_dataset() return to_netcdf( # type: ignore # mypy cannot resolve the overloads:( dataset, path, mode=mode, format=format, group=group, engine=engine, encoding=encoding, unlimited_dims=unlimited_dims, compute=compute, multifile=False, invalid_netcdf=invalid_netcdf, ) def to_dict(self, data: bool = True, encoding: bool = False) -> dict[str, Any]: """ 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, default: True Whether to include the actual data in the dictionary. When set to False, returns just the schema. encoding : bool, default: False Whether to include the Dataset's encoding in the dictionary. Returns ------- dict: dict See Also -------- DataArray.from_dict Dataset.to_dict """ d = self.variable.to_dict(data=data) d.update({"coords": {}, "name": self.name}) for k, coord in self.coords.items(): d["coords"][k] = coord.variable.to_dict(data=data) if encoding: d["encoding"] = dict(self.encoding) return d @classmethod def from_dict(cls: type[T_DataArray], d: Mapping[str, Any]) -> T_DataArray: """Convert a dictionary into an xarray.DataArray Parameters ---------- d : dict Mapping with a minimum structure of {"dims": [...], "data": [...]} Returns ------- obj : 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 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 array([10, 20, 30]) Coordinates: * t (t) int64 0 1 2 Attributes: title: air temperature """ coords = None if "coords" in d: try: coords = { k: (v["dims"], v["data"], v.get("attrs")) for k, v in d["coords"].items() } except KeyError as e: raise ValueError( "cannot convert dict when coords are missing the key " "'{dims_data}'".format(dims_data=str(e.args[0])) ) try: data = d["data"] except KeyError: raise ValueError("cannot convert dict without the key 'data''") else: obj = cls(data, coords, d.get("dims"), d.get("name"), d.get("attrs")) obj.encoding.update(d.get("encoding", {})) return obj @classmethod def from_series(cls, series: pd.Series, sparse: bool = False) -> DataArray: """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 """ temp_name = "__temporary_name" df = pd.DataFrame({temp_name: series}) ds = Dataset.from_dataframe(df, sparse=sparse) result = cast(DataArray, ds[temp_name]) result.name = series.name return result def to_cdms2(self) -> cdms2_Variable: """Convert this array into a cdms2.Variable""" from xarray.convert import to_cdms2 return to_cdms2(self) @classmethod def from_cdms2(cls, variable: cdms2_Variable) -> DataArray: """Convert a cdms2.Variable into an xarray.DataArray""" from xarray.convert import from_cdms2 return from_cdms2(variable) def to_iris(self) -> iris_Cube: """Convert this array into a iris.cube.Cube""" from xarray.convert import to_iris return to_iris(self) @classmethod def from_iris(cls, cube: iris_Cube) -> DataArray: """Convert a iris.cube.Cube into an xarray.DataArray""" from xarray.convert import from_iris return from_iris(cube) def _all_compat(self: T_DataArray, other: T_DataArray, compat_str: str) -> bool: """Helper function for equals, broadcast_equals, and identical""" def compat(x, y): return getattr(x.variable, compat_str)(y.variable) return utils.dict_equiv(self.coords, other.coords, compat=compat) and compat( self, other ) def broadcast_equals(self: T_DataArray, other: T_DataArray) -> bool: """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 : bool True if the two DataArrays are broadcast equal. 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 array([1, 2]) Dimensions without coordinates: X >>> b 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 eachother 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 """ try: return self._all_compat(other, "broadcast_equals") except (TypeError, AttributeError): return False def equals(self: T_DataArray, other: T_DataArray) -> bool: """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 : bool True if the two DataArrays are equal. 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 array([1, 2, 3]) Dimensions without coordinates: X >>> b array([1, 2, 3]) Dimensions without coordinates: X Attributes: units: m >>> c array([1, 2, 3]) Dimensions without coordinates: Y >>> d array([3, 2, 1]) Dimensions without coordinates: X >>> a.equals(b) True >>> a.equals(c) False >>> a.equals(d) False """ try: return self._all_compat(other, "equals") except (TypeError, AttributeError): return False def identical(self: T_DataArray, other: T_DataArray) -> bool: """Like equals, but also checks the array name and attributes, and attributes on all coordinates. Parameters ---------- other : DataArray DataArray to compare to. Returns ---------- equal : bool True if the two DataArrays are identical. 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 array([1, 2, 3]) Dimensions without coordinates: X Attributes: units: m >>> b array([1, 2, 3]) Dimensions without coordinates: X Attributes: units: m >>> c array([1, 2, 3]) Dimensions without coordinates: X Attributes: units: ft >>> a.equals(b) True >>> a.identical(b) True >>> a.equals(c) True >>> a.identical(c) False """ try: return self.name == other.name and self._all_compat(other, "identical") except (TypeError, AttributeError): return False def _result_name(self, other: Any = None) -> Hashable | None: # use the same naming heuristics as pandas: # https://github.com/ContinuumIO/blaze/issues/458#issuecomment-51936356 other_name = getattr(other, "name", _default) if other_name is _default or other_name == self.name: return self.name else: return None def __array_wrap__(self: T_DataArray, obj, context=None) -> T_DataArray: new_var = self.variable.