From bug-octave-request at bevo dot che dot wisc dot edu Tue Nov 18 12:29:43 2003
Subject: [patch] make statistics/base/iqr.m work with matrices
From: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
To: Danilo Piazzalunga <danilopiazza at libero dot it>
Cc: bug-octave at bevo dot che dot wisc dot edu
Date: Tue, 18 Nov 2003 12:29:40 -0600

On 16-Nov-2003, Danilo Piazzalunga <danilopiazza at libero dot it> wrote:

| Bug report for Octave 2.1.50 configured for i386-pc-linux-gnu
| 
| Description:
| -----------
| 
|     The `iqr' function gives an error when its argument is a matrix, while
|     in documentation it is stated that it should compute the IQR for each
|     column.
| 
| Repeat-By:
| ---------
| 
|     octave:1> a=randn(10,6);
|     octave:2> iqr(a)
|     error: empirical_inv: data must be a vector
| 
| Fix:
| ---
| 
| This patch solves the problem. It was tested succesfully with scalars,row and
| column vectors, matrices (only real values -- complex values not tested).
| Empty matrices were not considered.
| 
| 
| --- octave/scripts/statistics/base/iqr.m        2003-11-05 04:06:15.000000000 
| +0100
| +++ octave/scripts/statistics/base/iqr.m.new    2003-11-16 13:44:03.000000000 
| +0100
|  at @ -34,6 +34,15 @@
|      usage ("iqr (x)");
|    endif
|  
| -  y = empirical_inv (3/4, x) - empirical_inv (1/4, x);
| +  if (rows (x) == 1)
| +    x = x.';
| +  endif
| +
| +  [r, c] = size(x);
| +  y = zeros(1, c);
| +
| +  for i = (1:c);
| +    y(i) = empirical_inv (3/4, x(:,i)) - empirical_inv (1/4, x(:,i));
| +  endfor
|  
|  endfunction

I applied this patch, but it might be a better solution to have
empirical_inv handle matrix args.

jwe



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------