Clever Geek Handbook

Most elegant implementation of MATLAB "vec" function in NumPy

MATLAB has a function called vec

that takes a matrix and stacks the columns into one vector. For example, if we name the following matrix "X":

[1 2]
[3 4]

then vec(X)

will return a vector:

[1]
[3]
[2]
[4]

There seems to be no direct implementation of this, and " NumPy for MATLAB Users " has no direct equivalent.

So, given a numpy array (representing a matrix), what would a very elegant NumPy line be able to reproduce this result? Just curious to see how concise / elegant it can be. Thank!

+3


source to share


3 answers


You can use the "Fortran" option order

, for example. reshape

:



>>> import numpy as np
>>> a = np.array([[1, 2], [3, 4]])
>>> a.reshape((-1, 1), order="F")
array([[1],
       [3],
       [2],
       [4]])
+4


source


I think what you want flatten()

EG:



>>> import numpy as np
>>> a = np.array([[1, 2], [3, 4]])

>>> a.flatten('F')
>>> array([1, 3, 2, 4])

Thanks @jonrsharpe, I actually just looked! BTW: Transposing an array using a.T.flatten()

is an alternative to changing the order usingorder='F'

+3


source


For a one-dimensional result, use X.T.ravel()

or X.T.flatten()

. For a two dimensional column, use X.T.reshape(-1,1)

.

+1


source






More articles:


All Articles