__array_wrap__(obj, context) return self._replace(new_var) def __matmul__(self: T_DataArray, obj: T_DataArray) -> T_DataArray: return self.dot(obj) def __rmatmul__(self: T_DataArray, other: T_DataArray) -> T_DataArray: # currently somewhat duplicative, as only other DataArrays are # compatible with matmul return computation.dot(other, self) def _unary_op(self: T_DataArray, f: Callable, *args, **kwargs) -> T_DataArray: keep_attrs = kwargs.pop("keep_attrs", None) if keep_attrs is None: keep_attrs = _get_keep_attrs(default=True) with warnings.catch_warnings(): warnings.filterwarnings("ignore", r"All-NaN (slice|axis) encountered") warnings.filterwarnings( "ignore", r"Mean of empty slice", category=RuntimeWarning ) with np.errstate(all="ignore"): da = self.__array_wrap__(f(self.variable.data, *args, **kwargs)) if keep_attrs: da.attrs = self.attrs return da def _binary_op( self: T_DataArray, other: Any, f: Callable, reflexive: bool = False, ) -> T_DataArray: from xarray.core.groupby import GroupBy if isinstance(other, (Dataset, GroupBy)): return NotImplemented if isinstance(other, DataArray): align_type = OPTIONS["arithmetic_join"] self, other = align(self, other, join=align_type, copy=False) # type: ignore other_variable = getattr(other, "variable", other) other_coords = getattr(other, "coords", None) variable = ( f(self.variable, other_variable) if not reflexive else f(other_variable, self.variable) ) coords, indexes = self.coords._merge_raw(other_coords, reflexive) name = self._result_name(other) return self._replace(variable, coords, name, indexes=indexes) def _inplace_binary_op(self: T_DataArray, other: Any, f: Callable) -> T_DataArray: from xarray.core.groupby import GroupBy if isinstance(other, GroupBy): raise TypeError( "in-place operations between a DataArray and " "a grouped object are not permitted" ) # n.b. we can't align other to self (with other.reindex_like(self)) # because `other` may be converted into floats, which would cause # in-place arithmetic to fail unpredictably. Instead, we simply # don't support automatic alignment with in-place arithmetic. other_coords = getattr(other, "coords", None) other_variable = getattr(other, "variable", other) try: with self.coords._merge_inplace(other_coords): f(self.variable, other_variable) except MergeError as exc: raise MergeError( "Automatic alignment is not supported for in-place operations.\n" "Consider aligning the indices manually or using a not-in-place operation.\n" "See https://github.com/pydata/xarray/issues/3910 for more explanations." ) from exc return self def _copy_attrs_from(self, other: DataArray | Dataset | Variable) -> None: self.attrs = other.attrs plot = utils.UncachedAccessor(DataArrayPlotAccessor) def _title_for_slice(self, truncate: int = 50) -> str: """ If the dataarray has 1 dimensional coordinates or comes from a slice we can show that info in the title Parameters ---------- truncate : int, default: 50 maximum number of characters for title Returns ------- title : string Can be used for plot titles """ one_dims = [] for dim, coord in self.coords.items(): if coord.size == 1: one_dims.append( "{dim} = {v}{unit}".format( dim=dim, v=format_item(coord.values), unit=_get_units_from_attrs(coord), ) ) title = ", ".join(one_dims) if len(title) > truncate: title = title[: (truncate - 3)] + "..." return title def diff( self: T_DataArray, dim: Hashable, n: int = 1, label: Literal["upper", "lower"] = "upper", ) -> T_DataArray: """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 : DataArray The n-th order finite difference of this object. 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") array([0, 1, 0]) Coordinates: * x (x) int64 2 3 4 >>> arr.diff("x", 2) array([ 1, -1]) Coordinates: * x (x) int64 3 4 See Also -------- DataArray.differentiate """ ds = self._to_temp_dataset().diff(n=n, dim=dim, label=label) return self._from_temp_dataset(ds) def shift( self: T_DataArray, shifts: Mapping[Any, int] | None = None, fill_value: Any = dtypes.NA, **shifts_kwargs: int, ) -> T_DataArray: """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 DataArray with the same coordinates and attributes but shifted data. See Also -------- roll Examples -------- >>> arr = xr.DataArray([5, 6, 7], dims="x") >>> arr.shift(x=1) array([nan, 5., 6.]) Dimensions without coordinates: x """ variable = self.variable.shift( shifts=shifts, fill_value=fill_value, **shifts_kwargs ) return self._replace(variable=variable) def roll( self: T_DataArray, shifts: Mapping[Hashable, int] | None = None, roll_coords: bool = False, **shifts_kwargs: int, ) -> T_DataArray: """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 :py:func:`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 DataArray with the same attributes but rolled data and coordinates. See Also -------- shift Examples -------- >>> arr = xr.DataArray([5, 6, 7], dims="x") >>> arr.roll(x=1) array([7, 5, 6]) Dimensions without coordinates: x """ ds = self._to_temp_dataset().roll( shifts=shifts, roll_coords=roll_coords, **shifts_kwargs ) return self._from_temp_dataset(ds) @property def real(self: T_DataArray) -> T_DataArray: """ The real part of the array. See Also -------- numpy.ndarray.real """ return self._replace(self.variable.real) @property def imag(self: T_DataArray) -> T_DataArray: """ The imaginary part of the array. See Also -------- numpy.ndarray.imag """ return self._replace(self.variable.imag) def dot( self: T_DataArray, other: T_DataArray, dims: Dims = None, ) -> T_DataArray: """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. dims : ..., 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 : DataArray Array resulting from the dot product over all shared dimensions. 