Subtracting two lists but the values are not adding up in Python

Question

I want to perform the operation: roll_all-theta_all . Both of them are lists containing two values each and the result is supposed to be listed in the column named 'roll' on my panda.DataFrame . However when I do so, the 2nd row value of column 'roll' does not add up. Although it cannot be seen because I perform theta_all operation in other parts of my code, I obtain the values the same way I obtain the values of roll_all , ie, by creating an empty list and then append the values to it. Here is my code:

pyr = pd.DataFrame(columns = ['pitch','yaw','roll'])

for para in GTparas:  
    Mr = np.array([[np.cos(para[2]), -np.sin(para[2]), 0],[np.sin(para[2]), np.cos(para[2]), 0],[0, 0, 1]])
    My = np.array([[np.cos(para[1]), 0, np.sin(para[1])],[0, 1, 0],[-np.sin(para[1]), 0, np.cos(para[1])]])
    Mp = np.array([[1, 0, 0],[0, np.cos(para[0]), -np.sin(para[0])],[0, np.sin(para[0]), np.cos(para[0])]])
    M1 = np.dot(Mr,My)
    M = np.dot(M1,Mp)

    pitch = np.arctan(M[2,1]/M[2,2])
    yaw = np.arctan(M[2,0]/((((M[0,0])**2)+((M[1,0])**2))**(1/2)))
    roll = np.arctan(M[1,0]/M[0,0])
    
    roll_all = []
    roll_all.append(roll)
    print(roll_all)
    c = [a - b for a, b in zip(roll_all, theta_all)]
    #c = np.array(roll_all) - np.array(theta_all)
  
    pyr = pyr.append({'pitch':pitch,'yaw':yaw,'roll':c},ignore_index = True) # from Z-Y-X to X-Y-Z
print(pyr)

I am also attaching the values in my lists for theta_all and the second image shows the values in roll_all and what I get when I print pyr .

Answer 1

Summarizing:

pyr = pd.DataFrame(columns = ['pitch','yaw','roll'])

for para,theta in zip(GTparas, theta_all):
    Mr = np.array([[np.cos(para[2]), -np.sin(para[2]), 0],[np.sin(para[2]), np.cos(para[2]), 0],[0, 0, 1]])
    My = np.array([[np.cos(para[1]), 0, np.sin(para[1])],[0, 1, 0],[-np.sin(para[1]), 0, np.cos(para[1])]])
    Mp = np.array([[1, 0, 0],[0, np.cos(para[0]), -np.sin(para[0])],[0, np.sin(para[0]), np.cos(para[0])]])
    M1 = np.dot(Mr,My)
    M = np.dot(M1,Mp)

    pitch = np.arctan(M[2,1]/M[2,2])
    yaw = np.arctan(M[2,0]/((((M[0,0])**2)+((M[1,0])**2))**(1/2)))
    roll = np.arctan(M[1,0]/M[0,0]) - theta
      
    pyr = pyr.append({'pitch':pitch,'yaw':yaw,'roll':roll},ignore_index = True) 

print(pyr)