幻想索引与Numpy中的视图第二部分 [英] Fanccy Indexing vs View in Numpy part II
问题描述
在回答这个方程式时:解释了不同的习语会产生不同的结果.
In an answer to this equation: is is explained that different idioms will produce different results.
使用习惯式索引来选择值并将这些值设置为同一行中的新值，这意味着原始对象中的值将被更改.
Using the idiom where fancy indexing is to chose the values and said values are set to a new value in the same line means that the values in the original object will be changed in place.
但是下面的最后一个示例:
However the final example below:
https://scipy-cookbook.readthedocs.io/items/ViewsVsCopies.html
最后的练习"
该示例似乎使用了相同的习惯用法:
The example appears to use the same idiom:
a [x，:] [:, y] = 100
a[x, :][:, y] = 100
，但根据x是切片还是奇特的索引，它仍然会产生不同的结果(见下文):
but it still produces a different result depending on whether x is a slice or a fancy index (see below):
a = np.arange(12).reshape(3,4)
ifancy = [0,2]
islice = slice(0,3,2)
a[islice, :][:, ifancy] = 100
a
#array([[100, 1, 100, 3],
# [ 4, 5, 6, 7],
# [100, 9, 100, 11]])
a = np.arange(12).reshape(3,4)
ifancy = [0,2]
islice = slice(0,3,2)
a[ifancy, :][:, islice] = 100 # note that ifancy and islice are interchanged here
>>> a
array([[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11]])
我的直觉是，如果第一组花式索引是一个切片，则它将对象视为视图，因此原始对象中的值会更改.
My intuition is that if the first set of fancy indexes is a slice it treats the object like a view and therefore the values in the orignal object are changed.
在第二种情况下，第一组花式索引本身就是花式索引，因此它将对象视为创建原始对象副本的花式索引.这意味着复制对象的值更改时，原始对象不会更改.
Whereas in the second case the first set of fancy indexes is itself a fancy index so it treats the object as a fancy index creating a copy of the original object. This then means that the original object is not changed when the values of the copy object are changed.
我的直觉正确吗?
该示例提示人们应该考虑 getitem 和 setitem 的缺点，有人可以用这种方式向我正确解释吗?
The example hints that one should think of the sqeuence of getitem and setitem can someone explain it to my properly in theis way?
推荐答案
Python分别评估每组[]. a[x, :][:, y] = 100是2个操作.
Python evaluates each set of [] separately. a[x, :][:, y] = 100 is 2 operations.
temp = a[x,:] # getitem step
temp[:,y] = 100 # setitem step
第二行是否最终修改a取决于temp是视图还是副本.
Whether the 2nd line ends up modifying a depends on whether temp is a view or copy.
请记住，numpy是Python的附加组件.它不会修改基本的Python语法或解释.
Remember, numpy is an addon to Python. It does not modify basic Python syntax or interpretation.
这篇关于幻想索引与Numpy中的视图第二部分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！
