numpy equivalent of matlab matrix code

Question

I have the next code on Matlab:

cols = 7;
rows = 8;

redGrd = [0 0 0 0 0 1; 1 1 1 1 0 1; 0 0 0 0 0 1; 1 0 1 1 0 1];

redGrd(:,1)=-9999;
redGrd(1,:)=-9999;
redGrd(:,cols)=-9999;
redGrd(rows,:)=-9999

this is the result for matlab:

-9999  -9999  -9999  -9999  -9999  -9999  -9999
-9999      1      1      1      0      1  -9999
-9999      0      0      0      0      1  -9999
-9999      0      1      1      0      1  -9999
0      0      0      0      0      0      0
0      0      0      0      0      0      0
0      0      0      0      0      0      0
-9999  -9999  -9999  -9999  -9999  -9999  -9999

I want to do the same on python using numpy then I do this:

import numpy as np

cols = 7
rows = 8

redGrd = np.array([[0,0,0,0,0,1],[1,1,1,1,0,1],[0,0,0,0,0,1],[1,0,1,1,0,1]])

redGrd[:,0] = -9999
redGrd[0,:] = -9999
redGrd[:,cols] = -9999 
redGrd[rows,:] = -9999

But the two last command dont works.

Glorfindel · Accepted Answer · 2023-01-12 20:05:44Z

3

The first line (redGrd[:,cols] = -9999) won't work, because python uses zero-based indexing, whereas Matlab starts counting from 1. So the last column in your numpy n-dimensional array has the number cols-1, or just -1 (shorthand).

A very useful resource, if you're coming from Matlab, is NumPy for Matlab Users. It lists a good number of pitfalls, such as this one.

Now, the second line (redGrd[rows,:] = -9999) will fail because numpy will not attempt to increase the size of an array automatically, whereas Matlab does.

From a programmer's perspective, the former makes more sense, as it can prevent you from changing an array by accident.

If you want to "increase" the size of the array, you'll have to do it explicitly, with

    redGrd = np.vstack((redGrd, np.zeros(rows - redGrd.shape[0], redGrd.shape[1])))

edited Jan 12, 2023 at 20:05

Glorfindel

22.8k13 gold badges97 silver badges124 bronze badges

answered Dec 11, 2014 at 22:56

Oliver W.

13.6k3 gold badges41 silver badges52 bronze badges

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

2 Comments

Manuel Alejandro Over a year ago

Hi, I used the comand but I got an Error message: Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: data type not understood

Manuel Alejandro Over a year ago

I forgot say Thanks so much.

user252592 · Accepted Answer · 2014-12-11 23:04:55Z

1

With your initialization you are defining a 4x6 matrix:

redGrd = np.array([[0,0,0,0,0,1],[1,1,1,1,0,1],[0,0,0,0,0,1],[1,0,1,1,0,1]])

You can't access row 7 when you have only 4 rows. You should create a row x col matrix first, then fill the values as you wish. You can create the array like this:

redGrd = np.zeros((rows, cols))

Also, please note that in python you have the indices from 0, so you should use cols -1 instead of cols and rows -1 instead of rows

answered Dec 11, 2014 at 23:04

user252592

1413 bronze badges

Comments

unutbu · Accepted Answer · 2014-12-12 13:37:07Z

1

import numpy as np
cols, rows = 7, 8

redGrd = np.mat('0 0 0 0 0 1; 1 1 1 1 0 1; 0 0 0 0 0 1; 1 0 1 1 0 1')

# pad `redGrd[1:, 1:]` with -9999 on the top, left and right
redGrd = np.pad(redGrd[1:, 1:], ((1,0),(1,1)), mode='constant', constant_values=-9999)

# expand `redGrd` to the desired shape
redGrd.resize((rows, cols))    

# fill the last row with -9999
redGrd[rows-1, :] = -9999

print(redGrd)

yields

[[-9999 -9999 -9999 -9999 -9999 -9999 -9999]
 [-9999     1     1     1     0     1 -9999]
 [-9999     0     0     0     0     1 -9999]
 [-9999     0     1     1     0     1 -9999]
 [    0     0     0     0     0     0     0]
 [    0     0     0     0     0     0     0]
 [    0     0     0     0     0     0     0]
 [-9999 -9999 -9999 -9999 -9999 -9999 -9999]]

edited Dec 12, 2014 at 13:37

answered Dec 12, 2014 at 13:30

unutbu

886k197 gold badges1.9k silver badges1.7k bronze badges

3 Comments

Manuel Alejandro Over a year ago

I use this command but in the np.pad I have an error message: Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: 'module' object has no attribute 'pad'

Manuel Alejandro Over a year ago

I forgot say Thanks so much.

unutbu Over a year ago

np.pad was introduced in NumPy version 1.7.0. You could be getting that error if your NumPy version is older.

Collectives™ on Stack Overflow

numpy equivalent of matlab matrix code

3 Answers 3

2 Comments

Comments

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related