[英]Pythonic way of iterating over 3D array
I have a 3D array in Python and I need to iterate over all the cubes in the array. 我在Python中有一个3D数组,我需要迭代数组中的所有多维数据集。 That is, for all
(x,y,z)
in the array's dimensions I need to access the cube: 也就是说,对于数组维度中的所有
(x,y,z)
,我需要访问多维数据集:
array[(x + 0, y + 0, z + 0)]
array[(x + 1, y + 0, z + 0)]
array[(x + 0, y + 1, z + 0)]
array[(x + 1, y + 1, z + 0)]
array[(x + 0, y + 0, z + 1)]
array[(x + 1, y + 0, z + 1)]
array[(x + 0, y + 1, z + 1)]
array[(x + 1, y + 1, z + 1)]
The array is a Numpy array, though that's not really necessary. 该数组是一个Numpy数组,虽然这不是必需的。 I just found it very easy to read the data in with a one-liner using
numpy.fromfile()
. 我刚刚发现使用
numpy.fromfile()
使用numpy.fromfile()
读取数据非常容易。
Is there any more Pythonic way to iterate over these than the following? 是否有更多的Pythonic方法来迭代这些而不是以下? That simply looks like C using Python syntax.
这简直就像使用Python语法的C一样。
for x in range(x_dimension):
for y in range(y_dimension):
for z in range(z_dimension):
work_with_cube(array[(x + 0, y + 0, z + 0)],
array[(x + 1, y + 0, z + 0)],
array[(x + 0, y + 1, z + 0)],
array[(x + 1, y + 1, z + 0)],
array[(x + 0, y + 0, z + 1)],
array[(x + 1, y + 0, z + 1)],
array[(x + 0, y + 1, z + 1)],
array[(x + 1, y + 1, z + 1)])
Have a look at itertools , especially itertools.product . 看一下itertools ,特别是itertools.product 。 You can compress the three loops into one with
您可以将三个循环压缩为一个
import itertools
for x, y, z in itertools.product(*map(xrange, (x_dim, y_dim, z_dim)):
...
You can also create the cube this way: 您也可以这样创建多维数据集:
cube = numpy.array(list(itertools.product((0,1), (0,1), (0,1))))
print cube
array([[0, 0, 0],
[0, 0, 1],
[0, 1, 0],
[0, 1, 1],
[1, 0, 0],
[1, 0, 1],
[1, 1, 0],
[1, 1, 1]])
and add the offsets by a simple addition 并通过简单的添加添加偏移量
print cube + (10,100,1000)
array([[ 10, 100, 1000],
[ 10, 100, 1001],
[ 10, 101, 1000],
[ 10, 101, 1001],
[ 11, 100, 1000],
[ 11, 100, 1001],
[ 11, 101, 1000],
[ 11, 101, 1001]])
which would to translate to cube + (x,y,z)
in your case. 在你的情况下,它将转换为
cube + (x,y,z)
。 The very compact version of your code would be 你的代码非常紧凑的版本
import itertools, numpy
cube = numpy.array(list(itertools.product((0,1), (0,1), (0,1))))
x_dim = y_dim = z_dim = 10
for offset in itertools.product(*map(xrange, (x_dim, y_dim, z_dim))):
work_with_cube(cube+offset)
Edit : itertools.product
makes the product over the different arguments, ie itertools.product(a,b,c)
, so I have to pass map(xrange, ...)
with as *map(...)
编辑 :
itertools.product
使产品通过不同的参数,即itertools.product(a,b,c)
,所以我必须使用as *map(...)
传递map(xrange, ...)
*map(...)
import itertools
for x, y, z in itertools.product(xrange(x_size),
xrange(y_size),
xrange(z_size)):
work_with_cube(array[x, y, z])
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.