tir, 15 01 2008 kl. 06:06 -0800, skrev Rolf Fabian:
> Up to now I thought skipping NaN elements is the objective of the 'skipnan' -
> package.
>
> If this was implemented into mainstream Octave and changed into a 'feature'
> now, this should have been documented somewhere in order to prevent
> people from wasting their precious time.
>
> As long as it is not documented I consider this feature as a bug.
Isn't this just the way NaN works. I mean
octave:3> x=[1,NaN;NaN,4];
octave:4> x > 2
ans =
0 0
0 1
octave:5> x < 2
ans =
1 0
0 0
So, what else should min/max return that would be consistent with how
scalars and NaNs are compared?
Søren
>
>
> Rolf Fabian
> < r dot fabian at jacobs-university dot de>
>
>
> octave-3.0.0.exe:1> computer, version
> i686-pc-msdosmsvc
> ans = 3.0.0
> octave-3.0.0.exe:2> x=[1,NaN;NaN,4]
> x =
>
> 1 NaN
> NaN 4
>
> octave-3.0.0.exe:3> min(x)
> ans =
>
> 1 4
>
> octave-3.0.0.exe:4> max(x)
> ans =
>
> 1 4
>
> octave-3.0.0.exe:5> min(x.')
> ans =
>
> 1 4
>
> octave-3.0.0.exe:6> max(x.')
> ans =
>
> 1 4
>
> -----
> Rolf Fabian
> <r dot fabian at jacobs-university dot de>
>
_______________________________________________
Bug-octave mailing list
Bug-octave@...
https://www.cae.wisc.edu/mailman/listinfo/bug-octave
