sparse bug or unexplicable behaviour

[Francesco Potortì]

I observe behaviour that I cannot explicate on Debian Octave, either I
miss something obvious or it'a a bug.  But since that would be a bug in
a fundamental feature of Octave, I am puzzled.

I have three vectors from which I build a sparse matrix.  but I observe
that the min value of the third vector (whioch is negative) is different
from the min value of the sparse matrix:

octave-cli> version
ans = 3.8.2
octave-cli> clear p
octave-cli> size(prowidx)
ans =

   111263        1

octave-cli> min(prowidx)
ans =  1
octave-cli> max(prowidx)
ans =  2579
octave-cli> min(pcolidx)
ans =  1
octave-cli> max(pcolidx)
ans =  138
octave-cli> length(pt(:,2))
ans =  111263
octave-cli> p = full(sparse(prowidx, pcolidx, pt(:,2)));
octave-cli> size(p)
ans =

   2579    138

octave-cli> max(pt(:,2))
ans = -30
octave-cli> max(p(p!=0)(:))
ans = -30
octave-cli> min(pt(:,2))
ans = -94
octave-cli> min(p(:))
ans = -182

Is this a bug or am I missing something?

-- 
Francesco Potortì (ricercatore)        Voice:  +39.050.621.3058
ISTI - Area della ricerca CNR          Mobile: +39.348.8283.107
via G. Moruzzi 1, I-56124 Pisa         Skype:  wnlabisti
(entrance 20, 1st floor, room C71)     Web:    http://fly.isti.cnr.it