Multidimensional arrays

[dodzy]

Hi,

The code which I have posted is messy and may not make much sense to you and I apologise for this. It is supposed to be performing an obscure quantum - mechanical calculation, hence, the reason for all the messy factorials etc.

However, I have declared a number of arrays globally and among them are two two - dimensional arrays: wigner6j[20][20] and racah[20][20]. The number of the elements is not particularly important and is set randomly at 20 which will allow larger calculations to be made. The arrays are subsequently filled within the function Racah(), which follows the main() body, using 2 large for loops which encompass the complete function. The for loops are performed over integers from 0 to a value specified by the user, thus, giving the arrays the subscripts [m][n].

Upon compiling (with g++), I get the error messages: "invalid types 'float[int]' for array subscript" which are each referring to the appropriate lines where the two - dimensional arrays are filled at the end of the function Racah().

My question is simple: why can't I have a float array with integer subscripts? Have I understood the compiler's message? Can anyone help me with this problem?

Thank you for taking the time to read this.

dodzy

[quzah]

	if (ans1 = 'y'){

= for assignment.
== for equality.

		for (i = 1; i <= states; i++){
...
				scanf("%d %f", &J[i], &E[i]);

Arrays start at 0 and go through "size - 1".

				scanf("%d %f", &J[i], &E[i]);

All of those are what your errors / warnings are talking about.

That'll get you started.


Quzah.

[dodzy]

Thanks for your suggestions! I don't know how I could have been so sloppy.

Unfortunately, your suggestions have not solved the problem. The same error messages remain.

dodzy

[quzah]

Try posting your updated code. My psychic powers are resting.


Quzah.

[dodzy]

I already have. I just replaced the old code with the updated code.

You will find that my original post has been updated.

Cheers,

dodzy

[Salem]

Well here
calc[22] = J[m] + j_p + j_n + 1;
and here
intercalc[6] = calc[22] - i;
you're running off the ends of the arrays.

arr[N] has indices 0 to N-1, not 0 to N or 1 to N

[dwks]

That probably wouldn't cause an error message, though . . .

This computer can't open the attached file for some reason, but I'd guess that you're using a floating-point number for an array index. Something like this:

array[1.2] = 0;

[Salem]

One of the hazards of using globals I suppose.

You could try a different naming convention, say prefix all globals with g_ it you insist on using them. I notice all your functions have no parameters.

Anyway

float Wigner6j[20][20], racah[20][20];

And in the function

	float mincalc, wigner6j[20], Wigner6j, Racah, phase1, phase2;

That's right, the local variable with the same name isn't an array, so it barfs on the attempt to subscript it.


For your evening supper, try this

$ gcc -Wshadow -W -Wall -ansi -pedantic -O2 new.c
new.c:19: warning: return type defaults to `int'
new.c: In function `Racah':
new.c:118: warning: declaration of 'Wigner6j' shadows a global declaration
new.c:15: warning: shadowed declaration is here
new.c:118: warning: declaration of 'Racah' shadows a global declaration
new.c:111: warning: shadowed declaration is here
new.c:249: error: subscripted value is neither array nor pointer
new.c:262: error: subscripted value is neither array nor pointer
new.c:118: warning: unused variable `Racah'

That is, once you've turned all the // comments into proper C comments.

Having a local variable the same name as your function would cause an even more bizarre error message if you'd tried to call the function recursively.