大量对象属性 [英] Numpy array of object attributes
问题描述
我有一个多维的对象数组,类似:
I have a multi-dimensional array of objects, something like:
a = np.array([obj1,obj2,obj3])
对象是具有多个属性的类的实例.假设其中之一是高度,其中之一是长度.要获取相应的长度和高度的多维数组,我要做:
The objects are instances of a class which has several attributes. Let's say one of them is heights and one of them is lengths. To get the corresponding multi-dimensional array of lengths and heights I do:
lengths = np.array([obj1.length,obj2.length,obj3.length])
heights = np.array([obj1.height,obj2.height,obj3.height])
这开始使我的代码混乱很多.有更有效的方法吗?例如,我有类似
This is starting to clutter up my code quite a lot. Is there a more efficient way of doing this? For instance, I had something like
heights = a.height
记住
,但显然不起作用,因为a是我的对象而不是我的对象的数组.但是,有没有类似的事情我可以做,既高效又pythonic?我尝试过类似的
in mind but obviously it doesn't work because a is an array of my objects and not my object. But is there something similar I can do that is efficient and pythonic? I tried something like
for x in np.nditer(a,flags=['refs_ok']):
print x.length
看看会发生什么,但是它不起作用,因为nditer以某种方式返回了一个元组.
to see what would happen but it doesn't work because nditer returns a tuple somehow.
有什么想法吗?
推荐答案
您可以向量化函数:
>>> import numpy
>>>
>>> class Obj(object):
... def __init__(self, x, y):
... self.x = x
... self.y = y
...
>>> arr = numpy.array([Obj(1, 2), Obj(3, 4), Obj(5, 6)])
>>>
>>> vectorized_x = numpy.vectorize(lambda obj: obj.x)
>>>
>>> vectorized_x(arr)
array([1, 3, 5])
尽管我不确定您是否真的应该首先存储一个NumPy数组的Python对象.向量化没有比Python循环更有效.存储(n + 1)-D数组会更有效,因为我们可以简单地通过切片来轻松提取内容,这是本机操作,例如
Although I'm not sure if you should really store a NumPy array of Python objects in the first place. Vectorize is no more efficient than a Python loop. It would be more efficient to store an (n+1)-D array, as we could easily extract the contents simply by slicing which is a native operation, e.g.
>>> a = numpy.array([[(1, 2), (3, 4), (5, 6)], [(7, 8), (9, 10), (11, 12)], [(-13, -14), (-15, -16), (-17, -18)]])
>>> a[:,:,0]
array([[ 1, 3, 5],
[ 7, 9, 11],
[-13, -15, -17]])
>>> a[:,:,1]
array([[ 2, 4, 6],
[ 8, 10, 12],
[-14, -16, -18]])
这篇关于大量对象属性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!