Assignment of the last variable is being persistent, how to solve it?-CodePudding

I am trying to do assignments to different variables in my code, but for some reason, the last one is being persistent. I have tracked the problem until the example below. If someone is able to explain this behavior it will be great.

import numpy as np 
metric_a = metric_b = metric_c = metric_d= np.zeros(10) 

j = 0
temp_a = 1
temp_b = 2
temp_c = 1
temp_d = 5
          
print('temp_a=', temp_a, '  temp_b=', temp_b, ' temp_c=', temp_c, ' temp_d=', temp_d)   
print('metric_a[' str(j) ']=', metric_a[j], '   metric_b[' str(j) ']=', metric_b[j], '    metric_c[' str(j) ']=',metric_c[j], '   metric_d[' str(j) ']=',metric_d[j])
print('j:',j, '\n')

metric_a[j] = temp_a 
metric_b[j] = temp_b
metric_c[j] = temp_c
metric_d[j] = temp_d

print('temp_a=', temp_a, '     temp_b=', temp_b, '     temp_c=', temp_c, '     temp_d=', temp_d)   
print('metric_a[' str(j) ']=', metric_a[j], '   metric_b[' str(j) ']=', metric_b[j], '    metric_c[' str(j) ']=',metric_c[j], '   metric_d[' str(j) ']=',metric_d[j])

Here is the output

temp_a= 1   temp_b= 2  temp_c= 1  temp_d= 5
metric_a[0]= 0.0    metric_b[0]= 0.0     metric_c[0]= 0.0    metric_d[0]= 0.0
j: 0 

temp_a= 1      temp_b= 2      temp_c= 1      temp_d= 5
metric_a[0]= 5.0    metric_b[0]= 5.0     metric_c[0]= 5.0    metric_d[0]= 5.0

When I have been waiting for

metric_a[0]= 1.0    metric_b[0]= 2.0     metric_c[0]= 1.0    metric_d[0]= 5.0

CodePudding user response：

You can do that with astype like that :-

metric_a = metric_a.astype('float')
metric_a[metric_a == j] = temp_a 
print(metric_a[j])

output:-

1.0

CodePudding user response：

The reason was given by Jordan Hyatt in his comment. The problem is the declaration of the metric_X variables.

To build on his answer, be aware that just declaring it in an indivual line won't work. You have to make sure to use the np.zeros or at least use .copy().

What won't work as you are still referning the same memory space:

metric_a = np.zeros(10) 
metric_b = metric_a 
metric_c = metric_a 
metric_d = metric_a

What will work as it creates a different object for each declared variable:

metric_a = np.zeros(10) 
metric_b = metric_a.copy() # or np.zeros(10) 
metric_c = metric_a.copy() # or np.zeros(10) 
metric_d = metric_a.copy() # or np.zeros(10)