[c++] How do I declare a 2d array in C++ using new?

How do i declare a 2d array using new?

Like, for a "normal" array I would:

int* ary = new int[Size]

but

int** ary = new int[sizeY][sizeX]

a) doesn't work/compile and b) doesn't accomplish what:

int ary[sizeY][sizeX] 

does.

If your project is CLI (Common Language Runtime Support), then:

You can use the array class, not that one you get when you write:

#include <array>
using namespace std;

In other words, not the unmanaged array class you get when using the std namespace and when including the array header, not the unmanaged array class defined in the std namespace and in the array header, but the managed class array of the CLI.

with this class, you can create an array of any rank you want.

The following code below creates new two dimensional array of 2 rows and 3 columns and of type int, and I name it "arr":

array<int, 2>^ arr = gcnew array<int, 2>(2, 3);

Now you can access elements in the array, by name it and write only one squared parentheses [], and inside them, add the row and column, and separate them with the comma ,.

The following code below access an element in 2nd row and 1st column of the array I already created in previous code above:

arr[0, 1]

writing only this line is to read the value in that cell, i.e. get the value in this cell, but if you add the equal = sign, you are about to write the value in that cell, i.e. set the value in this cell. You also can use the +=, -=, *= and /= operators of course, for numbers only (int, float, double, __int16, __int32, __int64 and etc), but sure you know it already.

If your project is not CLI, then you can use the unmanaged array class of the std namespace, if you #include <array>, of course, but the problem is that this array class is different than the CLI array. Create array of this type is same like the CLI, except that you will have to remove the ^ sign and the gcnew keyword. But unfortunately the second int parameter in the <> parentheses specifies the length (i.e. size) of the array, not its rank!

There is no way to specify rank in this kind of array, rank is CLI array's feature only..

std array behaves like normal array in c++, that you define with pointer, for example int* and then: new int[size], or without pointer: int arr[size], but unlike the normal array of the c++, std array provides functions that you can use with the elements of the array, like fill, begin, end, size, and etc, but normal array provides nothing.

But still std array are one dimensional array, like the normal c++ arrays. But thanks to the solutions that the other guys suggest about how you can make the normal c++ one dimensional array to two dimensional array, we can adapt the same ideas to std array, e.g. according to Mehrdad Afshari's idea, we can write the following code:

array<array<int, 3>, 2> array2d = array<array<int, 3>, 2>();

This line of code creates a "jugged array", which is an one dimensional array that each of its cells is or points to another one dimensional array.

If all one dimensional arrays in one dimensional array are equal in their length/size, then you can treat the array2d variable as a real two dimensional array, plus you can use the special methods to treat rows or columns, depends on how you view it in mind, in the 2D array, that std array supports.

You also can use Kevin Loney's solution:

int *ary = new int[sizeX*sizeY];

// ary[i][j] is then rewritten as
ary[i*sizeY+j]

but if you use std array, the code must look different:

array<int, sizeX*sizeY> ary = array<int, sizeX*sizeY>();
ary.at(i*sizeY+j);

And still have the unique functions of the std array.

Note that you still can access the elements of the std array using the [] parentheses, and you don't have to call the at function. You also can define and assign new int variable that will calculate and keep the total number of elements in the std array, and use its value, instead of repeating sizeX*sizeY

You can define your own two dimensional array generic class, and define the constructor of the two dimensional array class to receive two integers to specify the number of rows and columns in the new two dimensional array, and define get function that receive two parameters of integer that access an element in the two dimensional array and returns its value, and set function that receives three parameters, that the two first are integers that specify the row and column in the two dimensional array, and the third parameter is the new value of the element. Its type depends on the type you chose in the generic class.

