使用Numpy的类似MATLAB的数组索引 [英] MATLAB-like array indexing with Numpy

问题描述

在MATLAB和Numpy中，都可以通过数组对数组进行索引.但是，行为是不同的.让我通过一个例子对此进行解释.

In both MATLAB and Numpy, arrays can be indexed by arrays. However, the behavior is different. Let me explain this by an example.

MATLAB:

>> A = rand(5,5)

A =

    0.1622    0.6020    0.4505    0.8258    0.1067
    0.7943    0.2630    0.0838    0.5383    0.9619
    0.3112    0.6541    0.2290    0.9961    0.0046
    0.5285    0.6892    0.9133    0.0782    0.7749
    0.1656    0.7482    0.1524    0.4427    0.8173

>> A([1,3,5],[1,3,5])

ans =

    0.1622    0.4505    0.1067
    0.3112    0.2290    0.0046
    0.1656    0.1524    0.8173

脾气暴躁:

In [2]: A = arange(25).reshape((5,5))

In [3]: A
Out[3]: 
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]])

In [6]: A[[0,2,4], [0,2,4]]
Out[6]: array([ 0, 12, 24])

换句话说:MATLAB选择行和列，Numpy压缩"两个索引数组，并使用元组指向条目.

In words: MATLAB selects rows and columns, Numpy "zips" the two index arrays and uses the tuples to point to entries.

如何通过Numpy获得MATLAB行为?

How can I get the MATLAB behavior with Numpy?

推荐答案

您可以执行以下操作:

A[[0,2,4],:][:,[0,2,4]]

这将提供您想要的类似MATLAB的结果.

which will give the MATLAB-like result you want.

值得一提的是，如果您使用切片来建立索引，则会产生类似MATLAB的结果，而不会受到任何黑客攻击，这是相当不一致的:

It's worth being aware that, rather inconsistently, if you use slices for indexing then you get MATLAB-like results without any such hackery:

>>> A[1:3,1:3]
array([[ 6, 7],
       [11,12]])

在numpy中，与MATLAB不同，1:3不仅是[1,2]或任何类型的缩写. (在这一点上，我不得不提一句您肯定已经知道的东西，即Python的1:3有点像[1,2]，而MATLAB的有点像[1,2,3]:右侧端点包含在MATLAB中而在Python中不包含在内. )

In numpy, unlike MATLAB, 1:3 is not just an abbreviation for [1,2] or anything of the kind. (At which point I feel obliged to mention something you surely know already, namely that Python's 1:3 is kinda like [1,2] whereas MATLAB's is kinda like [1,2,3]: the right-hand endpoint is included in MATLAB and excluded in Python.)

这篇关于使用Numpy的类似MATLAB的数组索引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！