Combining rows numpy to one single array

Question

I have a while statement that produces an array in numpy

while(i<int(sp)): 
   body=(np.bincount(un[i].ravel().astype('int'), weights=step_periods[i].ravel())) 
   print(body) 
   i=i+1

each iteration produces an array of like the following:

1st    [  0.   0.   0.  30.]

2nd    [  0.   0.  21.  18.  15.]

3rd    [  0.  24.  27.   0.   3.]

My first issue is that if the first array has "0" as the last value, it will leave it out of the array. Is there a way to convert it from:

[  0.   0.   0.  30.]

to:

[  0.   0.   0.  30.  0.]

From there I would like to simply append each array to a master array so that the final output is something similar to:

    [[  0.   0.   0.  30.  0.0],

    [  0.   0.  21.  18.  15.],

   [  0.  24.  27.   0.   3.]]

I've looked into appending and vstack, but can't get it to work in a "while" statement, or possibly because they aren't all the same size due to the ending "0" being ommited!

Thanks!

It would help to show the while loop that is generating these subarrays. — mgilson
– mgilson, Commented Jul 30, 2012 at 16:45
You should post that as an edit to the post. Putting it in comments makes it quite difficult to tell what is going on. — mgilson
– mgilson, Commented Jul 30, 2012 at 17:04
Sorry I'll do this right now, I'm new to these forums! I appreciate the help! — Eric Escobar
– Eric Escobar, Commented Jul 30, 2012 at 17:07
I'm not a numpy expert... but from itertools import izip_longest; np.vstack(izip_longest(a, b, c, fillvalue=0.)).T works... - requires Py2.6+ I believe... — Jon Clements
– Jon Clements, Commented Jul 30, 2012 at 17:13

mgilson · Accepted Answer · 2012-07-30 18:02:42Z

1

One solution -- if you can guarantee that the length of body is always going to be less than (or equal to) 5, you can pre-allocate the array:

import numpy as np
my_array = np.zeros((int(sp)-i,5),dtype=int)
for ii in range(i,int(sp)):
    body=(np.bincount(un[i].ravel().astype('int'), weights=step_periods[i].ravel())) 
    my_array[ii,:len(body)]=body

You could even add a check to see if body got too wide and create a new array, assigning the old portion into it ... But at that point, maybe a stacking operation is what is needed.

edited Jul 30, 2012 at 18:02

answered Jul 30, 2012 at 17:21

mgilson

312k70 gold badges656 silver badges722 bronze badges

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

8 Comments

Eric Escobar Over a year ago

Can you explain what ii is? and I get the error "operands could not be broadcast together with shapes (3) (4)"

Eric Escobar Over a year ago

Looking at the np.zeros i think that the 5, and the int(sp)-i should be switched. But when I do switch it, i get the error "invalid index"

Eric Escobar Over a year ago

after switching them, I realized my i=i+1 was not indented! THANK YOU! :) it works perfectly!

mgilson Over a year ago

@EricEscobar -- You're right, they should have been switched. (sorry about that)

Eric Escobar Over a year ago

could you point me somewhere I could learn more about what this means "my_array[ii,:len(body)]=body " it works well for what I need I just want to understand it more. Thank you again!

|

Collectives™ on Stack Overflow

Combining rows numpy to one single array

1 Answer 1

8 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

8 Comments

Your Answer

Sign up or log in

Post as a guest

Related