Allen, I'm surprised the Pick trick isn't working. I've used that technique for ages to block the "empty Pick" error. When empty, "is empty" is 1, so "is not empty" is 0, so Route should not execute, trapping the error of trying to use a non-existent value downstream.

Meanwhile, try Mark("connections"). Then the 'data' is a copy of the connections array, so what you see for the Picked 'data' should be the several integer indices into the positions.

You can also Glyph or Autoglyph the field with glyph type set to "text" (or more fancy: "text font=area" or whatever other font you prefer). This will make a sea of text assuming you don't run out of memory trying to render it all (in a really big mesh). You have to zoom in to read it. Picking is therefore a good idea, or Sample or Reduce or Slab, or whatever it takes to show less of the overall data set at a time.

Obviously, the same trick works to make any number array into visible strings: data, connections, positions, etc.


On Thursday, Feb 5, 2004, at 16:58 America/New_York, Allen H. Nugent wrote:

Márcio,

At 09:38 AM 4/02/04 -0800, you wrote:
Hi, does anybody know how to show the numbers of the elements and/or the nodes in a finite element grid? I found an example in dx manuals for structured grids, regarding row- and column- majority numbering, but it's not clear to me how to do it for unstructured grids. I think I can't use the majority command for unstructured grids, can I?

I have been trying to fix missing and wrong triangles in a 3D surface construction by interactively obtaining the coords of their vertices. I looked at PickPlot.net and PlotLine.net, then added the following to my network:

{Pick} --- {Inquire ("Is not empty")} --- {<1>Route} --- {<2>Format} --- {Caption} --- {Collect} --- {Image}
     |---------------------------------> {<2>Route}

and the 1st argument of {Format} is "coords = %f". I think the {Inquire}/{Route} combination is supposed to prevent an error when there is no Pick component (the 1st time the network is executed) but it doesn't work the way I've done it. Otherwise, the only problem is that the output is the coords of the node -- not its array index. I get the array index by searching the data file for the nearest coords (DX picks the closest interpolated point on the surface, not the actual point), so it's a pretty tedious method.

Please let me know if you improve on it.


Allen H. Nugent
Graduate School of Biomedical Engineering
University of New South Wales
Sydney NSW 2052 Australia
Tel: +61 2 9385 3916 Fax: +61 2 9663 2108



Reply via email to