Path: csiph.com!fu-berlin.de!uni-berlin.de!not-for-mail
From: Michael Selik <michael.selik@gmail.com>
Newsgroups: comp.lang.python
Subject: Re: reshape and keep x,y,z ordering
Date: Tue, 07 Jun 2016 13:16:37 +0000
Lines: 33
Message-ID: <mailman.59.1465305415.2306.python-list@python.org>
References: <5ef3691e-2317-41df-ad60-3d2450841413@googlegroups.com> <CAGgTfkMSF4Zq7BuCR4+wFXp-D4nJbJXL-YScsuW=pDmGWPiNgQ@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
In-Reply-To: <5ef3691e-2317-41df-ad60-3d2450841413@googlegroups.com>
Precedence: list
Xref: csiph.com comp.lang.python:109626

On Tue, Jun 7, 2016 at 7:31 AM Heli <hemla21@gmail.com> wrote:

> Hello,
> I have a question regarding reshaping numpy array.
>
> I either have a 1D array that I need to reshape to a 3D array or a 3D
> array to reshape to a 1d numpy array.
>
> In both of these cases it is assumed that data follows x,y,z ordering.
> and I use the following to reshape the numpy array.
>
> new_1d_array=np.reshape(3d.transpose(),(3d_nx*3d_ny*3d_nz,))
>
> new_3d_array=np.reshape(1d,((3d_x,3d_y,3d_z)).transpose())
>
> My question is if there is anyway that reshape would keep x,y,z ordering
> that would not require transpose? and if there is a better more efficient
> way to do this?


    >>> a = np.arange(9).reshape(3,3)
    >>> a
    array([[0, 1, 2],
           [3, 4, 5],
           [6, 7, 8]])
    >>> a.flatten()
    array([0, 1, 2, 3, 4, 5, 6, 7, 8])
    >>> a.flatten('F')
    array([0, 3, 6, 1, 4, 7, 2, 5, 8])

Does this work for you?
The flatten method normally goes row by row, but you can specify FORTRAN
style column by column flattening.