ufunc“减法”不包含签名匹配类型的循环

Question

import scipy as sy
import numpy as np
import re

from numpy import *

P = np.array(input('Please enter the P vector:\n'))
Q = np.array(input('\nPlease enter the Q vector:\n'))
R = np.array(input('\nPlease enter the R vector:\n'))


print('\n\nP: ',P)
print('Q: ',Q)
print('R: ',R)

#calculations for question 1
PQ=Q-P
magPQ= sy.sqrt(dot(PQ,PQ))
#calculations for question 2
PR= R-P
#calculations for question 3
QP = P-Q
QR = R-Q
magQP=sy.sqrt(dot(QP,QP))
magQR=sy.sqrt(dot(QR,QR))
angle=sy.arccos(dot(QP,QR)/(magQP*magQR))
angled=angle*180/sy.pi
#calculations for question 4
Area = sy.sqrt(dot(cross(PQ,QR),cross(PQ,QR)))
#calculations for question 5
magPR =sy.sqrt(dot(PR,PR))
perimeter = magPQ+magQR+magPR

######################OUTPUT#######################

print("Question 1. What is the distance between P and Q?")
print("Answer: ",round(q1,4))

print("\nQuestion 2. What is the distance vector from P to R?")
print('Answer: ',q2)

print("\nQuestion 3. What is the angle between QP and QR?")
print('Answer: ',angled)

print("\nQuestion 4. What is the area of the triangle PQR?")
print('Answer: ',Area)

print("\nQuestion 5. What is the perimeter of triangle PQR?")
print('Answer: ',perimeter)

上面是我的代码，当我尝试编译时，我得到了错误

Traceback (most recent call last):
File "Homework1-3.py", line 17, in <module>
PQ=Q-P
TypeError: ufunc 'subtract' did not contain a loop with signature matching types dtype('<U6') dtype('<U6') dtype('<U6')

能够输入我的数组非常重要，因为我希望这个程序能够解决问题。它似乎使用字符串而不是浮点数。我该如何解决这个问题？

Answer 1

在我尝试过的交互式 ipython(3) 会话中：

In [373]: x=input('')
123,232,232

In [374]: x
Out[374]: '123,232,232'   # I got a single string from input

In [375]: np.array(x)
Out[375]: 
array('123,232,232',    # trying to make that an array - still string
      dtype='<U11')

In [377]: np.array(x.split(','))
Out[377]: 
array(['123', '232', '232'],     # better - 3 strings 
      dtype='<U3')

In [378]: np.array(x.split(','),dtype=float)
Out[378]: array([ 123.,  232.,  232.])    # good

我强烈建议您启动一个交互式 Python 会话，并测试代码的每一步。让它一块一块地工作。编写一开始就有错误的整个脚本并不是一种高效的编程方式或学习方式。

而当使用numpy数组时，看看shape和dtype；不要假设数组是有意义或有用的。检查一下。