Python ValueError: 'axis' 条目超出范围 // numpy
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/19572178/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
ValueError: 'axis' entry is out of bounds // numpy
提问by Michael
Here is my 2D numpy array:
这是我的二维 numpy 数组：
returns = np.array([
[ -4.78878057e-03 9.79090927e-03 -2.06883581e-03 -1.25786164e-02]
[ 5.79128440e-03 -2.85791008e-03 1.69555388e-03 -5.63798220e-02]
[ 5.73427375e-05 2.45043133e-02 8.55025651e-03 -4.53257790e-02]
[ 6.75441635e-03 8.70168484e-03 1.07547532e-02 -1.36919315e-01]
[ 6.68332655e-03 6.76498174e-03 3.08225775e-03 0.00000000e+00]])
And when I'm trying to calculate the STD for each column:
当我尝试计算每列的 STD 时：
print np.std(returns, axis=1)
I'm getting the following error:
我收到以下错误：
ValueError: 'axis' entry is out of bounds
How can I fix that?
我该如何解决？
回答by Moe
Michael,
迈克尔，
I am not sure why you have the issue. It might be because of the way you are defining the 2D array (that's my only guess). I used the following code and it works properly:
我不确定你为什么有这个问题。这可能是因为您定义二维数组的方式（这是我唯一的猜测）。我使用了以下代码并且它工作正常：
import numpy as np
arr = np.array([ \
[ -4.78878057e-03, 9.79090927e-03, -2.06883581e-03, -1.25786164e-02], \
[ 5.79128440e-03, -2.85791008e-03, 1.69555388e-03, -5.63798220e-02], \
[ 5.73427375e-05, 2.45043133e-02, 8.55025651e-03, -4.53257790e-02], \
[ 6.75441635e-03, 8.70168484e-03, 1.07547532e-02, -1.36919315e-01], \
[ 6.68332655e-03, 6.76498174e-03, 3.08225775e-03, 0.00000000e+00] \
])
std = np.std(arr,axis=1)
print std
and the answer will be:
答案将是：
[ 0.00803178 0.02526721 0.02593599 0.06308687 0.00281146]
回答by Saullo G. P. Castro
Check if your array is a 2-D array querying a.ndim. It may happen that you have a 1-D array of objects that, when printed, looks like a 2-D array. In this case you can convert it to a 2-D array doing:
检查您的数组是否为二维数组查询a.ndim。您可能有一个一维对象数组，打印时看起来像一个二维数组。在这种情况下，您可以将其转换为二维数组：
a = np.array(list(a))
or
或者
a = np.array(tuple(a))
回答by dkv
Recently had this same problem:
最近遇到了同样的问题：
np.sum(myndarray, axis=0)
was producing the "ValueError: 'axis' entry is out of bounds" error when executed inside a function called from a script. Strangely, when executing the line from the IPython shell it ran without a hitch (I checked that the myndarray variable was the same in both contexts).
在从脚本调用的函数内执行时，会产生“ValueError: 'axis' entry is out of bounds”错误。奇怪的是，当从 IPython shell 执行该行时，它顺利运行（我检查了 myndarray 变量在两种上下文中是否相同）。
I was working with Spyder, so perhaps it somehow got stuck in a weird state as a result of using the Python console + pdb, and/or IPython + ipdb, with the same set of variables. Restarting the IDE solved the issue, which also explains the others' inability to reproduce the bug.
我正在使用 Spyder，所以也许由于使用 Python 控制台 + pdb 和/或 IPython + ipdb 以及相同的变量集，它以某种方式陷入了一种奇怪的状态。重新启动 IDE 解决了问题，这也解释了其他人无法重现该错误。
回答by Evan McFarland
This is for someone who stumbles across this thread having the same error for a different reason.
这是为了那些由于不同原因偶然发现这个线程有相同错误的人。
I had the same error for any axis other than 0. I used a.ndimand saw there were 3 dimensions in the array. The fix was array.reshape(x,y)which made it 2d and then calling from the array worked.
对于 0 以外的任何轴，我都有相同的错误。我使用a.ndim并看到数组中有 3 个维度。修复是array.reshape(x,y)使它成为 2d 然后从数组调用工作。
