Python Numpy赋值行为

b

是

array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
       -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
       -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
        0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
       -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
       -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])

a = b
a[a <= 0] = 0
a[a > 0] = 1

数组（[-0.06106568，-0.10843541，-0.0694688,0.02464023，-0.03686665，
-0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
-0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
-0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
-0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])
a=b
a[a0]=1

---

使用上面的代码，我想替换a中的元素，但b也会更改。。。您能解释一下错误在哪里吗？

通过使用表达式

a=b

您实际上将引用复制到了

b

。如果要复制其值，则应遍历所有

b

项，并将其值复制到

a

在numpy中，您应该使用

copy

功能

>>> import numpy
>>> b = numpy.array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
...        -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
...        -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
...         0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
...        -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
...        -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])
>>> a = numpy.copy(b)
>>> a
array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
       -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
       -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
        0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
       -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
       -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])
>>> a[a<= 0] = 0
>>> a[a> 0] = 1
>>> a
array([ 0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,
        0.,  1.,  1.,  0.,  0.,  0.,  0.,  0.,  1.,  0.,  1.,  0.,  0.,
        0.,  0.,  0.,  0.])
>>> b
array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
       -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
       -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
        0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
       -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
       -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])

导入numpy >>>b=numpy.数组（[-0.06106568，-0.10843541，-0.0694688,0.02464023，-0.03686665， ... -0.0582096 , -0.13476669, -0.08505708, 0.00391955, -0.12300518, ... -0.01183732, -0.05374973, -0.12300518, -0.05312849, 0.01963862, ... 0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149, ... -0.01777489, 0.01183732, -0.11575136, 0.04278603, -0.0694688 , ... -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411]) >>>a=numpy.copy（b） >>>a 阵列（[-0.06106568，-0.10843541，-0.0694688，0.02464023，-0.03686665， -0.0582096 , -0.13476669, -0.08505708, 0.00391955, -0.12300518, -0.01183732, -0.05374973, -0.12300518, -0.05312849, 0.01963862, 0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149, -0.01777489, 0.01183732, -0.11575136, 0.04278603, -0.0694688 , -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411]) >>>a[a>>a[a>0]=1 >>>a 数组（[0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,。， 0., 1., 1., 0., 0., 0., 0., 0., 1., 0., 1., 0., 0., 0., 0., 0., 0.]) >>>b 阵列（[-0.06106568，-0.10843541，-0.0694688，0.02464023，-0.03686665， -0.0582096 , -0.13476669, -0.08505708, 0.00391955, -0.12300518, -0.01183732, -0.05374973, -0.12300518, -0.05312849, 0.01963862, 0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149, -0.01777489, 0.01183732, -0.11575136, 0.04278603, -0.0694688 , -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411]) 有一些本地的方法可以做到这一点，但是如果你使用数学，强烈建议只使用numpy

更新

---

我认为没有任何非numpy方式会导致与您的案例完全兼容。

通过使用表达式

a=b

实际上是将引用复制到

b

。如果您想复制它的值，您应该遍历

b

的所有项目，并将它们的值复制到

a

在numpy中，您应该使用

copy

功能

>>> import numpy
>>> b = numpy.array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
...        -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
...        -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
...         0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
...        -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
...        -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])
>>> a = numpy.copy(b)
>>> a
array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
       -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
       -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
        0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
       -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
       -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])
>>> a[a<= 0] = 0
>>> a[a> 0] = 1
>>> a
array([ 0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,
        0.,  1.,  1.,  0.,  0.,  0.,  0.,  0.,  1.,  0.,  1.,  0.,  0.,
        0.,  0.,  0.,  0.])
>>> b
array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
       -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
       -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
        0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
       -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
       -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])

导入numpy >>>b=numpy.数组（[-0.06106568，-0.10843541，-0.0694688,0.02464023，-0.03686665， ... -0.0582096 , -0.13476669, -0.08505708, 0.00391955, -0.12300518, ... -0.01183732, -0.05374973, -0.12300518, -0.05312849, 0.01963862, ... 0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149, ... -0.01777489, 0.01183732, -0.11575136, 0.04278603, -0.0694688 , ... -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411]) >>>a=numpy.copy（b） >>>a 阵列（[-0.06106568，-0.10843541，-0.0694688，0.02464023，-0.03686665， -0.0582096 , -0.13476669, -0.08505708, 0.00391955, -0.12300518, -0.01183732, -0.05374973, -0.12300518, -0.05312849, 0.01963862, 0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149, -0.01777489, 0.01183732, -0.11575136, 0.04278603, -0.0694688 , -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411]) >>>a[a>>a[a>0]=1 >>>a 数组（[0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,。， 0., 1., 1., 0., 0., 0., 0., 0., 1., 0., 1., 0., 0., 0., 0., 0., 0.]) >>>b 阵列（[-0.06106568，-0.10843541，-0.0694688，0.02464023，-0.03686665， -0.0582096 , -0.13476669, -0.08505708, 0.00391955, -0.12300518, -0.01183732, -0.05374973, -0.12300518, -0.05312849, 0.01963862, 0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149, -0.01777489, 0.01183732, -0.11575136, 0.04278603, -0.0694688 , -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411]) 有一些本地的方法可以做到这一点，但是如果你使用数学，强烈建议只使用numpy

更新

---

我不认为有任何非numpy方式会导致与您的案例完全兼容。

当您执行

a=b

时，您对

b

进行了命名引用，因此

a

和

b

是同一对象上的视图，如果您想要副本，请使用

np

In [35]:
a = np.copy(b)
a[a<= 0] = 0
a[a> 0] = 1
a

Out[35]:
array([ 0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,  1.,  0.,  0.,  0.,  0.,
        0.,  1.,  1.,  0.,  0.,  0.,  0.,  0.,  1.,  0.,  1.,  0.,  0.,
        0.,  0.,  0.,  0.])

In [36]:
b

Out[36]:
array([-0.06106568, -0.10843541, -0.0694688 ,  0.02464023, -0.03686665,
       -0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
       -0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
        0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
       -0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
       -0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])

[35]中的

a=np.副本（b）
a[a0]=1
A.
出[35]：
数组（[0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,。，
0.,  1.,  1.,  0.,  0.,  0.,  0.,  0.,  1.,  0.,  1.,  0.,  0.,
0.,  0.,  0.,  0.])
在[36]中：
B
出[36]：
阵列（[-0.06106568，-0.10843541，-0.0694688，0.02464023，-0.03686665，
-0.0582096 , -0.13476669, -0.08505708,  0.00391955, -0.12300518,
-0.01183732, -0.05374973, -0.12300518, -0.05312849,  0.01963862,
0.00155719, -0.10843541, -0.08490177, -0.08505708, -0.02026149,
-0.01777489,  0.01183732, -0.11575136,  0.04278603, -0.0694688 ,
-0.06106568, -0.08755022, -0.01660802, -0.06087603, -0.06582411])
当您执行
a=b
时，您对
b
进行了命名引用，因此
a
和
b
是同一对象上的视图，如果您需要副本，请使用
np.c
import copy
a = copy.copy(b)

import numpy
a = np.copy(b)