python 3 and numpy array

Question

I am trying write a code to work with arrays. I have arrays for variables and I am trying to call certain numbers from those variables for a defined function. I have attached a simplified snip of what I am trying to accomplish. The code works fine when I just have the variables as a single number but when I incorporate arrays it doesn't work.

import numpy as np

noe = 3


# E of members ksi
E = 29000*np.ones((1, noe))
# Area of members in^2
A = 1000*np.array([20, 30, 20])
# length of members in
L = np.array([144, 240, 144])
# moment of inertia of members in^4
I = np.array([600, 1200, 600])

def k_frame_local_6x6(E, I, A, L):
    return (E/L)*np.array([[A, 0, 0, -A, 0, 0], [0, (12*I)/(L**2), (6*I)/L, 0, (-12*I)/(L**2), (6*I)/L], [0, (6*I)/L, 4*I, 0, (-6*I)/L, 2*I], [-A, 0, 0, A, 0, 0], [0, (-12*I)/(L**2), (-6*I)/L, 0, (12*I)/(L**2), (-6*I)/L], [0, (6*I)/L, 2*I, 0, (-6*I)/L, 4*I]])


m=k_frame_local_6x6(E[0, 1], I[0, 1], A[0, 1], L[0, 1])

print(m)

The Error I receive is "IndexError: too many indices for array"

When I manually enter the values I am trying to get the function to read it works, that looks like this:

def k_frame_local_6x6(E, I, A, L):
    return (E/L)*np.array([[A, 0, 0, -A, 0, 0], [0, (12*I)/(L**2), (6*I)/L, 0, (-12*I)/(L**2), (6*I)/L], [0, (6*I)/L, 4*I, 0, (-6*I)/L, 2*I], [-A, 0, 0, A, 0, 0], [0, (-12*I)/(L**2), (-6*I)/L, 0, (12*I)/(L**2), (-6*I)/L], [0, (6*I)/L, 2*I, 0, (-6*I)/L, 4*I]])

m=k_frame_local_6x6(29000, 1200, 30000, 240)

print(m)

and the results I get are:

[[  3.62500000e+06   0.00000000e+00   0.00000000e+00  -3.62500000e+06
    0.00000000e+00   0.00000000e+00]
 [  0.00000000e+00   3.02083333e+01   3.62500000e+03   0.00000000e+00
   -3.02083333e+01   3.62500000e+03]
 [  0.00000000e+00   3.62500000e+03   5.80000000e+05   0.00000000e+00
   -3.62500000e+03   2.90000000e+05]
 [ -3.62500000e+06   0.00000000e+00   0.00000000e+00   3.62500000e+06
    0.00000000e+00   0.00000000e+00]
 [  0.00000000e+00  -3.02083333e+01  -3.62500000e+03   0.00000000e+00
    3.02083333e+01  -3.62500000e+03]
 [  0.00000000e+00   3.62500000e+03   2.90000000e+05   0.00000000e+00
   -3.62500000e+03   5.80000000e+05]]

One thing I have just noticed is with my ones array I have two sets of brackets:

E = 29000*np.ones((1, noe))

and the result is:

array([[ 29000.,  29000.,  29000.]])

However with the rest of the arrays I only get one bracket set:

A = 1000*np.array([20, 30, 20])

gives me:

array([20000, 30000, 20000])

Let me know if I need to clarify anything else. Thank you all!!!

Sorry, I could not understand what you wrote. Could you post the script that does not work, the expected output and the error you are getting? — Andrea Corbellini
– Andrea Corbellini, Commented Jan 6, 2016 at 22:05
A, I and L are one-dimensional arrays, so eg. A[0, 1] fails. Take a look at NumPy Indexing — Finwood
– Finwood, Commented Jan 6, 2016 at 22:07
I shall repost again once I get back to my computer this time with some more clarity — Justin Kordas
– Justin Kordas, Commented Jan 6, 2016 at 22:30

hpaulj · Accepted Answer · 2016-01-07 02:38:29Z

1

Your k_frame_local_6x6 - which is ugly, virtually unreadable for ordinary humans, produces a 6x6 array when given 4 numbers.

But as best I can tell, none of the terms is designed to work with arrays:

def k_frame_local_6x6(E, I, A, L):
    temp = np.array([[A, 0, 0, -A, 0, 0], 
                     [0, (12*I)/(L**2), (6*I)/L, 0, (-12*I)/(L**2), (6*I)/L], 
                     [0, (6*I)/L, 4*I, 0, (-6*I)/L, 2*I], 
                     [-A, 0, 0, A, 0, 0], 
                     [0, (-12*I)/(L**2), (-6*I)/L, 0, (12*I)/(L**2), (-6*I)/L], 
                     [0, (6*I)/L, 2*I, 0, (-6*I)/L, 4*I]
                     ])
     return (E/L)*temp

Try, just for example the simplest row of terms

np.array([-A, 0,0,A,0,0])

That's fine if A is a number. But it does not work if A is an array. Well it does work, but does not return something meaningful

In [105]: A = 1000*np.array([20, 30, 20])

In [106]: np.array([A, 0, 0, -A, 0, 0])
Out[106]: 
array([array([20000, 30000, 20000]), 0, 0, array([-20000, -30000, -20000]),
       0, 0], dtype=object)

You try to pass A[0, 1]. Have you tried that expression?

In [107]: A[0,1]
---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-107-33f18b7a38c5> in <module>()
----> 1 A[0,1]

IndexError: too many indices

In [108]: A.shape
Out[108]: (3,)

It does not work because A is defined as 1d 3 element array. You can't index it with 2 values.

E is defined as 2d array, but all the rest are 1d. Is that intentional, or a mistake?

It looks like you are jumping into defining a complex array without understanding the basics of numpy array creation and indexing. Or are you coming from a MATLAB world where everything is 2d (or more)? In numpy arrays can be 1d.

edited Jan 7, 2016 at 2:38

answered Jan 6, 2016 at 23:43

hpaulj

233k14 gold badges260 silver badges392 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Justin Kordas Over a year ago

HAHA you caught me. I am coming from MATLAB. I am doing my Masters in Structural Engineering and we use MATLAB for all academic work but I have been told by many people that firms don't use MATLAB to much because of the price. I am trying to convert my MATLAB code to Python. The variables E, A, I, and L are meant to be 1xn matrices that correspond to individual elements. The function will go in a while loop and create a 3D matrix (6x6xnum_of_elements).

ali_m Over a year ago

@JustinKordas you can use np.r_ to emulate MATLAB's [] array construction syntax, e.g. np.r_[A, 0, 0, -A, 0, 0]. Also, read this.

Collectives™ on Stack Overflow

python 3 and numpy array

1 Answer 1

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related