You will be able to implement all this by using either the normal c++ array (pointers or without) or the std array and use one of the ideas that other people suggested, and make it easy to use like the cli array, or like the two dimensional array that you can define, assign and use in C#.

int** ary = new int[sizeY][sizeX]

should be:

int **ary = new int*[sizeY];
for(int i = 0; i < sizeY; ++i) {
    ary[i] = new int[sizeX];
}

and then clean up would be:

for(int i = 0; i < sizeY; ++i) {
    delete [] ary[i];
}
delete [] ary;

EDIT: as Dietrich Epp pointed out in the comments this is not exactly a light weight solution. An alternative approach would be to use one large block of memory:

int *ary = new int[sizeX*sizeY];

// ary[i][j] is then rewritten as
ary[i*sizeY+j]

declaring 2D array dynamically:

    #include<iostream>
    using namespace std;
    int main()
    {
        int x = 3, y = 3;

        int **ptr = new int *[x];

        for(int i = 0; i<y; i++)
        {
            ptr[i] = new int[y];
        }
        srand(time(0));

        for(int j = 0; j<x; j++)
        {
            for(int k = 0; k<y; k++)
            {
                int a = rand()%10;
                ptr[j][k] = a;
                cout<<ptr[j][k]<<" ";
            }
            cout<<endl;
        }
    }

Now in the above code we took a double pointer and assigned it a dynamic memory and gave a value of the columns. Here the memory allocated is only for the columns, now for the rows we just need a for loop and assign the value for every row a dynamic memory. Now we can use the pointer just the way we use a 2D array. In the above example we then assigned random numbers to our 2D array(pointer).Its all about DMA of 2D array.

If you want an 2d array of integers, which elements are allocated sequentially in memory, you must declare it like

int (*intPtr)[n] = new int[x][n]

where instead of x you can write any dimension, but n must be the same in two places. Example

int (*intPtr)[8] = new int[75][8];
intPtr[5][5] = 6;
cout<<intPtr[0][45]<<endl;

must print 6.

I presume from your static array example that you want a rectangular array, and not a jagged one. You can use the following:

int *ary = new int[sizeX * sizeY];

Then you can access elements as:

ary[y*sizeX + x]

Don't forget to use delete[] on ary.

The purpose of this answer is not to add anything new that the others don't already cover, but to extend @Kevin Loney's answer.

You could use the lightweight declaration:

int *ary = new int[SizeX*SizeY]

and access syntax will be:

ary[i*SizeY+j]     // ary[i][j]

but this is cumbersome for most, and can lead to confusion. So, you can define a macro as follows:

#define ary(i, j)   ary[(i)*SizeY + (j)]

Now you can access the array using the very similar syntax ary(i, j) // means ary[i][j]. This has the advantages of being simple and beautiful, and at the same time, using expressions in place of the indices is also simpler and less confusing.

To access, say, ary[2+5][3+8], you can write ary(2+5, 3+8) instead of the complex-looking ary[(2+5)*SizeY + (3+8)] i.e. it saves parentheses and helps readability.

Caveats:

• Although the syntax is very similar, it is NOT the same.
• In case you pass the array to other functions, SizeY has to be passed with the same name (or instead be declared as a global variable).

Or, if you need to use the array in multiple functions, then you could add SizeY also as another parameter in the macro definition like so:

#define ary(i, j, SizeY)  ary[(i)*(SizeY)+(j)]

You get the idea. Of course, this becomes too long to be useful, but it can still prevent the confusion of + and *.

This is not recommended definitely, and it will be condemned as bad practice by most experienced users, but I couldn't resist sharing it because of its elegance.

Edit:
If you want a portable solution that works for any number of arrays, you can use this syntax:

#define access(ar, i, j, SizeY) ar[(i)*(SizeY)+(j)]

and then you can pass on any array to the call, with any size using the access syntax:

access(ary, i, j, SizeY)      // ary[i][j]

P.S.: I've tested these, and the same syntax works (as both an lvalue and an rvalue) on g++14 and g++11 compilers.

A 2D array is basically a 1D array of pointers, where every pointer is pointing to a 1D array, which will hold the actual data.

