Re: Partial list of missing NDArray functionality

[David Bateman]

Another thing to add is all of the ::convert_to_string_internal
functions have to be converted. Otherwise toascii, etc won't
work. This is actually used in lots of places, that you might
not think of (for instance sortstr in octave-forge), and basically
makes NDArrays of strings not usuable..

D.

According to Alois Schloegl <address@hidden> (on 03/30/04):
> 
> Hi David,
> 
> Please add also the following functions to your (partial) list,
> REPMAT.M
> 
> From your list, the following functions are available in 
> octave-forge/extra/NaN/
> mod
> rem
> std
> var
> median
> 
> as well as a few others like
> normpdf
> normcdf
> norminv
> meandev
> moment
> mad
> skewness
> kurtosis
> center
> zscore
> harmmean
> geomean
> trimean
> tpdf
> tcdf
> tinv
> 
> 
> Regards,
>   Alois
> 
> 
> David Bateman wrote:
> 
> >Looking into missing NDArray functions compared to matlab in 2.1.57
> >(not including min/max and filter which I've converted) I've compiled
> >the following partial list 
> >
> >b = unwrap(a);
> >b = cplxpair(a);
> >b = mod(a,2);
> >b = rem(a,2);
> >b = airy(a);
> >b = besselj(1,a);
> >b = bessely(1,a);
> >b = besselh(1,a);
> >b = besseli(1,a);
> >b = besselk(1,a);
> >b = betainc(2,a,a);
> >b = erfinv(a);
> >b = gammainc(2,a);
> >b = legendre(0,a);
> >b = std(a);
> >b = var(a)
> >b = median(a);
> >b = convn(a,a);          % Doesn't exist
> >b = sort(a);
> >b = fftshift(a);
> >b = ifftshift(a);
> >
> >%% This one might be a bug in matlab...
> >b = nextpow2(a);
> >
> >%% Are these really N-D array issues
> >b = delaunayn(??);       % Doesn't exist
> >b = interp3(??);         % Doesn't exist
> >b = interpn(??);         % Doesn't exist
> >b = griddata3(??);       % Doesn't exist
> >b = griddatan(??);       % Doesn't exist
> >
> >Some of these functions are in octave-forge and some aren't strictly
> >NDArray issues. For example "nextpow2" treats matrix and NDArray args
> >as "nextpow2(length(x))" in matlab, whereas octave just crashes with
> >an error if the arg isn't a scalar or a vector. Perhaps we should not
> >do bug for bug compatiability of this one.
> >
> >However, most of the functions above are missing NDArray functionality.
> >There are probably other functions as well. 
> >
> >Note that this e-mail is not a criticism, as there are an amazing
> >number of functions already converted to be NDArray aware. But rather
> >a list of things to do for those that are interested. If you know of
> >any other functions to convert perhaps you can add them to the list.
> >
> >Regards
> >David
> >
> > 
> >

-- 
David Bateman                                address@hidden
Motorola CRM                                 +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax) 
91193 Gif-Sur-Yvette FRANCE

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary