What are the differences between using int[][] and int[,]?

Question

Coming from a perl background, I have always defined a 2D array using int[][]. I know you can use int[,] instead so what are the differences?

Answer 1

The difference here is that the first sample, int[][] creates a jagged array, while the second creates a rectangular array (of dimension 2). In a jagged array each "column" can be of a different size. In a true multidimensional array, each "column" (in a dimension) is the same size. For more complete information see the Array section of the C# Programming Guide.

Answer 2

Here's a good comparison

Basically int[][] is a "jagged" array, it looks like this:

[] -> [1, 2, 3]
[] -> [1, 2]
[] -> [1, 2, 3, 4]

While int[,] is a multidimentional array which always has the same width and height:

[1, 2, 3, 4
 5, 6, 7, 8
 9, 1, 2, 3]

Each have their uses depending on what you're trying to accomplish.

Answer 3

int[][] is an array of arrays or "jagged" array: you can use this when you want different sizes in the second dimension. For example, the first sub array can have 5 elements and the second can have 42.

int[,] is a two dimensional array: The second dimension is the same through out the array. With int[7, 42] the second dimension is 42 for all 7 lines.

Answer 4

int[][] is a jagged array, where int[,] is a two dimensional array.

plainly

var a = int[][]

allows you do have an array like this:

a[0] = new int[2];
a[1] = new int[5];

where as with int[,] you must always have the second portion of the array be the same:

var a = int[2,2];

a[0,0]
a[0,1]
a[1,0]
a[1,1]

you can't have a[2,2];

Answer 5

int[][] is called an array of arrays, it can have arbitrary length for each row.

int[,] is called rectangular array, where all the rows have the same length. it Can be simulated by the first one.

Answer 6

The best way to understand the difference is to look at two ways to create an nxn matrix:

const int n = 8, m = 8;

int[,] matrix1 = new int[n,m];

int[][] matrix2 = new int[n][];
for (int i = 0; i < matrix2.Length; i++) { matrix2[i] = new int[m]; }

matrix1[1,1] = matrix2[2][2];

As you can see the second one is a bit more involved because you need the for-loop to fully create it. It is often called a jagged array because the 2nd order arrays do not need to be all of the same length.

Answer 7

One thing to consider about a jagged array is that you are allocating non-contiguous chunks of memory. This is a good thing if you have a large array of large objects. I've seen SomeThing[n,n] create StackOverflow issues, but SomeThing[n][n] be ok.

Also, if an object gets > 85,000 bytes it goes to the LOH (Large Object Heap). .NET Collections and the Large Object Heap (LOH)

Answer 8

Here's an excellent article about arrays, covering this topic very well.

Answer 9

you can see int[][] as to (int[])[] (int[]) is a object