由于我们将相同的值（1D数组的每个元素）添加到2D数组中，这相当于将每个轴的总和加上1D数组**乘以2D数组第二个轴的维度。我们还必须向1D数组添加一个新的轴，否则形状不匹配。

import numpy as np

arr2d = np.random.randint(0, 10, size=(10000, 50))
arr1d = np.random.randint(0, 10, size=2000)
arr_sum = np.array([(arr2d+element).sum(axis=1) for element in arr1d]).T

out = (arr2d.sum(axis=1) + (arr1d * arr2d.shape[1])[:, np.newaxis]).T

print(out.shape == arr_sum.shape) # True
print(np.allclose(out, arr_sum))  # True

和

In [22]: %timeit np.array([(arr2d+element).sum(axis=1) for element in arr1d]).T
1.86 s ± 20.6 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)

In [23]: %timeit (arr2d.sum(axis=1) + (arr1d * arr2d.shape[1])[:, np.newaxis]).T
25.8 ms ± 226 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)

如果你想编写一个高性能的Numba实现，你必须编写简单的循环，或者在某些情况下使用BLAS函数。

实施例1

import numba as nb
import numpy as np

@nb.njit(fastmath=True,parallel=True,cache=True)
def cust_sum(TwoDArray,OneDArray):
    #allocate Output
    res=np.empty((TwoDArray.shape[0],OneDArray.shape[0]),dtype=TwoDArray.dtype)

    for i in nb.prange(TwoDArray.shape[0]):
        for j in range(OneDArray.shape[0]):
            acc=0
            for k in range(TwoDArray.shape[1]):
                acc+=TwoDArray[i,k]+OneDArray[j]
            res[i,j]=acc
    return res

%timeit sum_2=cust_sum(TwoDArray,OneDArray)
#12.8 ms ± 498 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)

编辑：示例2

与一些想法从弥敦道炉回答这可以修改为

@nb.njit(fastmath=True,parallel=True,cache=True)
def cust_sum_2(TwoDArray,OneDArray):
    #allocate Output
    res=np.empty((TwoDArray.shape[0],OneDArray.shape[0]),dtype=TwoDArray.dtype)

    for i in nb.prange(TwoDArray.shape[0]):
        acc=0
        for k in range(TwoDArray.shape[1]):
            acc+=TwoDArray[i,k]
        for j in range(OneDArray.shape[0]):
            res[i,j]=acc+OneDArray[j]*TwoDArray.shape[1]
    return res

%timeit sum_2=cust_sum_2(TwoDArray,OneDArray)
#7.25 ms ± 815 µs per loop (mean ± std. dev. of 7 runs, 1 loop each)

print(np.allclose(cust_sum_2(TwoDArray,OneDArray),Sum))
#True

Change - Sum = np.array（[（TwoDArray+element）.sum（axis=1）for element in OneDArray]）.T

import numpy as np

TwoDArray = np.random.randint(0, 10, size=(10000, 50))
OneDArray = np.random.randint(0, 10, size=(2000))

Sum = TwoDArray.dot(np.ones((50, 2000))) + OneDArray

print(Sum.shape)