@nm The subsindets call applies its action (its 3rd argument, here an anonymous procedure) to all subexpressions of type `+`. That is how that procedure gets called more than once.

Here is one way to see its actions, using option trace on that anonymous procedure.

Download subsindets_examp_explained.mw

You could also assign that inner procedure to a name and pass that as the 3rd argument to the outer subsindets. Then you could trace or debug it by name.

I can't seem to edit previous replies right now...

Even terser, but perhaps more obscure (which often I don't prefer),

R := expr -> subsindets(expr,`+`,
  ee -> local u := selectremove(type,ee,specfunc(f));
        u[2] + `if`(nops(u[1])=1,u[1],f(map(op,u[1])))):

Messing around can make it important to test all the corner cases.

This is a terser way to implement that idea,

R := expr -> subsindets(expr,`+`,
       proc(ee)
         local u := selectremove(type,ee,specfunc(f));
         u[2] + `if`(u[1]=0,0,f(subsindets(u[1],specfunc(f),op)));
       end proc):

In either implementation of this idea the key part is to separate all the function calls to f, as a first step. That allows you to avoid using the type `&+`, for which the dependency of the relative positions of its arguments was getting in your way.

@emendes Using Maple 2019 (or later),

[for t,v in eval(T) do `if`(v<>NULL,t,NULL); end do];

   [7,11,15,19,16,22,21,26,25,30,28,37,38,32,41,48,49,60,58,57]

[for t,v in eval(T) do if v<>NULL then t end end];

   [7,11,15,19,16,22,21,26,25,30,28,37,38,32,41,48,49,60,58,57]

@AE98 How tragic; I had it repaired with contents visible. It must have re-corrupted when I saved the file. In fact I now see that happening, even inducing some additional loss.

You could try the following attachment:
Assignment2_ac2.mw

@Earl I don't see how any of the other responses can directly search Maple's 2D Input, since that is stored in a .mw file as a base64 encoding of a .m (dotm) Maple encoding.

@Earl You have not yet clarified whether you need to deal with searching 2D Input.

Here is a rough illustration of the approach I mentioned.

Download file_search.mw

Of course, for plaintext .mpl or .txt files there are more efficient tools. But that's beside the point if you need to search through 2D Inputs.

note: Worksheet:-WorksheetToMapleText has some limitations, eg. it cannot handle embedded components (and possibly GUI Tables).

Please don't spawn a completely separate Question thread for a close followup query about this system.

You could add your followup details here, instead.

@Carl Love It did contain Danish characters.

The past four or five corruption reports have all been of the same flavour: an unclosed and empty Equation XML element. I'd never seen that particular variant prior to Maple 2022.2.

I have submitted them all as bug reports. In each report for this past group (five of this same variant, Dec2021-Feb2022) I have mentioned two distinct flaws:
  1) that it happens at all
  2) that the GUI is incapable of recovering as much as possible from the corrupted file, and sometimes very little of what is possible.

ps. These corrupted files do not have the "bad characters" for which Joe Riel made a repair script some years ago. That was a different flavor of problem -- which I do not recall seeing for several years.

Why do you think that your procedure's do-loop would ever end?

@Ahmed111 Sorry, I was looking at the wrong file -- you gave the same name to both your original and your revised attachments.

There are still mix-ups involving names like lambda__1 and lambda[1], H__34 and H[34], etc, with both present in some expressions, even in the 2D Input of assignments.

That happens in (at least) the 2D Input of the assignments to vvalue, A2snum, and valphipsi12.

@zenterix Someone (perhaps you, the OP) tagged this Question as "pointplot3d". I suggest that makes the Maple Help page for the pointplot3d command a decent starting place for help with programmatic options for that very command.

As I explained, the general programmatic 3D plotting options Help page is linked from there, in two ways.

The Portal pages are very much skewed towards mouse-pointer based context-menu approaches. It's almost as if they were written with the idea that programmatic approaches are undesirable.

I'm not suggesting that you've done anything wrong; I'm simply trying to assist you with some friendly advice in your future programmatic Maple endeavors.

ps. If I search for "light" in the Maple help browser then the mentioned plot3d,option topic page is the second hit down in the returned Search Results. Perhaps it ought to be the top hit.

@zenterix Here is a way to sort a Matrix by a given column. I'll use the second column.

Download sort_Matrix_by_column.mw

I would agree that it might be generally useful to have a Help page that contained common Matrix&Vector manipulations.

There are some manipulation commands (for Arrays, Matrices, & Vectors) in the ArrayTools package.

@zenterix It is not difficult to restructure the Matrix/Array of data. For the kind of regularly structured data you've described, the convertArrayToTable approach is very poor (relatively speaking). I was surprised that you returned to using it.

In fact I already showed you one way to do it directly and easily, using just seq and the angle-bracket constructor. The resulting Matrix can work directly with either surfdata or listdensityplot.

There are other ways to reassemble that data in a way that can be plotted directly, eg. the data can be rewritten as mxnx3, say, and "transposed. And there are still more ways, involving top-level easy commands. You seem so insistent that this quite basic task is difficult that's I think I should give up now.