Numpy array - stack multiple columns into one using reshape

Question

For a 2D array like this:

table = np.array([[11,12,13],[21,22,23],[31,32,33],[41,42,43]])

Is it possible to use np.reshape on table to get an array single_column where each column of table is stacked vertically? This can be accomplished by splitting table and combining with vstack.

single_column = np.vstack(np.hsplit(table , table .shape[1]))

Reshape can combine all the rows into a single row, I'm wondering if it can combine the columns as well to make the code cleaner and possibly faster.

single_row = table.reshape(-1)

score 2 · Accepted Answer · answered Mar 31 '19 at 19:47

2

You can transpose first, then reshape:

table.T.reshape(-1, 1)

array([[11],
       [21],
       [31],
       [41],
       [12],
       [22],
       [32],
       [42],
       [13],
       [23],
       [33],
       [43]])

answered Mar 31 '19 at 19:47

cs95

379,657
97
704
746

1

This is much more concise than my hasty realization. – rovyko Mar 31 '19 at 19:50

score 1 · Answer 2 · answered Mar 31 '19 at 21:02

A few more approaches are:

1) flattening using Fotran order, followed by explicit promotion as a column vector
2) reshaping using Fortran order, followed by explicit promotion as a column vector

# using approach 1
In [200]: table.flatten(order='F')[:, np.newaxis]
Out[200]: 
array([[11],
       [21],
       [31],
       [41],
       [12],
       [22],
       [32],
       [42],
       [13],
       [23],
       [33],
       [43]])

# using approach 2
In [202]: table.reshape(table.size, order='F')[:, np.newaxis]
Out[202]: 
array([[11],
       [21],
       [31],
       [41],
       [12],
       [22],
       [32],
       [42],
       [13],
       [23],
       [33],
       [43]])

Numpy array - stack multiple columns into one using reshape

2 Answers2

Linked