Assigning an array of arrays in C/C++

Question

Hi I'm not a "native" C/C++ programmer. I'm able to write some basic things and do a coding that is common to other languages. But I have this problem and I don't know even how to ask different then explaining it (thus google searching won't fit me).

I've got in my code

typedef float point3[3];

And now I'm initilizing a bunch of points in 3D (x,y,z) by this:

point3 cpoint = {computeX(u,v),computeY(u,v)-5,computeZ(u,v)};

What functions does and values of u and v are irrevelent to my question (but I can provide code if asked).

Now I want to declare an array (one dimensional) of point3. So when I call array[0] I will get a point3 type variable. How I can do that?

EDIT:

I provided insufficient information. My bad. I need to retain: typedef float point3[3]; because I'm using OpenGL with GLUT and using this function glVertex3fv(cpoint); where cpoint is point3 type. So I'm pretty sure I can't use struct.

You can provide a cast or conversion from a struct to a float* which you then pass to the function... Also, there is no such thing as "C/C++". They are different languages, and while it's possible to write things that would compile as either, you are shooting yourself in the foot by trying to do so. Pick one and stick with it. C++ provides many, many useful tools that you are denying yourself. — Karl Knechtel
– Karl Knechtel, Commented Dec 13, 2010 at 10:42

Andreas Wong · Accepted Answer · 2010-12-13 09:20:41Z

2

point3 cpoint[2] = {
   {computeX(a, b), computeY(a,b)-5, computeZ(a,b)},
   {computeX(c, d), computeY(c,d)-5, computeZ(c,d)}
}

answered Dec 13, 2010 at 9:20

Andreas Wong

60.7k19 gold badges112 silver badges123 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Ertai Over a year ago

If this code is something like this: point3 test[2]; point3 cpoint = {computeX(u,v),computeY(u,v)-5,computeZ(u,v)}; point3 test[0] = cpoint;? If yes that isn't working and that's my case.

Jens Gustedt Over a year ago

@Ertai: You must distinguish between initialization and assignment. Initialization of arrays is possible and should be done, assignment as a whole not. I recently wrote things up arround that theme: gustedt.wordpress.com/2010/12/12/assignment-and-initialization

Ertai Over a year ago

Ok, so basicly my problem is that I don't have the right assignment operator. I think I could use an assigment If I get my pointers right... problem is I don't know how to do this, and learning about pointers in C++ is a bit of hard thing to do (and a bit of overkill for simple assigment that I have) just to assign a value (array) to an array of values.

Component 10 · Accepted Answer · 2010-12-13 09:29:41Z

2

Assuming C++, the simplest way would be to create a class to encapsulate your point structure and then create a vector of that type:

struct Point
{
    Point(float x, float y, float z) : x_(x), y_(y), z_(z)
    float x_;
    float y_;
    float z_;
};

typedef std::vector<Point> Points;

Then you can create a collection for storing the points:

Points points;

then later in the code, once populated you can access using:

Point a = points[ 3 ]; // for example

answered Dec 13, 2010 at 9:29

Component 10

10.6k7 gold badges52 silver badges69 bronze badges

2 Comments

Ertai Over a year ago

I provided insufficient information. My bad. I need to retain: typedef float point3[3]; because I'm using OpenGL with GLUT and using this function glVertex3fv(cpoint); where cpoint is point3 type. So I'm pretty sure I can't use struct.

Mephane Over a year ago

@Ertai: I've added an answer which expands upon this solution with automatic conversion.

Mihran Hovsepyan · Accepted Answer · 2010-12-13 09:22:19Z

1

At first declare you own point3 like this

struct point3
{
 float x;
 float y;
 float z;
};

then declare your array

point3 points[2] = {{calculateX(a1, b1), calculateY(a1, b1), calculateZ(a1, b1)},{calculateX(a2, b2), calculateY(a2, b2), calculateZ(a2, b2)}};

answered Dec 13, 2010 at 9:22

Mihran Hovsepyan

11.2k15 gold badges69 silver badges113 bronze badges

2 Comments

Ertai Over a year ago

I provided insufficient information. My bad. I need to retain: typedef float point3[3]; because I'm using OpenGL with GLUT and using this function glVertex3fv(cpoint); where cpoint is point3 type. So I'm pretty sure I can't use struct.

Pawel Zubrycki Over a year ago

Have you tried? Not so sure if you really can't use pointer to struct.

Mephane · Accepted Answer · 2010-12-13 10:26:49Z

I'd like to expand Robin Welch's answer in a way which allows to stay compatible with the point3 typedef:

typedef float point3[3]; 


struct Point
{
    Point(float x, float y, float z)
    : x_(values_[0]), y_(values_[1]), z_(values_[2])
    {
        values_[0] = x;
        values_[1] = y;
        values_[2] = z;
    }

    float& x_;
    float& y_;
    float& z_;
    point3 values_;

    operator const point3&() const { return values_; }
    operator point3&() { return values_; }
};

Now you can just use this simple struct, and pass it to OpenGL functions just as if it weren't this struct at all. It contains the original point3 array for OpenGL, yet convenient access and a constructor for yourself.

pmg · Accepted Answer · 2010-12-13 09:29:35Z

0

In C, I wouldn't initialize the array when declaring it unless it's a small array with constant values.

typedef float point3[3]; /* point is a type */
point3 *pp;              /* pp points to objects of type point3 */
point3 ap[NELEMS];       /* each (of NELEMS) element of ap is of type point3 */

pp = malloc(NELEMS * sizeof *pp);
if (pp) {
    /* use pp */
    pp[0][0] = computeX(); pp[0][1] = computeY(); pp[0][2] = computeZ();
    (*(pp+42))[0] = computeX(); (*(pp+42))[1] = computeY(); (*(pp+42))[2] = computeZ();
    free(pp);
}

ap[0][0] = computeX(); ap[0][1] = computeY(); ap[0][2] = computeZ();

answered Dec 13, 2010 at 9:29

pmg

109k14 gold badges132 silver badges203 bronze badges

7 Comments

Ertai Over a year ago

I'm assuming that (*(pp+42)) is like array[42]?

pmg Over a year ago

Yes, same thing. You can use the array subscript syntax instead. I just wanted to show the pointer syntax.

Ertai Over a year ago

This line pp = malloc(NELEMS * sizeof pp); in my code is like this pp = malloc(NN * sizeof pp); where N is const int N = 20; And I've got "invalid conversion from 'void' to 'float (*)[3]'"

JeremyP Over a year ago

@Ertai: It sounds like you are using C++ not C. You need to do an explicit cast with C++ on malloc. In fact, you should be using new instead if you are using C++.

Ertai Over a year ago

I've modified that line to do casting: pp = (point3*)malloc(NN * sizeof *pp); and it compiles... but something is very wrong as I don't get my points right. Could you provide me how a new syntax would be equal to the code above? I've got this: pp = new point3[NN]; and still my points are all wrong (it compiles thou).

|

Collectives™ on Stack Overflow

Assigning an array of arrays in C/C++

5 Answers 5

3 Comments

2 Comments

2 Comments

Comments

7 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

5 Answers 5

3 Comments

2 Comments

2 Comments

Comments

7 Comments

Your Answer

Sign up or log in

Post as a guest

Related