How to access a NumPy array by column

Accessing a NumPy-based array by a specific Column index can be achieved by indexing. NumPy follows standard 0-based indexing in Python.

Example:

Given array: 1 13 6
             9  4 7
             19 16 2

Input: print(NumPy_array_name[ :,2])
Output: [6 7 2]
Explanation: printing 3rd column

Access i_th column of a 2D Numpy Array in Python

Printing 1st row and 2nd column.

For column : numpy_Array_name[  : ,column] 
For row : numpy_Array_name[ row, :  ]

import numpy as np

array = [[1, 13, 6], [9, 4, 7], [19, 16, 2]]

# defining array
arr = np.array(array)

print('printing 0th row')
print(arr[0, :])

print('printing 2nd column')
print(arr[:, 2])

# multiple columns or rows can be selected as well
print('selecting 0th and 1st row simultaneously')
print(arr[:,[0,1]])

 Output :

printing 0th row
[ 1 13  6]

printing 2nd column
[6 7 2]

selecting 0th and 1st row simultaneously
[[ 1 13]
 [ 9  4]
 [19 16]]

Access the i_th column of a Numpy array using transpose

Transpose of the given array using the .T property and pass the index as a slicing index to print the array.

import numpy as np

arr = np.array([[1, 13, 6], [9, 4, 7], [19, 16, 2]])

# Access the ith column of a 2D NumPy array
column_i = arr.T[2]

#printing the column
print(column_i)

Output:

[6 7 2]

Access the i_th column of a Numpy array using list comprehension

Here, we access the ith element of the row and append it to a list using the list comprehension and printed the col.

import numpy as np

arr = np.array([[1, 13, 6], [9, 4, 7], [19, 16, 2]])

# Access the ith column of a 2D NumPy array
col = [row[1] for row in arr]

# printing the column
print(col)

Output:

[13, 4, 16]

Access the i_th column of a Numpy array using Ellipsis

Pass the ith index along with the ellipsis to print the returned array column.

For column : numpy_Array_name[...,column]
For row : numpy_Array_name[row,...]
where '...' represents no of elements in the given row or column 

Note: This is not a very practical method but one must know as much as one can.

import numpy as np

# defining array
array = [[1, 13, 6], [9, 4, 7], [19, 16, 2]]

# converting to numpy array
arr = np.array(array)

print('selecting 0th column')
print(arr[..., 0])

print('selecting 1st row')
print(arr[1, ...])

Output:

selecting 0th column
[ 1  9 19]

selecting 1st row
[9 4 7]