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') """ if isinstance(other, Dataset): raise NotImplementedError( "dot products are not yet supported with Dataset objects." ) if not isinstance(other, DataArray): raise TypeError("dot only operates on DataArrays.") return computation.dot(self, other, dims=dims) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def sortby( self, variables: Hashable | DataArray | Sequence[Hashable | DataArray], ascending: bool = True, ) -> DataArray: """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, or sequence of Hashable or DataArray 1D DataArray objects or name(s) of 1D variable(s) in coords whose values are used to sort this array. ascending : bool, default: True Whether to sort by ascending or descending order. Returns ------- sorted : DataArray A new dataarray where all the specified dims are sorted by dim labels. See Also -------- Dataset.sortby numpy.sort pandas.sort_values pandas.sort_index Examples -------- >>> da = xr.DataArray( ... np.random.rand(5), ... coords=[pd.date_range("1/1/2000", periods=5)], ... dims="time", ... ) >>> da array([0.5488135 , 0.71518937, 0.60276338, 0.54488318, 0.4236548 ]) Coordinates: * time (time) datetime64[ns] 2000-01-01 2000-01-02 ... 2000-01-05 >>> da.sortby(da) array([0.4236548 , 0.54488318, 0.5488135 , 0.60276338, 0.71518937]) Coordinates: * time (time) datetime64[ns] 2000-01-05 2000-01-04 ... 2000-01-02 """ ds = self._to_temp_dataset().sortby(variables, ascending=ascending) return self._from_temp_dataset(ds) def quantile( self: T_DataArray, q: ArrayLike, dim: Dims = None, method: QuantileMethods = "linear", keep_attrs: bool | None = None, skipna: bool | None = None, interpolation: QuantileMethods | None = None, ) -> T_DataArray: """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 :py:func:`numpy.quantile` or [1]_ for details. The "method" argument was previously called "interpolation", renamed in accordance with numpy version 1.22.0. (*) These methods require numpy version 1.22 or newer. 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 : DataArray 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. 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=...) array(0.7) Coordinates: quantile float64 0.0 >>> da.quantile(0, dim="x") array([0.7, 4.2, 2.6, 1.5]) Coordinates: * y (y) float64 1.0 1.5 2.0 2.5 quantile float64 0.0 >>> da.quantile([0, 0.5, 1]) array([0.7, 3.4, 9.4]) Coordinates: * quantile (quantile) float64 0.0 0.5 1.0 >>> da.quantile([0, 0.5, 1], dim="x") 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 1.0 1.5 2.0 2.5 * quantile (quantile) float64 0.0 0.5 1.0 References ---------- .. [1] R. J. Hyndman and Y. Fan, "Sample quantiles in statistical packages," The American Statistician, 50(4), pp. 361-365, 1996 """ ds = self._to_temp_dataset().quantile( q, dim=dim, keep_attrs=keep_attrs, method=method, skipna=skipna, interpolation=interpolation, ) return self._from_temp_dataset(ds) def rank( self: T_DataArray, dim: Hashable, pct: bool = False, keep_attrs: bool | None = None, ) -> T_DataArray: """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 DataArray with the same coordinates and dtype 'float64'. Examples -------- >>> arr = xr.DataArray([5, 6, 7], dims="x") >>> arr.rank("x") array([1., 2., 3.]) Dimensions without coordinates: x """ ds = self._to_temp_dataset().rank(dim, pct=pct, keep_attrs=keep_attrs) return self._from_temp_dataset(ds) def differentiate( self: T_DataArray, coord: Hashable, edge_order: Literal[1, 2] = 1, datetime_unit: DatetimeUnitOptions = None, ) -> T_DataArray: """ 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 : {"Y", "M", "W", "D", "h", "m", "s", "ms", \ "us", "ns", "ps", "fs", "as", None}, optional Unit to compute gradient. Only valid for datetime coordinate. Returns ------- differentiated: 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 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) float64 0.0 0.1 1.1 1.2 Dimensions without coordinates: y >>> >>> da.differentiate("x") array([[30. , 30. , 30. ], [27.54545455, 27.54545455, 27.54545455], [27.54545455, 27.54545455, 27.54545455], [30. , 30. , 30. ]]) Coordinates: * x (x) float64 0.0 0.1 1.1 1.2 Dimensions without coordinates: y """ ds = self._to_temp_dataset().differentiate(coord, edge_order, datetime_unit) return self._from_temp_dataset(ds) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def integrate( self, coord: Hashable | Sequence[Hashable] = None, datetime_unit: DatetimeUnitOptions = None, ) -> DataArray: """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 : {'Y', 'M', 'W', 'D', 'h', 'm', 's', 'ms', 'us', 'ns', \ 'ps', 'fs', 'as', None}, optional Specify the unit if a datetime coordinate is used. Returns ------- integrated : 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 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) float64 0.0 0.1 1.1 1.2 Dimensions without