Python Numpy Flat函数 [英] Python Numpy Flat Function

问题描述

numpy中的flat函数如何工作?另外，这些索引如何工作?请解释以下所有代码.

How does the flat function in numpy work? Also, how do these indexes work? Please explain all the following code.

res = zeros((n, n), v.dtype)  
res[:n-k].flat[i::n+1] = v  

推荐答案

flat的简单用法:

In [410]: res = np.zeros((2,3), dtype=int)
In [411]: res
Out[411]: 
array([[0, 0, 0],
       [0, 0, 0]])

In [413]: res.flat[::2]=1
In [414]: res
Out[414]: 
array([[1, 0, 1],
       [0, 1, 0]])
In [415]: res.ravel()
Out[415]: array([1, 0, 1, 0, 1, 0])

flat是flatten和ravel的变体.在这里，我将其分配给扁平化数组的每个其他元素1.您可以在ravel表达式中看到这种情况.在同一数组的2d视图中，它不太明显.

flat is a variant on flatten and ravel. Here I use it to assign 1 to every other element of the flattened array. You can see that happening in the ravel expression. It's a bit less obvious in the 2d view of the same array.

在res[:n-k].flat[i::n+1] = v中，第一个[:n-k]选择res的某些行.在我的示例中，flat[]起作用，将v中的值分配给展平数组中的任何n+1元素.

In res[:n-k].flat[i::n+1] = v, the first [:n-k] selects some rows of res. The flat[] acts in my example, assigning values from v to ever n+1 element in the flattened array.

再次通过一个小例子进行测试:

Again testing with a small example:

In [417]: res = np.zeros((5,5), dtype=int)
In [418]: res[:3]
Out[418]: 
array([[0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0]])
In [419]: res[:3].flat[2::6]
Out[419]: array([0, 0, 0])
In [420]: res[:3].flat[2::6]=[1,2,3]
In [421]: res
Out[421]: 
array([[0, 0, 1, 0, 0],
       [0, 0, 0, 2, 0],
       [0, 0, 0, 0, 3],
       [0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0]])

使用[i::n+1]索引最终会在对角线上设置值.

The use of [i::n+1] indexing ends up setting values on an diagonal.

或

In [422]: res = np.zeros((5,5), dtype=int)
In [424]: res.flat[0::6]
Out[424]: array([0, 0, 0, 0, 0])
In [425]: res.flat[0::6]=np.arange(5)
In [426]: res
Out[426]: 
array([[0, 0, 0, 0, 0],
       [0, 1, 0, 0, 0],
       [0, 0, 2, 0, 0],
       [0, 0, 0, 3, 0],
       [0, 0, 0, 0, 4]])

这篇关于Python Numpy Flat函数的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！