Passing a multidimensional array to a function

[DanSandberg]

Hi All - I have an array of data, x,y,z coordinates for natoms as a function of time. The data is in mydata[ntime][natom][2] where the last index 0,1,2 is x,y,z respectively. I would like to pass the entire array to a function but the function won't allow me to pass an array with two variable dimensions. There MUST be a way around this. As a matter of fact I used to be a proficient C programmer but haven't been coding in C++ for about 5 years now and I just forget. Any help is greatly appreciated.

[eidge]

You could pass a pointer to the first element and a count of how many elements are in the array.

Something like this:

return_type my_function(int *arr, int n_elems);

int xyz[2][2][2];

my_function(&xyz[0][0][0], 2*2*2);

There are problably much better aprouchs for this, but I don't use arrays anyway.

You could emulate a 3d array using a 1d one. Or much simplier, use vectors.

[andywestken]

Might it be might be better if the function was?

return_type my_function(int *arr, int x_count, int y_count, int z_count);

Andy

P.S. But in C++ I'd prob. use a class which implementes a GetAt(x, y, z), etc. method.

[Computergeek01]

I might be misunderstanding your question so I'll provide a few ideas.

1) Simplify your function. Don't worry about passing the entire 3-D array at once, if you can pass each element of the array as seperate arguments and process them with in the function that way. This may not be a valid suggestion but you don't tell us what you function actually does.

2) For the sake of organization how about you don't use multi-demensional arrays and you instead stick to objects (struct's or class's) instead?

3) You could pass a pointer to the first element in each dimension of the array to your function as seperate arguments then assign them to a dynamically created local array within the function. This is ugly C-style programming and you have to be careful but it could work.

EDIT: How much code do you have? Can we see this function?

[DanSandberg]

Thanks for the suggestions guys. The function is designed to generate the dihedral angle defined by 4 atoms. Here are the functions:

void cross(double a[2], double b[2], double c[2])
{
c[0]=a[1]*b[2]-a[2]*b[1];
c[1]=a[2]*b[0]-a[0]*b[2];
c[2]=a[0]*b[1]-a[1]*b[0];
}

double dp(double a[2], double b[2])
{
return(a[0]*b[0]+a[1]*b[1]+a[2]*b[2]);
}

double Sdihedral(int a1, int a2, int a3, int a4, double mydat[][82][2], int mytime){
double v1[2], v2[2], v3[2];
double size;
double v2v1[2];
double cp23[2],cp12[2],dp1,dp2;
double retval;
int myi;
for (myi=0;myi<3;myi++){
v1[myi]=mydat[mytime][a2][myi]-mydat[mytime][a1][myi];
v2[myi]=mydat[mytime][a3][myi]-mydat[mytime][a2][myi];
v3[myi]=mydat[mytime][a4][myi]-mydat[mytime][a3][myi];
}
size=pow((v2[0]*v2[0]+v2[1]*v2[1]+v2[2]*v2[2]),0.5);
v2v1[0]=v1[0]*size;
v2v1[1]=v1[1]*size;
v2v1[2]=v1[2]*size;

cross(v2,v2,cp23);
cross(v1,v2,cp12);
dp1=dp(v2v1,cp23);
dp2=dp(cp12,cp23);

retval=57.2957795*atan2(dp1,dp2);
if (retval<0){
retval=retval+360;
}
if (retval>360){
retval=retval-360;
}
return (retval);
}

so for the array 'mydat' in the dihedral function I assume that 82 atoms are being read into the array. Sounds like there are much better ways to do this though. I'll have to look into classes, etc.

Thanks again.

[andywestken]

void cross(double a[2], double b[2], double c[2]);

looks wrong, given you're accessing elements with an index of 2.

You need double a[3], etc.

a[3] has three elements: a[0], a[1], a[2]

a[2] only two: a[0], a[1]

[andywestken]

Seeing your code, it does look like a class based approach would make a lot of sense. Starting with 3d vectors and matrices.

Or would you consider using a library? If so, I have been looking about in this kind of area for my own reasons, and some of the libraries I found (the first two seem to be the most popular):

IT++
http://sourceforge.net/apps/wordpress/itpp/

Armadillo
http://arma.sourceforge.net/

The Blitz++ Library / An array library for C++
http://www.oonumerics.org/

boost::ublas
http://www.boost.org/doc/libs/1_41_0/libs/numeric/ublas/doc/index.htm

(Also see this list: http://techlogbook.wordpress.com/c-numerical-libraries/)