coordinates: y >>> >>> da.integrate("x") array([5.4, 6.6, 7.8]) Dimensions without coordinates: y """ ds = self._to_temp_dataset().integrate(coord, datetime_unit) return self._from_temp_dataset(ds) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def cumulative_integrate( self, coord: Hashable | Sequence[Hashable] = None, datetime_unit: DatetimeUnitOptions = None, ) -> DataArray: """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 : {'Y', 'M', 'W', 'D', 'h', 'm', 's', 'ms', 'us', 'ns', \ 'ps', 'fs', 'as', None}, optional Specify the unit if a datetime coordinate is used. Returns ------- integrated : 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 array([[ 0, 1, 2], [ 3, 4, 5], [ 6, 7, 8], [ 9, 10, 11]]) Coordinates: * x (x) float64 0.0 0.1 1.1 1.2 Dimensions without coordinates: y >>> >>> da.cumulative_integrate("x") 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 0.0 0.1 1.1 1.2 Dimensions without coordinates: y """ ds = self._to_temp_dataset().cumulative_integrate(coord, datetime_unit) return self._from_temp_dataset(ds) def unify_chunks(self) -> DataArray: """Unify chunk size along all chunked dimensions of this DataArray. Returns ------- DataArray with consistent chunk sizes for all dask-array variables See Also -------- dask.array.core.unify_chunks """ return unify_chunks(self)[0] def map_blocks( self, func: Callable[..., T_Xarray], args: Sequence[Any] = (), kwargs: Mapping[str, Any] | None = None, template: DataArray | Dataset | None = None, ) -> T_Xarray: """ Apply a function to each block of this DataArray. .. warning:: This method is experimental and its signature may change. Parameters ---------- func : callable User-provided function that accepts a DataArray as its first parameter. The function will receive a subset or 'block' of this DataArray (see below), corresponding to one chunk along each chunked dimension. ``func`` will be executed as ``func(subset_dataarray, *subset_args, **kwargs)``. This function must return either a single DataArray or a single Dataset. This function cannot add a new chunked dimension. args : sequence Passed to func after unpacking and subsetting any xarray objects by blocks. xarray objects in args must be aligned with this object, otherwise an error is raised. kwargs : mapping Passed verbatim to func after unpacking. xarray objects, if any, will not be subset to blocks. Passing dask collections in kwargs is not allowed. template : DataArray or Dataset, optional xarray object representing the final result after compute is called. If not provided, the function will be first run on mocked-up data, that looks like this object but has sizes 0, to determine properties of the returned object such as dtype, variable names, attributes, new dimensions and new indexes (if any). ``template`` must be provided if the function changes the size of existing dimensions. When provided, ``attrs`` on variables in `template` are copied over to the result. Any ``attrs`` set by ``func`` will be ignored. Returns ------- A single DataArray or Dataset with dask backend, reassembled from the outputs of the function. Notes ----- This function is designed for when ``func`` needs to manipulate a whole xarray object subset to each block. Each block is loaded into memory. In the more common case where ``func`` can work on numpy arrays, it is recommended to use ``apply_ufunc``. If none of the variables in this object is backed by dask arrays, calling this function is equivalent to calling ``func(obj, *args, **kwargs)``. See Also -------- dask.array.map_blocks, xarray.apply_ufunc, xarray.Dataset.map_blocks xarray.DataArray.map_blocks Examples -------- Calculate an anomaly from climatology using ``.groupby()``. Using ``xr.map_blocks()`` allows for parallel operations with knowledge of ``xarray``, its indices, and its methods like ``.groupby()``. >>> def calculate_anomaly(da, groupby_type="time.month"): ... gb = da.groupby(groupby_type) ... clim = gb.mean(dim="time") ... return gb - clim ... >>> time = xr.cftime_range("1990-01", "1992-01", freq="M") >>> month = xr.DataArray(time.month, coords={"time": time}, dims=["time"]) >>> np.random.seed(123) >>> array = xr.DataArray( ... np.random.rand(len(time)), ... dims=["time"], ... coords={"time": time, "month": month}, ... ).chunk() >>> array.map_blocks(calculate_anomaly, template=array).compute() array([ 0.12894847, 0.11323072, -0.0855964 , -0.09334032, 0.26848862, 0.12382735, 0.22460641, 0.07650108, -0.07673453, -0.22865714, -0.19063865, 0.0590131 , -0.12894847, -0.11323072, 0.0855964 , 0.09334032, -0.26848862, -0.12382735, -0.22460641, -0.07650108, 0.07673453, 0.22865714, 0.19063865, -0.0590131 ]) Coordinates: * time (time) object 1990-01-31 00:00:00 ... 1991-12-31 00:00:00 month (time) int64 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12 Note that one must explicitly use ``args=[]`` and ``kwargs={}`` to pass arguments to the function being applied in ``xr.map_blocks()``: >>> array.map_blocks( ... calculate_anomaly, kwargs={"groupby_type": "time.year"}, template=array ... ) # doctest: +ELLIPSIS dask.array<-calculate_anomaly, shape=(24,), dtype=float64, chunksize=(24,), chunktype=numpy.ndarray> Coordinates: * time (time) object 1990-01-31 00:00:00 ... 