[Octave-bug-tracker] [bug #60873] Feature request: combine system and dlmread if viable

[Philip Nienhuis]

Follow-up Comment #4, bug #60873 (project octave):

(@OP, gathered from your quite terse description)
> ... each solution would have tens of thousands of variables being listed...
> The length of each solution can be anywhere from 8 to 300 elements, all
separated by spaces, one solution per line. Each element is a positive integer
being the index of the true variable in question.
I get it that you do some preselection based on the solver's output (or the
solver does it) - is that meant to simplify feeding it to Octave? If so, I'm
wondering why you would do that in the first place.
Neither dlmread nor textscan nor Octave are put off by tens of thousands of
values. Octave is ideally suited for digging around in large datasets, much
easier than doing it using file I/O. Just let Octave sort it out for you.
Assuming the total nr of variables is the same for each run of the solver, why
not also add e.g., zeros for the "false" variables. Then textscan, or dlmread,
can read it all into a rectangular array from which the true variables can
easily be extracted.
If not, Markus' suggestion looks quite suitable to me.
But OK I very well might misunderstand what you're trying to accomplish.

Anyway, IMO this is a typical "x-y problem" that is more amenable for the help
forum than a bug report, let alone a feature request.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?60873>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/