Re: 0-1 sparse representations?

[E R]

Thanks for all of the replies.

In my case the matrix may not be square and may have duplicate rows,
so it may not be a permutation matrix. The boolean sparse matrices
look promising.

As a suggestion for Octave - perhaps it would be a good idea to add
standard library functions which creates these classes of 0-1 sparse
matrices, i.e. permutation matricies, matrices with only one 1 per
row/column, etc.

For example, the current advice for creating a permutation matrix is
to use the expression eye(n)(list,:)  (c.f.
http://www.gnu.org/software/octave/doc/interpreter/Creating-Permutation-Matrices.html
) However, if some sort of `permutation_matrix()` function were a
built-in, users wouldn't have to remember this construction. Moreover,
it would give Octave more freedom to change the implementation of
matrices in the future.


2011/11/20 Jordi Gutiérrez Hermoso <address@hidden>:
> 2011/11/20 c. <address@hidden>:
>>
>> On 19 Nov 2011, at 22:33, Jordi Gutiérrez Hermoso wrote:
>>
>>> No. Sparse matrices must be of float types.
>>
>> Do they?
>
> Oops. Well, there are bool sparse matrices. Aren't I embarrassed.
>
> - Jordi G. H.
> _______________________________________________
> Help-octave mailing list
> address@hidden
> https://mailman.cae.wisc.edu/listinfo/help-octave
>