CODEWITHSUNDEEP
pythonlistmatrix
Zigii Wong
Zigii Wong

Reputation: 7826

Accessing the column of the matrix

New to python. I have a matrix likes: 

[['A', 'A', 'A', 'A'],
 ['B', 'Z', 'A', 'A'],
 ['0', 'B', 'A', 'A'],
 ['0', 'C', 'A', 'A']
]

Now how can I access the specific column ?

likes [A B 0 0] [A Z B C]

Thanks!

Upvotes: 0

Views: 204

Answers (3)

tom10
tom10

Reputation: 69182

To get a single column you can do (if x is your original array):

x_0 = [r[0] for r in x]

or if you do this a lot and can use numpy, it would look like this:

y = np.array(x)
y_0 = y[:,0]

An advantage of this approach over using zip and/or dstack is that there is minimal copying of data.

Upvotes: 1

Kasravnd
Kasravnd

Reputation: 107287

You can use zip built-in function to get the columns of an array :

>>> a=[['A', 'A', 'A', 'A'],
...  ['B', 'Z', 'A', 'A'],
...  ['0', 'B', 'A', 'A'],
...  ['0', 'C', 'A', 'A']
... ]
>>> 
>>> zip(*a)
[('A', 'B', '0', '0'), ('A', 'Z', 'B', 'C'), ('A', 'A', 'A', 'A'), ('A', 'A', 'A', 'A')]

Or if you have a matrix in numpy you can use dstack:

>>> import numpy as np
>>> a=np.array([['A', 'A', 'A', 'A'],['B', 'Z', 'A', 'A'],['0', 'B', 'A', 'A'],['0', 'C', 'A', 'A']])
>>> np.dstack(a)
array([[['A', 'B', '0', '0'],
        ['A', 'Z', 'B', 'C'],
        ['A', 'A', 'A', 'A'],
        ['A', 'A', 'A', 'A']]], 
      dtype='|S1')

Also you can just use indexing to get a specific column :

>>> a[:,0]
array(['A', 'B', '0', '0'], 
      dtype='|S1')

Upvotes: 1

Abhijit
Abhijit

Reputation: 63707

If you need to access a 2D matrix column wise, you have to transpose the matrix such that the columns can be accessed as rows. Python's zip builtin can be used for matrix transpose.

>>> zip(*mat)[0]
('A', 'B', '0', '0')

On the contrary, you are using a numpy array (recommended if you are performing an extended matrix manipulation), and use numpy indexing. This would be natural

>>> import numpy as np
>>> npmat = np.asarray(mat)
>>> npmat[:,0]
array(['A', 'B', '0', '0'], 
      dtype='|S1')

based on your requirement, you can either transpose the entire matrix using numpy.dstack (if you are using numpy) or use zip to transpose before indexing column wise

>>> np.dstack(mat)
array([[['A', 'B', '0', '0'],
        ['A', 'Z', 'B', 'C'],
        ['A', 'A', 'A', 'A'],
        ['A', 'A', 'A', 'A']]], 
      dtype='|S1')

Upvotes: 2

Related Questions