Here N is row and M is column.

dynamic allocation

int** ary = new int*[N];
  for(int i = 0; i < N; i++)
      ary[i] = new int[M];

fill

for(int i = 0; i < N; i++)
    for(int j = 0; j < M; j++)
      ary[i][j] = i;

print

for(int i = 0; i < N; i++)
    for(int j = 0; j < M; j++)
      std::cout << ary[i][j] << "\n";

free

for(int i = 0; i < N; i++)
    delete [] ary[i];
delete [] ary;

Why not use STL:vector? So easy, and you don't need to delete the vector.

int rows = 100;
int cols = 200;
vector< vector<int> > f(rows, vector<int>(cols));
f[rows - 1][cols - 1] = 0; // use it like arrays

You can also initialise the 'arrays', just give it a default value

const int DEFAULT = 1234;
vector< vector<int> > f(rows, vector<int>(cols, DEFAULT));

Source: How to Create 2, 3 (or Multi) Dimensional Arrays in C/C++?

I used this not elegant but FAST,EASY and WORKING system. I do not see why can not work because the only way for the system to allow create a big size array and access parts is without cutting it in parts:

#define DIM 3
#define WORMS 50000 //gusanos

void halla_centros_V000(double CENW[][DIM])
{
    CENW[i][j]=...
    ...
}


int main()
{
    double *CENW_MEM=new double[WORMS*DIM];
    double (*CENW)[DIM];
    CENW=(double (*)[3]) &CENW_MEM[0];
    halla_centros_V000(CENW);
    delete[] CENW_MEM;
}

Although this popular answer will give you your desired indexing syntax, it is doubly inefficient: big and slow both in space and time. There's a better way.

Why That Answer is Big and Slow

The proposed solution is to create a dynamic array of pointers, then initializing each pointer to its own, independent dynamic array. The advantage of this approach is that it gives you the indexing syntax you're used to, so if you want to find the value of the matrix at position x,y, you say:

int val = matrix[ x ][ y ];

This works because matrix[x] returns a pointer to an array, which is then indexed with [y]. Breaking it down:

int* row = matrix[ x ];
int  val = row[ y ];

Convenient, yes? We like our [ x ][ y ] syntax.

But the solution has a big disadvantage, which is that it is both fat and slow.

Why?

The reason that it's both fat and slow is actually the same. Each "row" in the matrix is a separately allocated dynamic array. Making a heap allocation is expensive both in time and space. The allocator takes time to make the allocation, sometimes running O(n) algorithms to do it. And the allocator "pads" each of your row arrays with extra bytes for bookkeeping and alignment. That extra space costs...well...extra space. The deallocator will also take extra time when you go to deallocate the matrix, painstakingly free-ing up each individual row allocation. Gets me in a sweat just thinking about it.

There's another reason it's slow. These separate allocations tend to live in discontinuous parts of memory. One row may be at address 1,000, another at address 100,000—you get the idea. This means that when you're traversing the matrix, you're leaping through memory like a wild person. This tends to result in cache misses that vastly slow down your processing time.

So, if you absolute must have your cute [x][y] indexing syntax, use that solution. If you want quickness and smallness (and if you don't care about those, why are you working in C++?), you need a different solution.

A Different Solution

The better solution is to allocate your whole matrix as a single dynamic array, then use (slightly) clever indexing math of your own to access cells. The indexing math is only very slightly clever; nah, it's not clever at all: it's obvious.

class Matrix
{
    ...
    size_t index( int x, int y ) const { return x + m_width * y; }
};

Given this index() function (which I'm imagining is a member of a class because it needs to know the m_width of your matrix), you can access cells within your matrix array. The matrix array is allocated like this:

array = new int[ width * height ];

So the equivalent of this in the slow, fat solution:

array[ x ][ y ]

...is this in the quick, small solution:

array[ index( x, y )]

Sad, I know. But you'll get used to it. And your CPU will thank you.

How to allocate a contiguous multidimensional array in GNU C++? There's a GNU extension that allows the "standard" syntax to work.

It seems the problem come from operator new []. Make sure you use operator new instead :

double (* in)[n][n] = new (double[m][n][n]);  // GNU extension

And that's all : you get a C-compatible multidimensional array...