1991-12-31 00:00:00 month (time) int64 dask.array """ from xarray.core.parallel import map_blocks return map_blocks(func, self, args, kwargs, template) def polyfit( self, dim: Hashable, deg: int, skipna: bool | None = None, rcond: float | None = None, w: Hashable | Any | None = None, full: bool = False, cov: bool | Literal["unscaled"] = False, ) -> Dataset: """ 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 : Dataset 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`) See Also -------- numpy.polyfit numpy.polyval xarray.polyval """ return self._to_temp_dataset().polyfit( dim, deg, skipna=skipna, rcond=rcond, w=w, full=full, cov=cov ) def pad( self: T_DataArray, pad_width: Mapping[Any, int | tuple[int, int]] | None = None, mode: PadModeOptions = "constant", stat_length: int | tuple[int, int] | Mapping[Any, tuple[int, int]] | None = None, constant_values: float | tuple[float, float] | Mapping[Any, tuple[float, float]] | None = None, end_values: int | tuple[int, int] | Mapping[Any, tuple[int, int]] | None = None, reflect_type: PadReflectOptions = None, keep_attrs: bool | None = None, **pad_width_kwargs: Any, ) -> T_DataArray: """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 DataArray with the padded coordinates and data. 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) array([0, 5, 6, 7, 0, 0]) Coordinates: * x (x) float64 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) array([[nan, nan, nan, nan], [ 0., 1., 2., 3.], [10., 11., 12., 13.], [nan, nan, nan, nan]]) Coordinates: * x (x) float64 nan 0.0 1.0 nan * y (y) int64 10 20 30 40 z (x) float64 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) array([[ 1, 1, 1, 1], [ 0, 1, 2, 3], [10, 11, 12, 13], [ 1, 1, 1, 1]]) Coordinates: * x (x) float64 nan 0.0 1.0 nan * y (y) int64 10 20 30 40 z (x) float64 nan 100.0 200.0 nan """ ds = self._to_temp_dataset().pad( pad_width=pad_width, mode=mode, stat_length=stat_length, constant_values=constant_values, end_values=end_values, reflect_type=reflect_type, keep_attrs=keep_attrs, **pad_width_kwargs, ) return self._from_temp_dataset(ds) def idxmin( self, dim: Hashable | None = None, skipna: bool | None = None, fill_value: Any = dtypes.NA, keep_attrs: bool | None = None, ) -> DataArray: """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 :py:meth:`~DataArray.argmin`, this returns the coordinate label while :py:meth:`~DataArray.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 : DataArray New `DataArray` object with `idxmin` applied to its data and the indicated dimension removed. 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() array(-2) >>> array.argmin(...) {'x': array(4)} >>> array.idxmin() array('e', dtype='>> 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") array([-2., -4., 1.]) Coordinates: * y (y) int64 -1 0 1 >>> array.argmin(dim="x") array([4, 0, 2]) Coordinates: * y (y) int64 -1 0 1 >>> array.idxmin(dim="x") array([16., 0., 4.]) Coordinates: * y (y) int64 -1 0 1 """ return computation._calc_idxminmax( array=self, func=lambda x, *args, **kwargs: x.argmin(*args, **kwargs), dim=dim, skipna=skipna, fill_value=fill_value, keep_attrs=keep_attrs, ) def idxmax( self, dim: Hashable = None, skipna: bool | None = None, fill_value: Any = dtypes.NA, keep_attrs: bool | None = None, ) -> DataArray: """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 :py:meth:`~DataArray.argmax`, this returns the coordinate label while :py:meth:`~DataArray.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 : DataArray New `DataArray` object with `idxmax` applied to its data and the indicated dimension removed. 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() array(2) >>> array.argmax(...) {'x': array(1)} >>> array.idxmax() array('b', dtype='>> 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") array([2., 2., 1.]) Coordinates: * y (y) int64 -1 0 1 >>> array.argmax(dim="x") array([0, 2, 2]) Coordinates: * y (y) int64 -1 0 1 >>> array.idxmax(dim="x") array([0., 4., 4.]) Coordinates: * y (y) int64 -1 0 1 """ return computation._calc_idxminmax( array=self, func=lambda x, *args, **kwargs: x.argmax(*args, **kwargs), dim=dim, skipna=skipna, fill_value=fill_value, keep_attrs=keep_attrs, ) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def argmin( self, dim: Dims = None, axis: int | None = None, keep_attrs: bool | None = None, skipna: bool | None = None, ) -> DataArray | dict[Hashable, DataArray]: """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 : DataArray or dict of DataArray See Also -------- Variable.argmin, DataArray.idxmin Examples -------- >>> array = xr.DataArray([0, 2, -1, 3], dims="x") >>> array.min() array(-1) >>> array.argmin(...) {'x': array(2)} >>> array.isel(array.argmin(...)) 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") array([[ 1, 2, 1], [ 2, -5, 1], [ 2, 1, 1]]) Dimensions without coordinates: y, z >>> array.argmin(dim="x") array([[1, 0, 0], [1, 1, 1], [0, 0, 1]]) Dimensions without coordinates: y, z >>> array.argmin(dim=["x"]) {'x': array([[1, 0, 0], [1, 1, 1], [0, 0, 1]]) Dimensions without coordinates: y, z} >>> array.min(dim=("x", "z")) array([ 1, -5, 1]) Dimensions without coordinates: y >>> array.argmin(dim=["x", "z"]) {'x': array([0, 1, 0]) Dimensions without coordinates: y, 'z': array([2, 1, 1]) Dimensions without coordinates: y} >>> array.isel(array.argmin(dim=["x", "z"])) array([ 1, -5, 1]) Dimensions without coordinates: y """ result = self.variable.argmin(dim, axis, keep_attrs, skipna) if isinstance(result, dict): return {k: self._replace_maybe_drop_dims(v) for k, v in result.items()} else: return self._replace_maybe_drop_dims(result) # change type of self and return to T_DataArray once # https://github.com/python/mypy/issues/12846 is resolved def argmax( self, dim: Dims = None, axis: int | None = None, keep_attrs: bool | None = None, skipna: bool | None = None, ) -> DataArray | dict[Hashable, DataArray]: """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 : DataArray or dict of DataArray See Also -------- Variable.argmax, DataArray.idxmax Examples -------- >>> array = xr.DataArray([0, 2, -1, 3], dims="x") >>> array.max() array(3) >>> array.argmax(...) {'x': array(3)} >>> array.isel(array.argmax(...)) 