I've got XML data with multiple elements that I'm trying to sort so I can show them in a particular order. I don't know what my upper bound on an array could be, but I can do xmlcount(<element>) to find that.

I define a global variable: define(&ServiceOrder,arrayinteger,15)

Then a global function @OrderedArray():arrayinteger
Within the function,
define(&temparray,arrayinteger,xmlcount(<element>)

and, after sorting,

&result:=&temparray
endfunction()

Then call the function with this line: &ServiceOrder:=@OrderedArray

What Does &ServiceOrder look like now? Does it have xmlcount() elements, or the original 15?

Toby,

In Design time, if you do this, it appears as though it resized the array, and it now has a different amount of items. However, when trying to preview, it just won't compile.

PlanetPress Press does not support dynamic arrays. Arrays have to be of a predetermines size and cannot be dynamically resized in a PressTalk code.

Regards,
Raphaël Lalonde Lefebvre

That's so not cool. So, can I pass the arraystring out of the function, and if so, I need to make that arraystring the same size as the global variable, right?

Toby,

To avoid those kinds of issues altogether, you should pass your array as a parameter to your Sort function:

That way, your OrderedArray function doesn't have to declare another array (because it already receives a copy of the original one as its parameter). The function body would therefore look something like:

I like that, Phillipe. It has it's limitations, but it is simple.

I tried it out and am gratified to see that as long as the &temparray is the same size as the global array, everything works. The remaining element's values are already defaulted to 0, so I don't have to do that manually. Now I just need to code the loop to stop when it sees a 0 value. Or, more effectively, at the end of xmlcount() iterations.

Ahhh... now I think I understand your issue: global variables cannot be initialized with dynamic values because they get initialized before the data is read. So once you've picked a length for your Global Array, you're stuck with it. Unfortunately, there's no way around this restriction.
Either you don't use Global Arrays for this task, or you do exactly what you do now: pick a default size that's large enough to accomodate any other array.

Yes, you've got it. That's a buzzkill. I remember way back when when global variables could be set by the data....

Thus far, it all seems to be working. I'm testing the existing data (an accurately representative sample) to see just what the maximum size needs to be for speed's sake.

I prefer to recall even further, when there were just no global variables...