Try doing this:

int **ary = new int* [sizeY];
for (int i = 0; i < sizeY; i++)
    ary[i] = new int[sizeX];

In C++11 it is possible:

auto array = new double[M][N]; 

This way, the memory is not initialized. To initialize it do this instead:

auto array = new double[M][N]();

Sample program (compile with "g++ -std=c++11"):

#include <iostream>
#include <utility>
#include <type_traits>
#include <typeinfo>
#include <cxxabi.h>
using namespace std;

int main()
{
    const auto M = 2;
    const auto N = 2;

    // allocate (no initializatoin)
    auto array = new double[M][N];

    // pollute the memory
    array[0][0] = 2;
    array[1][0] = 3;
    array[0][1] = 4;
    array[1][1] = 5;

    // re-allocate, probably will fetch the same memory block (not portable)
    delete[] array;
    array = new double[M][N];

    // show that memory is not initialized
    for(int r = 0; r < M; r++)
    {
        for(int c = 0; c < N; c++)
            cout << array[r][c] << " ";
        cout << endl;
    }
    cout << endl;

    delete[] array;

    // the proper way to zero-initialize the array
    array = new double[M][N]();

    // show the memory is initialized
    for(int r = 0; r < M; r++)
    {
        for(int c = 0; c < N; c++)
            cout << array[r][c] << " ";
        cout << endl;
    }

    int info;
    cout << abi::__cxa_demangle(typeid(array).name(),0,0,&info) << endl;

    return 0;
}

Output:

2 4 
3 5 

0 0 
0 0 
double (*) [2]

I'm using this when creating dynamic array. If you have a class or a struct. And this works. Example:

struct Sprite {
    int x;
};

int main () {
   int num = 50;
   Sprite **spritearray;//a pointer to a pointer to an object from the Sprite class
   spritearray = new Sprite *[num];
   for (int n = 0; n < num; n++) {
       spritearray[n] = new Sprite;
       spritearray->x = n * 3;
  }

   //delete from random position
    for (int n = 0; n < num; n++) {
        if (spritearray[n]->x < 0) {
      delete spritearray[n];
      spritearray[n] = NULL;
        }
    }

   //delete the array
    for (int n = 0; n < num; n++) {
      if (spritearray[n] != NULL){
         delete spritearray[n];
         spritearray[n] = NULL;
      }
    }
    delete []spritearray;
    spritearray = NULL;

   return 0;
  } 

I have left you with a solution which works the best for me, in certain cases. Especially if one knows [the size of?] one dimension of the array. Very useful for an array of chars, for instance if we need an array of varying size of arrays of char[20].

int  size = 1492;
char (*array)[20];

array = new char[size][20];
...
strcpy(array[5], "hola!");
...
delete [] array;

The key is the parentheses in the array declaration.

This problem has bothered me for 15 years, and all the solutions supplied weren't satisfactory for me. How do you create a dynamic multidimensional array contiguously in memory? Today I finally found the answer. Using the following code, you can do just that:

#include <iostream>

int main(int argc, char** argv)
{
    if (argc != 3)
    {
        std::cerr << "You have to specify the two array dimensions" << std::endl;
        return -1;
    }

    int sizeX, sizeY;

    sizeX = std::stoi(argv[1]);
    sizeY = std::stoi(argv[2]);

    if (sizeX <= 0)
    {
        std::cerr << "Invalid dimension x" << std::endl;
        return -1;
    }
    if (sizeY <= 0)
    {
        std::cerr << "Invalid dimension y" << std::endl;
        return -1;
    }

    /******** Create a two dimensional dynamic array in continuous memory ******
     *
     * - Define the pointer holding the array
     * - Allocate memory for the array (linear)
     * - Allocate memory for the pointers inside the array
     * - Assign the pointers inside the array the corresponding addresses
     *   in the linear array
     **************************************************************************/

    // The resulting array
    unsigned int** array2d;

    // Linear memory allocation
    unsigned int* temp = new unsigned int[sizeX * sizeY];

    // These are the important steps:
    // Allocate the pointers inside the array,
    // which will be used to index the linear memory
    array2d = new unsigned int*[sizeY];

    // Let the pointers inside the array point to the correct memory addresses
    for (int i = 0; i < sizeY; ++i)
    {
        array2d[i] = (temp + i * sizeX);
    }



    // Fill the array with ascending numbers
    for (int y = 0; y < sizeY; ++y)
    {
        for (int x = 0; x < sizeX; ++x)
        {
            array2d[y][x] = x + y * sizeX;
        }
    }



