删除x数组数据集的值和扁平DataArrays

Question

是否可以删除包含在xarray数据集中的DataArrays值，以便在下面的示例代码中将xr_dataset转换为xr_flat？

import xarray as xr
import numpy as np
n = np.nan

a = np.array([[n,n,1],
              [1,2,2],
              [2,n,n]], dtype='float32')
b = np.random.rand(3,3)
xr_dataset = xr.Dataset({'a': xr.DataArray(a, dims=['x', 'y']),
                         'b': xr.DataArray(b, dims=['x', 'y'])})

a_flat = a[np.isfinite(a)]
b_flat = b[np.isfinite(a)]

xr_flat = xr.Dataset({'a': xr.DataArray(a_flat),
                      'b': xr.DataArray(b_flat)})

Answer 1

您可以使用xarray的stack和where方法来完成这一任务。为了与您的示例保持一致，我还删除了x/y坐标，但这不是严格要求的。关键部分是：

1. 使用堆栈将x/y简化为dim_0
2. 使用where和drop=True掩蔽并只选择有限元素

举个例子，从你停止的地方开始.

In [2]: ds_stack = xr_dataset.stack(dim_0=('x', 'y'))

In [3]: ds_stack = ds_stack.reset_index('dim_0').drop(['x', 'y'])

In [4]: ds_stack.where(np.isfinite(ds_stack['a']), drop=True)
Out[4]:
<xarray.Dataset>
Dimensions:  (dim_0: 5)
Dimensions without coordinates: dim_0
Data variables:
    a        (dim_0) float32 1.0 1.0 2.0 2.0 2.0
    b        (dim_0) float64 0.8642 0.05446 0.3728 0.7797 0.9501

In [5]: xr_flat
Out[5]:
<xarray.Dataset>
Dimensions:  (dim_0: 5)
Dimensions without coordinates: dim_0
Data variables:
    a        (dim_0) float32 1.0 1.0 2.0 2.0 2.0
    b        (dim_0) float64 0.8642 0.05446 0.3728 0.7797 0.9501