给定通用维度的开始和结束索引，对NumPy数组进行切片 [英] Slicing NumPy array given start and end indices for generic dimensions

问题描述

给出一个形状为(N_1...N_k)的numpy数组x，其中k是任意的，还有2个数组:

Given a numpy array x of shape (N_1...N_k) where k is arbitrary, and 2 arrays :

start_indices=[a_1,...,a_k], end_indices=[b_1,...b_k], where `0<=a_i<b_i<=N_i`.

我想对x进行如下切片:x[a_1:b_1,...,a_k:b_k].

I want to slice x as follows: x[a_1:b_1,...,a_k:b_k].

让我们说:

x is of shape `(1000, 1000, 1000)`
start_indices=[450,0,400]
end_indices=[550,1000,600].

我希望输出等于x[450:550,0:1000,400:600].

例如，我尝试定义:

slice_arrays = (np.arange(start_indices[i], end_indices[i]) for i in range(k))

并使用

x[slice_arrays]

但是没有用.

推荐答案

您可以使用slice表示法创建可用于索引的索引元组-

You can use slice notation to create an indexing tuple that could be used for the indexing -

indexer = tuple([slice(i,j) for (i,j) in zip(start_indices,end_indices)])
out = x[indexer]

或者，用速记 np.s_ -

Alternatively, with shorthand np.s_ -

indexer = tuple([np.s_[i:j] for (i,j) in zip(start_indices,end_indices)])

或与map组合在一起-

indexer = tuple(map(slice,start_indices,end_indices))

这篇关于给定通用维度的开始和结束索引，对NumPy数组进行切片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！