    // Code for testing
    // Print the addresses
    for (int y = 0; y < sizeY; ++y)
    {
        for (int x = 0; x < sizeX; ++x)
        {
            std::cout << std::hex << &(array2d[y][x]) << ' ';
        }
    }
    std::cout << "\n\n";

    // Print the array
    for (int y = 0; y < sizeY; ++y)
    {
        std::cout << std::hex << &(array2d[y][0]) << std::dec;
        std::cout << ": ";
        for (int x = 0; x < sizeX; ++x)
        {
            std::cout << array2d[y][x] << ' ';
        }
        std::cout << std::endl;
    }



    // Free memory
    delete[] array2d[0];
    delete[] array2d;
    array2d = nullptr;

    return 0;
}

When you invoke the program with the values sizeX=20 and sizeY=15, the output will be the following:

0x603010 0x603014 0x603018 0x60301c 0x603020 0x603024 0x603028 0x60302c 0x603030 0x603034 0x603038 0x60303c 0x603040 0x603044 0x603048 0x60304c 0x603050 0x603054 0x603058 0x60305c 0x603060 0x603064 0x603068 0x60306c 0x603070 0x603074 0x603078 0x60307c 0x603080 0x603084 0x603088 0x60308c 0x603090 0x603094 0x603098 0x60309c 0x6030a0 0x6030a4 0x6030a8 0x6030ac 0x6030b0 0x6030b4 0x6030b8 0x6030bc 0x6030c0 0x6030c4 0x6030c8 0x6030cc 0x6030d0 0x6030d4 0x6030d8 0x6030dc 0x6030e0 0x6030e4 0x6030e8 0x6030ec 0x6030f0 0x6030f4 0x6030f8 0x6030fc 0x603100 0x603104 0x603108 0x60310c 0x603110 0x603114 0x603118 0x60311c 0x603120 0x603124 0x603128 0x60312c 0x603130 0x603134 0x603138 0x60313c 0x603140 0x603144 0x603148 0x60314c 0x603150 0x603154 0x603158 0x60315c 0x603160 0x603164 0x603168 0x60316c 0x603170 0x603174 0x603178 0x60317c 0x603180 0x603184 0x603188 0x60318c 0x603190 0x603194 0x603198 0x60319c 0x6031a0 0x6031a4 0x6031a8 0x6031ac 0x6031b0 0x6031b4 0x6031b8 0x6031bc 0x6031c0 0x6031c4 0x6031c8 0x6031cc 0x6031d0 0x6031d4 0x6031d8 0x6031dc 0x6031e0 0x6031e4 0x6031e8 0x6031ec 0x6031f0 0x6031f4 0x6031f8 0x6031fc 0x603200 0x603204 0x603208 0x60320c 0x603210 0x603214 0x603218 0x60321c 0x603220 0x603224 0x603228 0x60322c 0x603230 0x603234 0x603238 0x60323c 0x603240 0x603244 0x603248 0x60324c 0x603250 0x603254 0x603258 0x60325c 0x603260 0x603264 0x603268 0x60326c 0x603270 0x603274 0x603278 0x60327c 0x603280 0x603284 0x603288 0x60328c 0x603290 0x603294 0x603298 0x60329c 0x6032a0 0x6032a4 0x6032a8 0x6032ac 0x6032b0 0x6032b4 0x6032b8 0x6032bc 0x6032c0 0x6032c4 0x6032c8 0x6032cc 0x6032d0 0x6032d4 0x6032d8 0x6032dc 0x6032e0 0x6032e4 0x6032e8 0x6032ec 0x6032f0 0x6032f4 0x6032f8 0x6032fc 0x603300 0x603304 0x603308 0x60330c 0x603310 0x603314 0x603318 0x60331c 0x603320 0x603324 0x603328 0x60332c 0x603330 0x603334 0x603338 0x60333c 0x603340 0x603344 0x603348 0x60334c 0x603350 0x603354 0x603358 0x60335c 0x603360 0x603364 0x603368 0x60336c 0x603370 0x603374 0x603378 0x60337c 0x603380 0x603384 0x603388 0x60338c 0x603390 0x603394 0x603398 0x60339c 0x6033a0 0x6033a4 0x6033a8 0x6033ac 0x6033b0 0x6033b4 0x6033b8 0x6033bc 0x6033c0 0x6033c4 0x6033c8 0x6033cc 0x6033d0 0x6033d4 0x6033d8 0x6033dc 0x6033e0 0x6033e4 0x6033e8 0x6033ec 0x6033f0 0x6033f4 0x6033f8 0x6033fc 0x603400 0x603404 0x603408 0x60340c 0x603410 0x603414 0x603418 0x60341c 0x603420 0x603424 0x603428 0x60342c 0x603430 0x603434 0x603438 0x60343c 0x603440 0x603444 0x603448 0x60344c 0x603450 0x603454 0x603458 0x60345c 0x603460 0x603464 0x603468 0x60346c 0x603470 0x603474 0x603478 0x60347c 0x603480 0x603484 0x603488 0x60348c 0x603490 0x603494 0x603498 0x60349c 0x6034a0 0x6034a4 0x6034a8 0x6034ac 0x6034b0 0x6034b4 0x6034b8 0x6034bc 

