A better way to act on a dim of an N-d array?

[John W. Eaton]

On 31-Jan-2008, Ben Abbott wrote:

| 
| On Jan 29, 2008, at 10:35 PM, Ben Abbott wrote:
| 
| >
| > On Jan 29, 2008, at 8:30 PM, Ben Abbott wrote:
| >
| >> While working on quantile and prctile functions for Octave, I am in  
| >> need to functionally operate on the dim-th dimension of a N-d array.
| 
| What I'm looking to do is allow a specific dimension of an N-d array  
| to be acted upon by a function that returns a vector of a consistent  
| length.
| 
| The algirithms for calculating the quanitles of a sample are  
| sufficiently varied that the methods used by mean(), median(), etc, to  
| support N-d arrays, are insufficient for my application.
| 
| As I am not a competent c/c++ programmer, I chose to do the job in an  
| m-file ... but is there a better way to do this?

I'm not sure, but it does seem fairly complex.  But if it is necessary
to use your dimfunc, then you should probably use

  function y = dimfunc (func, x, dim, varargin)

instead of 

  function y = dimfunc (func, x, dim, arg1, arg2, arg3, arg4, arg5)

so that dimfunc can handle any number of additional arguments, not
just 5, and so you can replace

    if (nargin == 3)
      temp = func (xdim(:));
    elseif (nargin == 4)
      temp = func (xdim(:), arg1);
    elseif (nargin == 5)
      temp = func (xdim(:), arg1, arg2);
    elseif (nargin == 6)
      temp = func (xdim(:), arg1, arg2, arg3);
    elseif (nargin == 7)
      temp = func (xdim(:), arg1, arg2, arg3, arg4);
    elseif (nargin == 8)
      temp = func (xdim(:), arg1, arg2, arg3, arg4, arg5);
    endif

with

   temp = func (xdim(:), varargin{:});

jwe