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") array([[3, 3, 2], [3, 5, 2], [2, 3, 3]]) Dimensions without coordinates: y, z >>> array.argmax(dim="x") array([[0, 1, 1], [0, 1, 0], [0, 1, 0]]) Dimensions without coordinates: y, z >>> array.argmax(dim=["x"]) {'x': array([[0, 1, 1], [0, 1, 0], [0, 1, 0]]) Dimensions without coordinates: y, z} >>> array.max(dim=("x", "z")) array([3, 5, 3]) Dimensions without coordinates: y >>> array.argmax(dim=["x", "z"]) {'x': array([0, 1, 0]) Dimensions without coordinates: y, 'z': array([0, 1, 2]) Dimensions without coordinates: y} >>> array.isel(array.argmax(dim=["x", "z"])) array([3, 5, 3]) Dimensions without coordinates: y """ result = self.variable.argmax(dim, axis, keep_attrs, skipna) if isinstance(result, dict): return {k: self._replace_maybe_drop_dims(v) for k, v in result.items()} else: return self._replace_maybe_drop_dims(result) def query( self, queries: Mapping[Any, Any] | None = None, parser: QueryParserOptions = "pandas", engine: QueryEngineOptions = None, missing_dims: ErrorOptionsWithWarn = "raise", **queries_kwargs: Any, ) -> DataArray: """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 : DataArray A new DataArray with the same contents as this dataset, indexed by the results of the appropriate queries. See Also -------- DataArray.isel Dataset.query pandas.eval Examples -------- >>> da = xr.DataArray(np.arange(0, 5, 1), dims="x", name="a") >>> da array([0, 1, 2, 3, 4]) Dimensions without coordinates: x >>> da.query(x="a > 2") array([3, 4]) Dimensions without coordinates: x """ ds = self._to_dataset_whole(shallow_copy=True) ds = ds.query( queries=queries, parser=parser, engine=engine, missing_dims=missing_dims, **queries_kwargs, ) return ds[self.name] def curvefit( self, coords: str | DataArray | Iterable[str | DataArray], func: Callable[..., Any], reduce_dims: Dims = None, skipna: bool = True, p0: dict[str, Any] | None = None, bounds: dict[str, Any] | None = None, param_names: Sequence[str] | None = None, kwargs: dict[str, Any] | None = None, ) -> Dataset: """ 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 none or only some parameters are passed, the rest will be assigned initial values following the default scipy behavior. bounds : dict-like or None, optional Optional dictionary of parameter names to bounding values passed to the `curve_fit` `bounds` arg. 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. **kwargs : optional Additional keyword arguments to passed to scipy curve_fit. Returns ------- curvefit_results : Dataset A single dataset which contains: [var]_curvefit_coefficients The coefficients of the best fit. [var]_curvefit_covariance The covariance matrix of the coefficient estimates. See Also -------- DataArray.polyfit scipy.optimize.curve_fit """ return self._to_temp_dataset().curvefit( coords, func, reduce_dims=reduce_dims, skipna=skipna, p0=p0, bounds=bounds, param_names=param_names, kwargs=kwargs, ) def drop_duplicates( self: T_DataArray, dim: Hashable | Iterable[Hashable], keep: Literal["first", "last", False] = "first", ) -> T_DataArray: """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. Returns ------- 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 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 0 0 1 2 3 * y (y) int64 0 1 2 3 3 >>> da.drop_duplicates(dim="x") 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 0 1 2 3 * y (y) int64 0 1 2 3 3 >>> da.drop_duplicates(dim="x", keep="last") 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 0 1 2 3 * y (y) int64 0 1 2 3 3 Drop all duplicate dimension values: >>> da.drop_duplicates(dim=...) array([[ 0, 1, 2, 3], [10, 11, 12, 13], [15, 16, 17, 18], [20, 21, 22, 23]]) Coordinates: * x (x) int64 0 1 2 3 * y (y) int64 0 1 2 3 """ deduplicated = self._to_temp_dataset().drop_duplicates(dim, keep=keep) return self._from_temp_dataset(deduplicated) def convert_calendar( self, calendar: str, dim: str = "time", align_on: str | None = None, missing: Any | None = None, use_cftime: bool | None = None, ) -> DataArray: """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 :py:func:`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 :py:class:`numpy.datetime64` values if the date range permits it, and :py:class:`cftime.datetime` objects if not. If False, it uses :py:class:`numpy.datetime64` or fails. Returns ------- DataArray 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`. Notes ----- Passing a value to `missing` is only usable if the source's time coordinate as an inferable frequencies (see :py:func:`~xarray.