0x603010: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 
0x603060: 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 
0x6030b0: 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 
0x603100: 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 
0x603150: 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 
0x6031a0: 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 
0x6031f0: 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 
0x603240: 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 
0x603290: 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 
0x6032e0: 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 
0x603330: 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 
0x603380: 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 
0x6033d0: 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 
0x603420: 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 
0x603470: 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299

As you can see, the multidimensional array lies contiguously in memory, and no two memory addresses are overlapping. Even the routine for freeing the array is simpler than the standard way of dynamically allocating memory for every single column (or row, depending on how you view the array). Since the array basically consists of two linear arrays, only these two have to be (and can be) freed.

This method can be extended for more than two dimensions with the same concept. I won't do it here, but when you get the idea behind it, it is a simple task.

I hope this code will help you as much as it helped me.

There are two general techniques that I would recommend for this in C++11 and above, one for compile time dimensions and one for run time. Both answers assume you want uniform, two-dimensional arrays (not jagged ones).

Compile time dimensions

Use a std::array of std::array and then use new to put it on the heap:

// the alias helps cut down on the noise:
using grid = std::array<std::array<int, sizeX>, sizeY>;
grid * ary = new grid;

Again, this only works if the sizes of the dimensions are known at compile time.

Run time dimensions

The best way to accomplish a 2 dimensional array with sizes only known at runtime is to wrap it into a class. The class will allocate a 1d array and then overload operator [] to provide indexing for the first dimension. This works because in C++ a 2D array is row-major:

^{(Taken from http://eli.thegreenplace.net/2015/memory-layout-of-multi-dimensional-arrays/)}

A contiguous sequence of memory is good for performance reasons and is also easy to clean up. Here's an example class that omits a lot of useful methods but shows the basic idea:

#include <memory>

class Grid {
  size_t _rows;
  size_t _columns;
  std::unique_ptr<int[]> data;

public:
  Grid(size_t rows, size_t columns)
      : _rows{rows},
        _columns{columns},
        data{std::make_unique<int[]>(rows * columns)} {}

  size_t rows() const { return _rows; }

  size_t columns() const { return _columns; }

  int *operator[](size_t row) { return row * _columns + data.get(); }

  int &operator()(size_t row, size_t column) {
    return data[row * _columns + column];
  }
}

So we create an array with std::make_unique<int[]>(rows * columns) entries. We overload operator [] which will index the row for us. It returns an int * which points to the beginning of the row, which can then be dereferenced as normal for the column. Note that make_unique first ships in C++14 but you can polyfill it in C++11 if necessary.

It's also common for these types of structures to overload operator() as well:

  int &operator()(size_t row, size_t column) {
    return data[row * _columns + column];
  }

^{Technically I haven't used new here, but it's trivial to move from std::unique_ptr<int[]> to int * and use new/delete.}

typedef is your friend

After going back and looking at many of the other answers I found that a deeper explanation is in order, as many of the other answers either suffer from performance problems or force you to use unusual or burdensome syntax to declare the array, or access the array elements ( or all the above ).

