1. 程式人生 > 實用技巧 >numpy花式索引與ix_()

numpy花式索引與ix_()

花式索引(Fancy indexing)是Numpy的一個術語,指的是利用整數陣列進行索引。(不僅是1維,也可以是多維)

用法與例子如下:

建立 arr 陣列

>>> arr1 = np.empty((8,4))# 建立一個8行4列的二維陣列

>>> for i in range(8):# 每一行賦值為0~7
arr1[i] = i

>>> arr1

array([[ 0., 0., 0., 0.],
[ 1., 1., 1., 1.],
[ 2., 2., 2., 2.],
[ 3., 3., 3., 3.],

[ 4., 4., 4., 4.],
[ 5., 5., 5., 5.],
[ 6., 6., 6., 6.],
[ 7., 7., 7., 7.]])


1、用1維陣列進行索引

>>> arr1[[4,3,0,6]]
# 選取第4行、第3行、第0行、第6行

array([[ 4., 4., 4., 4.],
[ 3., 3., 3., 3.],
[ 0., 0., 0., 0.],
[ 6., 6., 6., 6.]])
2、用有負數的1維陣列進行索引,就是從末尾開始選取行

>>> arr1[[-3,-5,-7]]
# 選取倒數第3行,倒數第5行,倒數第7行

array([[ 5., 5., 5., 5.],
[ 3., 3., 3., 3.],
[ 1., 1., 1., 1.]])
在這裡必須注意!

順序選取是從0開始數的,a[0]代表第一個;而逆序選取是從1開始數的,a[-1]是倒數第一個

新建一個數組 arr2

>>> arr2 = np.arange(32).reshape((8,4))

>>> arr2

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, 26, 27],
[28, 29, 30, 31]])


3、按座標選取每一個數

>>> arr2[[1,5,7,2],[0,3,1,2]]
# 意思就是,取座標所對應的數(1,0)——4,(5,3)——23,(7,1)——29,(2,2)——10,然後返回一個數組

array([ 4, 23, 29, 10])
4、希望先按我們要求選取行,再按順序將列排序,獲得一個矩形

>>> arr2[[1,5,7,2]][:,[0,3,1,2]]

array([[ 4, 7, 5, 6],
[20, 23, 21, 22],
[28, 31, 29, 30],
[ 8, 11, 9, 10]])
先按先選取第1、5、2、7行,每一行再按第0個、第3個、第1個、第2個排序

5、np.ix_函式,能把兩個一維陣列 轉換為 一個用於選取方形區域的索引器

實際意思就是,直接往np.ix_()裡扔進兩個一維陣列[1,5,7,2],[0,3,1,2],就能先按我們要求選取行,再按順序將列排序,跟上面得到的結果一樣,而不用寫“[ : , [0,3,1,2] ]”

原理:np.ix_函式就是輸入兩個陣列,產生笛卡爾積的對映關係

>>> arr2[np.ix_([1,5,7,2],[0,3,1,2])]

array([[ 4, 7, 5, 6],
[20, 23, 21, 22],
[28, 31, 29, 30],
[ 8, 11, 9, 10]])
例如就這個例子,np.ix_函式,將陣列[1,5,7,2]和陣列[0,3,1,2]產生笛卡爾積,就是得到(1,0),(1,3),(1,1),(1,2);(5,0),(5,3),(5,1),(5,2);(7,0),(7,3),(7,1),(7,2);(2,0),(2,3),(2,1),(2,2);

就是按照座標(1,0),(1,3),(1,1),(1,2)取得 arr2 所對應的元素4,7,5,6

(5,0),(5,3),(5,1),(5,2)取得 arr2 所對應的元素20,23,21,22

如此類推。

原文:https://blog.csdn.net/weixin_40001181/article/details/79775792

下面是官方解釋:

numpy.ix_

numpy.ix_(*args)[source]

Construct an open mesh from multiple sequences.

This function takes N 1-D sequences and returns N outputs with N dimensions each, such that the shape is 1 in all but one dimension and the dimension with the non-unit shape value cycles through all N dimensions.

Usingix_one can quickly construct index arrays that will index the cross product.a[np.ix_([1,3],[2,5])]returns the array[[a[1,2]a[1,5]],[a[3,2]a[3,5]]].

Parameters
args1-D sequences

Each sequence should be of integer or boolean type. Boolean sequences will be interpreted as boolean masks for the corresponding dimension (equivalent to passing innp.nonzero(boolean_sequence)).

Returns
outtuple of ndarrays

N arrays with N dimensions each, with N the number of input sequences. Together these arrays form an open mesh.

See also

ogrid,mgrid,meshgrid

Examples

>>>
>>> a = np.arange(10).reshape(2, 5)
>>> a
array([[0, 1, 2, 3, 4],
       [5, 6, 7, 8, 9]])
>>> ixgrid = np.ix_([0, 1], [2, 4])
>>> ixgrid
(array([[0],
       [1]]), array([[2, 4]]))
>>> ixgrid[0].shape, ixgrid[1].shape
((2, 1), (1, 2))
>>> a[ixgrid]
array([[2, 4],
       [7, 9]])
>>>
>>> ixgrid = np.ix_([True, True], [2, 4])
>>> a[ixgrid]
array([[2, 4],
       [7, 9]])
>>> ixgrid = np.ix_([True, True], [False, False, True, False, True])
>>> a[ixgrid]
array([[2, 4],
       [7, 9]])