Contiguous Array

[Stéfan van der Walt]

Hi Marc

On Wed, Jul 31, 2013 at 9:39 AM, Marc de Klerk <deklerkmc at gmail.com> wrote:
> I've been using np.ravel(). This morning I tried to lookup the difference
> between np.ravel() and np.ascontiguousarray(). Does anybody know?

np.ravel docstring:

A 1-D array, containing the elements of the input, is returned.  A copy is
made only if needed.

np.ascontiguousarray:

Return a contiguous array in memory (C order).

In [13]: x = np.array(([1, 2], [3, 4]))

In [14]: np.ravel(x)
Out[14]: array([1, 2, 3, 4])

In [15]: np.ascontiguousarray(x)
Out[15]:
array([[1, 2],
       [3, 4]])

Since the default "order" parameter for ravel is "C", and 1-D arrays
with order "C" can only be contiguous, you happen to get contiguous
memory layout.  However, if you want to preserve shape, you cannot use
ravel.

Stéfan