NumPy迭代数组的实现

>>>a = np.arange(12).reshape(3, 4)
>>>for x in np.nditer(a):
            print(x, end=' ')
0 1 2 3 4 5 6 7 8 9 10 11 

# 以上实例不是使用标准 C 或者 Fortran 顺序，选择的顺序是和数组内存布局一致的，
# 这样做是为了提升访问的效率，默认是行序优先（row-major order，或者说是 C-order）。
# 这反映了默认情况下只需访问每个元素，而无需考虑其特定顺序。
# 我们可以通过迭代上述数组的转置来看到这一点，
# 并与以 C 顺序访问数组转置的 copy 方式做对比，如下实例：
>>>for x in np.nditer(a.T):
            print(x, end=' ')
0 1 2 3 4 5 6 7 8 9 10 11 

>>>for x in np.nditer(a.T.copy(order='C')):
            print(x, end=' ')
0 4 8 1 5 9 2 6 10 3 7 11 

>>>a = np.arange(12).reshape(3, 4)
>>>for x in np.nditer(a, order='C'):
        print(x, end=' ')
0 1 2 3 4 5 6 7 8 9 10 11 

>>>a = np.arange(12).reshape(3, 4)
>>>for x in np.nditer(a, order='F'):
        print(x, end=' ')
0 4 8 1 5 9 2 6 10 3 7 11 

>>>a = np.arange(12).reshape(3, 4)
>>>for x in np.nditer(a, order='K'):
        print(x, end=' ')
0 1 2 3 4 5 6 7 8 9 10 11 

>>>a = np.arange(12).reshape(3, 4)
>>>for x in np.nditer(a, order='A'):
        print(x, end=' ')
0 1 2 3 4 5 6 7 8 9 10 11 

>>>a = np.arange(12).reshape(3, 4)
>>>print(a)
>>>with np.nditer(a, op_flags=['readwrite']) as it:
        for x in it:
            x += 10
>>>print(a)
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
[[10 11 12 13]
 [14 15 16 17]
 [18 19 20 21]]

# 默认情况下保持本机的内存顺序，迭代器提供单一的一维数组
# 'external_loop' 给出的值是具有多个值的一维数组，而不是零维数组
>>>a = np.arange(12).reshape(3, 4)
>>>print(a)
>>>for x in np.nditer(a, flags=['external_loop']):
        print(x, end=' ')
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
[ 0  1  2  3  4  5  6  7  8  9 10 11], 

# 设定 'F' 顺序
>>>a = np.arange(12).reshape(3, 4)
>>>print(a)
>>>for x in np.nditer(a, flags=['external_loop'], order='F'):
        print(x, end=' ')
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
[0 4 8], [1 5 9], [ 2  6 10], [ 3  7 11], 

# 'c_index' 可以通过 it.index 跟踪 'C‘ 顺序的索引
>>>a = np.arange(12).reshape(3, 4)
>>>print(a)
>>>it = np.nditer(a, flags=['c_index'])
>>>for x in it:
            print("{}: ({})".format(x, it.index))
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
0: (0)
1: (1)
2: (2)
3: (3)
4: (4)
5: (5)
6: (6)
7: (7)
8: (8)
9: (9)
10: (10)
11: (11)

# 'f_index' 可以通过 it.index 跟踪 'F‘ 顺序的索引
>>>a = np.arange(12).reshape(3, 4)
>>>print(a)
>>>it = np.nditer(a, flags=['c_index'])
>>>for x in it:
            print("{}: ({})".format(x, it.index))
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
0: (0)
1: (3)
2: (6)
3: (9)
4: (1)
5: (4)
6: (7)
7: (10)
8: (2)
9: (5)
10: (8)
11: (11)

# 'multi_index' 可以通过 it.multi_index 跟踪数组索引
>>>a = np.arange(12).reshape(3, 4)
>>>print(a)
>>>it = np.nditer(a, flags=['multi_index'])
>>>for x in it:
        print("{}: {}".format(x, it.multi_index))
[[ 0  1  2  3]
 [ 4  5  6  7]
 [ 8  9 10 11]]
0: (0, 0)
1: (0, 1)
2: (0, 2)
3: (0, 3)
4: (1, 0)
5: (1, 1)
6: (1, 2)
7: (1, 3)
8: (2, 0)
9: (2, 1)
10: (2, 2)
11: (2, 3)

# 临时副本
>>>a = np.arange(12).reshape(3, 4)
>>>print(a.dtype)
>>>it = np.nditer(a, op_flags=['readonly', 'copy'],op_dtypes=[np.float64])
>>>for x in it:
        print("{}".format(x), end=', ')
int32
0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0,

# 缓冲模式

>>>a = np.arange(12).reshape(3, 4)
>>>print(a.dtype)
>>>it = np.nditer(a, flags=['buffered'],op_dtypes=[np.float64])
>>>for x in it:
        print("{}".format(x), end=', ')
int32
0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 

>>> a = np.arange(3)
>>> b = np.arange(6).reshape(2,3)
>>> for x, y in np.nditer([a,b]):
        print("%d:%d" % (x,y), end=' ')
0:0 1:1 2:2 0:3 1:4 2:5