There is a bug affecting 2d interpolation in dx 4.1.0 ... is this what
you are using?
see
http://opendx.watson.ibm.com/cgi-bin/dx/cvsweb/dx/src/exec/libdx/trisRI2D.c
It is fixed in cvs (you can make the change to just this file and
recompile if you wish).
Pete
"Mr. Roman Putanowicz" wrote:
> Hi,
>
> I was trying some more combinations with the script and data I have
> posted
> on the list (12.02.01, "Strange Map behaviour").
> Apparently when I add the Connect command the script produces correct
> results.
> So it brings the question which module fails or where I make mistake?
> Any ideas?
>
> Regards
> Roman Putanowicz
>
> PS. I include the script with Connect.
> (Original grid was 100x100 and it takes a lot ot time to triangulate
> so you may generate smaller grid, eg. 10x10. In that case
> set the point to [0.7, 0.5] to see the wrong results without Connect
> and correct one with Connect)
>
> data = Import ("header.general");
> // Create line parallel to 'y' axis
> // set point to [0.96 0.5] to see the correct profile
> // set point to [0.97 0.5] to see the incorrect profile
> line = Grid ([0.97 0.5],"line", [0 1], {30});
> // adding Connect brings again the correct results even for 0.97
> data = Connect (data, "triangulation", [0,0,1]);
> data = Map (line, data);
> data = Mark (data, "positions");
> data = Compute ("$0.y", data);
> data = Unmark (data, "positions");
> data = Color(data, "medium turquoise");
> data = Options (data, "mark", "triangle", "mark every", 3,
> "label","fluid speed");
> image = Plot(data, {"y" "fluid speed"});
> camera = AutoCamera (image);
> image = Render (image, camera);
> Display (image);