First off, this answer assumes you know the dimensions of the array at compile time. If you do, then this is the best solution as it will both give the best performance and allows you to use standard array syntax to access the array elements.

The reason this gives the best performance is because it allocates all of the arrays as a contiguous block of memory meaning that you are likely to have less page misses and better spacial locality. Allocating in a loop may cause the individual arrays to end up scattered on multiple non-contiguous pages through the virtual memory space as the allocation loop could be interrupted ( possibly multiple times ) by other threads or processes, or simply due to the discretion of the allocator filling in small, empty memory blocks it happens to have available.

The other benefits are a simple declaration syntax and standard array access syntax.

In C++ using new:

#include <stdio.h>
#include <stdlib.h>

int main(int argc, char **argv) {

typedef double (array5k_t)[5000];

array5k_t *array5k = new array5k_t[5000];

array5k[4999][4999] = 10;
printf("array5k[4999][4999] == %f\n", array5k[4999][4999]);

return 0;
}

Or C style using calloc:

#include <stdio.h>
#include <stdlib.h>

int main(int argc, char **argv) {

typedef double (*array5k_t)[5000];

array5k_t array5k = calloc(5000, sizeof(double)*5000);

array5k[4999][4999] = 10;
printf("array5k[4999][4999] == %f\n", array5k[4999][4999]);

return 0;
}

I don't know for sure if the following answer wasn't provided but I decided to add some local optimizations to the allocation of 2d array (e.g., a square matrix is done through only one allocation): int** mat = new int*[n]; mat[0] = new int [n * n];

However, deletion goes like this because of linearity of the allocation above: delete [] mat[0]; delete [] mat;

Here, I have two options. The first one shows the concept of an array of arrays or pointer of pointers. I prefer the second one because the addresses are contiguous, as you can see in the image.

#include <iostream>

using namespace std;


int main(){

    int **arr_01,**arr_02,i,j,rows=4,cols=5;

    //Implementation 1
    arr_01=new int*[rows];

    for(int i=0;i<rows;i++)
        arr_01[i]=new int[cols];

    for(i=0;i<rows;i++){
        for(j=0;j<cols;j++)
            cout << arr_01[i]+j << " " ;
        cout << endl;
    }


    for(int i=0;i<rows;i++)
        delete[] arr_01[i];
    delete[] arr_01;


    cout << endl;
    //Implementation 2
    arr_02=new int*[rows];
    arr_02[0]=new int[rows*cols];
    for(int i=1;i<rows;i++)
        arr_02[i]=arr_02[0]+cols*i;

    for(int i=0;i<rows;i++){
        for(int j=0;j<cols;j++)
            cout << arr_02[i]+j << " " ;
        cout << endl;
    }

    delete[] arr_02[0];
    delete[] arr_02;


    return 0;
}

Start by defining the array using pointers (Line 1):

int** a = new int* [x];     //x is the number of rows
for(int i = 0; i < x; i++)
    a[i] = new int[y];     //y is the number of columns

This question was bugging me - it's a common enough problem that a good solution should already exist, something better than the vector of vectors or rolling your own array indexing.

When something ought to exist in C++ but doesn't, the first place to look is boost.org. There I found the Boost Multidimensional Array Library, multi_array. It even includes a multi_array_ref class that can be used to wrap your own one-dimensional array buffer.

Below example may help,

int main(void)
{
    double **a2d = new double*[5]; 
    /* initializing Number of rows, in this case 5 rows) */
    for (int i = 0; i < 5; i++)
    {
        a2d[i] = new double[3]; /* initializing Number of columns, in this case 3 columns */
    }

    for (int i = 0; i < 5; i++)
    {
        for (int j = 0; j < 3; j++)
        {
            a2d[i][j] = 1; /* Assigning value 1 to all elements */
        }
    }

    for (int i = 0; i < 5; i++)
    {
        for (int j = 0; j < 3; j++)
        {
            cout << a2d[i][j] << endl;  /* Printing all elements to verify all elements have been correctly assigned or not */
        }
    }

    for (int i = 0; i < 5; i++)
        delete[] a2d[i];

    delete[] a2d;


    return 0;
}