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. """ return convert_calendar( self, calendar, dim=dim, align_on=align_on, missing=missing, use_cftime=use_cftime, ) def interp_calendar( self, target: pd.DatetimeIndex | CFTimeIndex | DataArray, dim: str = "time", ) -> DataArray: """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. Return ------ DataArray The source interpolated on the decimal years of target, """ return interp_calendar(self, target, dim=dim) def groupby( self, group: Hashable | DataArray | IndexVariable, squeeze: bool = True, restore_coord_dims: bool = False, ) -> DataArrayGroupBy: """Returns a DataArrayGroupBy object for performing grouped operations. Parameters ---------- group : Hashable, DataArray or IndexVariable Array whose unique values should be used to group this array. If a Hashable, must be the name of a coordinate contained in this dataarray. squeeze : bool, default: True If "group" is a dimension of any arrays in this dataset, `squeeze` controls whether the subarrays have a dimension of length 1 along that dimension or if the dimension is squeezed out. restore_coord_dims : bool, default: False If True, also restore the dimension order of multi-dimensional coordinates. Returns ------- grouped : DataArrayGroupBy A `DataArrayGroupBy` object patterned after `pandas.GroupBy` that can be iterated over in the form of `(unique_value, grouped_array)` pairs. Examples -------- Calculate daily anomalies for daily data: >>> da = xr.DataArray( ... np.linspace(0, 1826, num=1827), ... coords=[pd.date_range("2000-01-01", "2004-12-31", freq="D")], ... dims="time", ... ) >>> da array([0.000e+00, 1.000e+00, 2.000e+00, ..., 1.824e+03, 1.825e+03, 1.826e+03]) Coordinates: * time (time) datetime64[ns] 2000-01-01 2000-01-02 ... 2004-12-31 >>> da.groupby("time.dayofyear") - da.groupby("time.dayofyear").mean("time") array([-730.8, -730.8, -730.8, ..., 730.2, 730.2, 730.5]) Coordinates: * time (time) datetime64[ns] 2000-01-01 2000-01-02 ... 2004-12-31 dayofyear (time) int64 1 2 3 4 5 6 7 8 ... 359 360 361 362 363 364 365 366 See Also -------- DataArray.groupby_bins Dataset.groupby core.groupby.DataArrayGroupBy pandas.DataFrame.groupby """ from xarray.core.groupby import DataArrayGroupBy # While we don't generally check the type of every arg, passing # multiple dimensions as multiple arguments is common enough, and the # consequences hidden enough (strings evaluate as true) to warrant # checking here. # A future version could make squeeze kwarg only, but would face # backward-compat issues. if not isinstance(squeeze, bool): raise TypeError( f"`squeeze` must be True or False, but {squeeze} was supplied" ) return DataArrayGroupBy( self, group, squeeze=squeeze, restore_coord_dims=restore_coord_dims ) def groupby_bins( self, group: Hashable | DataArray | IndexVariable, bins: ArrayLike, right: bool = True, labels: ArrayLike | Literal[False] | None = None, precision: int = 3, include_lowest: bool = False, squeeze: bool = True, restore_coord_dims: bool = False, ) -> DataArrayGroupBy: """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 : bool, default: True If "group" is a dimension of any arrays in this dataset, `squeeze` controls whether the subarrays have a dimension of length 1 along that dimension or if the dimension is squeezed out. restore_coord_dims : bool, default: False If True, also restore the dimension order of multi-dimensional coordinates. Returns ------- grouped : DataArrayGroupBy 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. See Also -------- DataArray.groupby Dataset.groupby_bins core.groupby.DataArrayGroupBy pandas.DataFrame.groupby References ---------- .. [1] http://pandas.pydata.org/pandas-docs/stable/generated/pandas.cut.html """ from xarray.core.groupby import DataArrayGroupBy return DataArrayGroupBy( self, group, squeeze=squeeze, bins=bins, restore_coord_dims=restore_coord_dims, cut_kwargs={ "right": right, "labels": labels, "precision": precision, "include_lowest": include_lowest, }, ) def weighted(self, weights: DataArray) -> DataArrayWeighted: """ Weighted DataArray operations. Parameters ---------- weights : DataArray An array of weights associated with the values in this Dataset. Each value in the data contributes to the reduction operation according to its associated weight. Notes ----- ``weights`` must be a DataArray and cannot contain missing values. Missing values can be replaced by ``weights.fillna(0)``. Returns ------- core.weighted.DataArrayWeighted See Also -------- Dataset.weighted """ from xarray.core.weighted import DataArrayWeighted return DataArrayWeighted(self, weights) def rolling( self, dim: Mapping[Any, int] | None = None, min_periods: int | None = None, center: bool | Mapping[Any, bool] = False, **window_kwargs: int, ) -> DataArrayRolling: """ 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. **window_kwargs : optional The keyword arguments form of ``dim``. One of dim or window_kwargs must be provided. Returns ------- core.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 array([ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11.]) Coordinates: * time (time) datetime64[ns] 1999-12-15 2000-01-15 ... 2000-11-15 >>> da.rolling(time=3, center=True).mean() array([nan, 1., 2., 3., 4., 5., 6., 7., 8., 9., 10., nan]) Coordinates: * time (time) datetime64[ns] 1999-12-15 2000-01-15 ... 2000-11-15 Remove the NaNs using ``dropna()``: >>> da.rolling(time=3, center=True).mean().dropna("time") array([ 1., 2., 3., 4., 5., 6., 7., 8., 9., 10.]) Coordinates: * time (time) datetime64[ns] 2000-01-15 2000-02-15 ... 