Complicated growable array

[breakerbyte]

Hi codeguru users!

I am basically looking to make a growable array thats basically as follows:


{ [ ((PT1.X,PT1.Y,PT2.X,PT3.X),(PT1.X,PT1.Y,PT2.X,PT3.X),(PT1.X,PT1.Y,PT2.X,PT3.X) )] }

the arrays size grows a lot.
(pt1.x/y are points i am using via opencv)
the larger picture ie the outside "()" represent an object
the second set of "()" represent the iterations
and the individual elements of the iterations (the points) are points

if you help me i will love you forever

EDIT: the array increases by each of the above (marked with "[]")

[laserlight]

Have you considered using std::vector?

[Lindley]

struct Point
{
    float x, y;
};

std::vector< std::pair<Point, Point> > growablearray;

[breakerbyte]

yeah, but apparently Its hard for me to find/understand exactly what is happening.

would it be:

vector<vector<vector<double>>> detection_points;

not sure how I would set the sizes or anything.

or how to use the push_back function correctly for it

[breakerbyte]

struct Point
{
    float x, y;
};

std::vector< std::pair<Point, Point> > growablearray;

holy wow! going to try this method out.

[breakerbyte]

not working :/

could I combine both methods and store each "Point"
lets say that (Point(iteration 0),Point(iteration 1),Point(iteration 2) is the equivalence of object

(object,object,object)

[laserlight]

Okay, before talking about an array, what is the class definition of the objects that you plan to store?

[breakerbyte]

they are either doubles or CvRect if they would work

[laserlight]

they are either doubles or CvRect if they would work

Well, I don't really understand you, actually. I felt that Lindley made a good stab at an answer, but it seems like there is a communication breakdown somewhere. Let's do this step by step.

Firstly, instead of dumping this:

{ [ ((PT1.X,PT1.Y,PT2.X,PT3.X),(PT1.X,PT1.Y,PT2.X,PT3.X),(PT1.X,PT1.Y,PT2.X,PT3.X) )] }

Do a breakdown of it. You say that "pt1.x/y are points i am using via opencv". What do you mean by "the second set of "()" represent the iterations"? Perhaps you want to create a class here, and then objects of this class would then be members of another class, and then the dynamic array would contain objects of this second class?

Also, you should state how does Lindley's example not work.

[breakerbyte]

Sorry bout the confusion.

<vector<vector<CvRect>>foo;
Changing everything to CvRect.

need foo to grow 5 rows and 1 column on demand. thats the real issue

[Paul McKenzie]

Sorry bout the confusion.

<vector<vector<CvRect>>foo;
Changing everything to CvRect.

need foo to grow 5 rows and 1 column on demand. thats the real issue

It's easier if you use typedef.

#include <vector>
typedef std::vector<CvRect> CvRectArray;
typedef std::vector<CvRectArray> CvRectArray2D;
//...
int main()
{
    // Assuming that CVRect is default constructible.
    CVRectArray2D foo(5, CVRectArray(10));  A 5x10 array of CVRect

    // resize to 10x20

    // first, resize the existing rows
    for (CVRectArray2D::size_type i = 0; i < foo.size(); ++i )
        foo[i].resize(20);

    // now add on new rows 
    foo.resize(10, CVRectArray(20));
}

Regards,

Paul McKenzie