Nice solution, but I would not apply it in this case. Mostly for maintenance reason (and personal preferences), I prefer to see the MAX_GROUP where it is used, ie in the array declaration. Especially since those table does not evolve that frequently. But if I had code that would take a long time to compile, or working with multiple coder that would frequently modify the table I would definitely use it.

And since we are in the array declaration problems, how do you do with multiple diemsion array. I think that you are forced to declare a size for the third column as soon as you have a 3 dimensional array. (Or at least the compiler would not let me go without it)

If you know the size of the array as a constant, then there isn't much point in using sizeof.

Sizeof is useful in cases like what Tyche showed, where you know the array is a fixed size, but you have not defined that size as a constant.

There's no way for the compiler to calculate the size of the 2nd and 3rd dimensions.
So you can only leave off the first dimension.
This is legal:
int foo[][3][4] = {1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24};
int bar[][3] = {1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24};
Because the compiler can calculate that the first dimension of foo is 2 and bar is 8.

Now one would think a C compiler could calculate the dimensions if you include the correct scalar initializers…
int foo [][][] = {{{1,2,3,4},{5,6,7,8},{9,10,11,12}},{{13,14,15,16},{17,18,19,20},{21,22,23,24}}};
…but gcc and vc++ do not. Maybe because they don't force the use scalar initializer notation, only validate it.

Tyche's example showed exactly what I meant.