2000-10-15 See Also -------- core.rolling.DataArrayRolling Dataset.rolling """ from xarray.core.rolling import DataArrayRolling dim = either_dict_or_kwargs(dim, window_kwargs, "rolling") return DataArrayRolling(self, dim, min_periods=min_periods, center=center) def coarsen( self, dim: Mapping[Any, int] | None = None, boundary: CoarsenBoundaryOptions = "exact", side: SideOptions | Mapping[Any, SideOptions] = "left", coord_func: str | Callable | Mapping[Any, str | Callable] = "mean", **window_kwargs: int, ) -> DataArrayCoarsen: """ Coarsen object for DataArrays. Parameters ---------- dim : mapping of hashable to int, optional Mapping from the dimension name to the window size. boundary : {"exact", "trim", "pad"}, default: "exact" If 'exact', a ValueError will be raised if dimension size is not a multiple of the window size. If 'trim', the excess entries are dropped. If 'pad', NA will be padded. side : {"left", "right"} or mapping of str to {"left", "right"}, default: "left" coord_func : str or mapping of hashable to str, default: "mean" function (name) that is applied to the coordinates, or a mapping from coordinate name to function (name). Returns ------- core.rolling.DataArrayCoarsen Examples -------- Coarsen the long time series by averaging over every four days. >>> da = xr.DataArray( ... np.linspace(0, 364, num=364), ... dims="time", ... coords={"time": pd.date_range("1999-12-15", periods=364)}, ... ) >>> da # +doctest: ELLIPSIS array([ 0. , 1.00275482, 2.00550964, 3.00826446, 4.01101928, 5.0137741 , 6.01652893, 7.01928375, 8.02203857, 9.02479339, 10.02754821, 11.03030303, ... 356.98071625, 357.98347107, 358.9862259 , 359.98898072, 360.99173554, 361.99449036, 362.99724518, 364. ]) Coordinates: * time (time) datetime64[ns] 1999-12-15 1999-12-16 ... 2000-12-12 >>> da.coarsen(time=3, boundary="trim").mean() # +doctest: ELLIPSIS array([ 1.00275482, 4.01101928, 7.01928375, 10.02754821, 13.03581267, 16.04407713, 19.0523416 , 22.06060606, 25.06887052, 28.07713499, 31.08539945, 34.09366391, ... 349.96143251, 352.96969697, 355.97796143, 358.9862259 , 361.99449036]) Coordinates: * time (time) datetime64[ns] 1999-12-16 1999-12-19 ... 2000-12-10 >>> See Also -------- core.rolling.DataArrayCoarsen Dataset.coarsen """ from xarray.core.rolling import DataArrayCoarsen dim = either_dict_or_kwargs(dim, window_kwargs, "coarsen") return DataArrayCoarsen( self, dim, boundary=boundary, side=side, coord_func=coord_func, ) def resample( self, indexer: Mapping[Any, str] | None = None, skipna: bool | None = None, closed: SideOptions | None = None, label: SideOptions | None = None, base: int | None = None, offset: pd.Timedelta | datetime.timedelta | str | None = None, origin: str | DatetimeLike = "start_day", keep_attrs: bool | None = None, loffset: datetime.timedelta | str | None = None, restore_coord_dims: bool | None = None, **indexer_kwargs: str, ) -> DataArrayResample: """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, 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. base : int, optional For frequencies that evenly subdivide 1 day, the "origin" of the aggregated intervals. For example, for "24H" frequency, base could range from 0 through 23. 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. loffset : timedelta or str, optional Offset used to adjust the resampled time labels. Some pandas date offset strings are supported. restore_coord_dims : bool, optional If True, also restore the dimension order of multi-dimensional coordinates. **indexer_kwargs : str The keyword arguments form of ``indexer``. One of indexer or indexer_kwargs must be provided. Returns ------- resampled : core.resample.DataArrayResample This object resampled. 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 array([ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9., 10., 11.]) Coordinates: * time (time) datetime64[ns] 1999-12-15 2000-01-15 ... 2000-11-15 >>> da.resample(time="QS-DEC").mean() array([ 1., 4., 7., 10.]) Coordinates: * time (time) datetime64[ns] 1999-12-01 2000-03-01 2000-06-01 2000-09-01 Upsample monthly time-series data to daily data: >>> da.resample(time="1D").interpolate("linear") # +doctest: ELLIPSIS 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 , ... 10.80645161, 10.83870968, 10.87096774, 10.90322581, 10.93548387, 10.96774194, 11. ]) Coordinates: * time (time) datetime64[ns] 1999-12-15 1999-12-16 ... 2000-11-15 Limit scope of upsampling method >>> da.resample(time="1D").nearest(tolerance="1D") array([ 0., 0., nan, ..., nan, 11., 11.]) Coordinates: * time (time) datetime64[ns] 1999-12-15 1999-12-16 ... 2000-11-15 See Also -------- Dataset.resample pandas.Series.resample pandas.DataFrame.resample References ---------- .. [1] http://pandas.pydata.org/pandas-docs/stable/timeseries.html#offset-aliases """ from xarray.core.resample import DataArrayResample return self._resample( resample_cls=DataArrayResample, indexer=indexer, skipna=skipna, closed=closed, label=label, base=base, offset=offset, origin=origin, keep_attrs=keep_attrs, loffset=loffset, restore_coord_dims=restore_coord_dims, **indexer_kwargs, ) # this needs to be at the end, or mypy will confuse with `str` # https://mypy.readthedocs.io/en/latest/common_issues.html#dealing-with-conflicting-names str = utils.UncachedAccessor(StringAccessor["DataArray"])