XrDiscreteRV.rvs¶

XrDiscreteRV.rvs(*args, size=None, random_state=None, dims=None, coords=None, apply_kwargs=None, **kwargs)¶

Method wrapping .rvs of the input distribution with xarray.apply_ufunc

Usage examples available at Probability distributions.

Parameters:

*argsscalar or array_like or xarray.DataArray, optional: Passed to the scipy distribution after broadcasting.
sizeint of sequence of int, optional: The number of samples to draw per array element. If the distribution parameters broadcast to a (4, 10, 6) shape and size=(5, 3) then the output shape is (5, 3, 4, 10, 6). This differs from the scipy implementation. Here, all broadcasting and alignment is done for you, you give the dimensions the right names, and broadcasting just happens. If size followed scipy behaviour, you’d be forced to broadcast to provide a valid value which would defeat the xarray_einstats goal of handling all alignment and broadcasting for you.
random_stateint or np.Generator, optional: Passed as is to the wrapped scipy distribution
dimshashable or sequence of hashable, optional: Dimension names for the dimensions created due to size. If present it must have the same length as size.
coordsmapping of {hashablearray_like or xarray.DataArray}, optional: Mapping of coordinate name to coordinate values.
apply_kwargsdict, optional: Passed to xarray.apply_ufunc
**kwargs: Passed to the scipy distribution after broadcasting using the same key.

Returns: