Re: forpart of zero and fixed length

Hi Bill,

I understand your explanation, but this design decision looks odd to me.
You make a distinction between length and size, but personally I care only
about size in your terminology.
The somewhat ugly workaround for generating partitions of a given size
whether or not amin is zero is
forpart(p=n, print(p), [amin,amax], if(amin>0,[psize,psize],psize))
Also, there seems to be no simple way to generate all partitions with zero
parts for a given range of sizes, as specifying the size range (oddly)
disables the zero parts.

Regards,
Max


On Tue, May 20, 2025 at 3:30 PM Bill Allombert <
[email protected]> wrote:

> On Tue, May 20, 2025 at 02:52:20PM -0400, Max Alekseyev wrote:
> > On a related note, I do not understand why I'm getting different results
> in
> > the following two examples (fist one looks like a bug):
> >
> > ? forpart(p=5,print(p),[0,5],[2,2])
> > Vecsmall([1, 4])
> > Vecsmall([2, 3])
>
> This is correct: the length is the number of nonzero entries.
>
> > ? forpart(p=5,print(p),[0,5],2)
> > Vecsmall([0, 5])
> > Vecsmall([1, 4])
> > Vecsmall([2, 3])
> >
> > Also, in the second case, shouldn't 2 be interpreted as nmax and
> partitions
> > with smaller parts (=1) be also generated?
>
>    By default, parts are positive and we remove zero entries unless amin
> <= 0, in which case we fix the size #X = nmax.
>
> Cheers,
> Bill.
>
>