why not using something like numpy.repeat? In [18]: B = numpy.random.rand(4,3) In [19]: A = numpy.repeat(B[:,:,numpy.newaxis],2,axis=2) In [20]: B.shape Out[20]: (4, 3) In [21]: A.shape Out[21]: (4, 3, 2) In [22]: numpy.all(A[:,:,0] == A[:,:,1]) Out[22]: True hth, L. On Fri, Apr 25, 2008 at 12:09 PM, Matthieu Brucher < matthieu.brucher@gmail.com> wrote:
2008/4/25, tournesol <tournesol33@gmail.com>:
Hi All.
I just want to conver Fortran 77 source to Python.
Here is my F77 source.
DIMENSION A(25,60,13),B(25,60,13)
open(15,file='data.dat') DO 60 K=1,2 READ(15,1602) ((B(I,J),J=1,60),I=1,25) 60 CONTINUE 1602 FORMAT(15I4)
DO 63 K=1,10 DO 62 I=1,25 DO 62 J=1,60 A(I,J,K)=B(I,J) 62 CONTINUE 63 CONTINUE END
Q1: Fortran-contiguous is ARRAY(row,colum,depth). How about the Python-contiguous ? array(depth,row,colum) ?
Default is C-contiguous, but you can you Fortran contiguous arrays.
Q2: How can I insert 1D to a 2D array and make it to
3D array. ex:) B:25x60 ==> A: 10X25X60
I don't understand what you want to do, but broadcasting allows copying several instances of an array into another one.
Matthieu -- French PhD student Website : http://matthieu-brucher.developpez.com/ Blogs : http://matt.eifelle.com and http://blog.developpez.com/?blog=92 LinkedIn : http://www.linkedin.com/in/matthieubrucher _______________________________________________ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion
-- Lorenzo Bolla lbolla@gmail.com http://lorenzobolla.emurse.com/