np.linalg.norm AxisError: axis 1 is out of bounds for array of dimension 1

Question

I want to use np.linalg.norm to calculate the norm of a row vector, and then use this norm to normalize the row vector, as I wrote in the code. I give an initial value to the vector x, but after I run this code I always get:

AxisError: axis 1 is out of bounds for array of dimension 1

So I'm confused and don't know what's the problem. Here is my code:

import numpy as np
    
def normalizeRows(x):
    x_norm = np.linalg.norm(x, axis=1, keepdims=True)
    x_normalized = x / x_norm
    
    return x_normalized
 
x = np.array([1, 2, 3]) 
print(normalizeRows(x)) 

And here is the error:

---------------------------------------------------------------------------
AxisError                                 Traceback (most recent call last)
<ipython-input-16-155a4cdd9bf8> in <module>
     10 
     11 x = np.array([1, 2, 3])
---> 12 print(normalizeRows(x))
     13 
     14 

<ipython-input-16-155a4cdd9bf8> in normalizeRows(x)
      4 
      5 def normalizeRows(x):
----> 6     x_norm = np.linalg.norm(x, axis=1, keepdims=True)
      7     x_normalized = x / x_norm
      8 

<__array_function__ internals> in norm(*args, **kwargs)

d:\programs\python39\lib\site-packages\numpy\linalg\linalg.py in norm(x, ord, axis, keepdims)
   2558             # special case for speedup
   2559             s = (x.conj() * x).real
-> 2560             return sqrt(add.reduce(s, axis=axis, keepdims=keepdims))
   2561         # None of the str-type keywords for ord ('fro', 'nuc')
   2562         # are valid for vectors

AxisError: axis 1 is out of bounds for array of dimension 1

Could somebody tell me why this is wrong and how to fix it? Thanks!

Answer 1

This gives you axis error because your array is a 1d array, and since :
"In a multi-dimensional NumPy array, axis 1 is the second axis. When we're talking about 2-d and multi-dimensional arrays, axis 1 is the axis that runs horizontally across the columns" quoted from link

All you have to do is change your axis to 0

import numpy as np

def normalizeRows(x):
    x_norm = np.linalg.norm(x, axis=0, keepdims=True)
    x_normalized = x / x_norm
    
    return x_normalized
 
x = np.array([1, 2, 3]) 
print(normalizeRows(x)) 

[0.26726124 0.53452248 0.80178373]