On this page


VEX includes an array datatype. This is useful in several places:

  • Supporting ramp parameters.

  • Reading capture data from surface nodes using the import() function.

  • General programming, wherever arrays would be useful.


Currently VEX does not support multi-dimensional arrays.


A function that returns an array defined inside a Snippet, Wrangle, or other function should have the function keyword.

This example shows off some of the crazy things that you can do with arrays:

      string maps[] = { "Mandril.rat", "default.pic" };
      export float alength = 0;
    vector texclr, av[];

    texclr = texture(maps[s+t > 1], s, t);
    av = array( {1,0,0}, vector(nrandom()), t, texclr, {.5,0,0});

    if (fit(noise(s*8), 0, 1, .3, .7) > t)
        av = array(1, {0,1,0}, 0);

    Cf = spline("linear", s, av);
    alength = len(av);

Declaring array types

To declare an array variable, the general form is member_type var_name[]:

// my_array is an array of floats
float   my_array[];

// v is a single vector, vector_array is an array of vectors
vector  v, vector_array[];

// str_array is an array of strings
string  str_array[];

You can optionally put a size inside the square brackets, but the VEX compiler currently ignores it.

To declare a function that returns an array:

// A function which returns an array of vectors
// DOES NOT WORK IN A WRANGLE/SNIPPET: use function keyword then.
vector[] rgb_array()

It is ambiguous if you are a nested function, however. Note that Wrangles and Snippets are always implicitly nested. To declare a nested function that returns an array:

// A function which returns an array of vectors
    // Use the optional 'function' keyword to avoid type ambiguity
    function vector[] rgb_array()

To specify a literal array, use curly braces, with the array members separated by commas:

vector an_array[] = { {1, 2, 3}, {2, 3, 4}, {4, 5, 6} };

vector[] rgb_array()
    return { {1, 0, 0}, {0, 1, 0}, {0, 0, 1} };


Literal arrays are constructed at compile time, so they cannot include variables.

For example, this is an error:

int arr[] = { my_var, other_var + 2 }; // Error

To avoid this error, use the array() function which constructs the array at runtime from any number of arguments:

int arr[] = array( my_var, other_var + 2 );

If you specify scalars where a vector is expected, the compiler assigns the scalar value to all components of the vector:

vector an_array[] = { 1, 2, 3};
// an_array[] == { {1, 1, 1}, {2, 2, 2}, {3, 3, 3} }

The array() function creates an array from its arguments.

int my_array[] = array(1, 2, 3, 4, 5);

You can use array() to generate an array of any type. To force array() to generate vectors (for example):

vector (array (value1, value2, ...) );

Accessing and setting array values

Use arrayname[index] to look up a value by its position in the array.

vector bw[] = { 0, 1 };
// bw[] == { {0, 0, 0}, {1, 1, 1} }
Cf = bw[index];

Array bounds are checked at run time. Reading out of bounds will return 0 or "". This may generate a warning or optional run-time error in the future. Writing past the end of an array will resize the array to include the index written to. The new entries will be set to 0 or "".

Python-style indexing is used. This means negative indices refer to positions from the end of the array.

int nums[] = { 0, 1, 2, 3, 4, 5 };
int n = nums[10];  // Returns 0
int b = nums[-2];  // Returns 4

string strs[] = { };
string s = strs[20];  // Returns ""

You can also assign values using the square brackets notation:

float nums[] = { };
nums[0] = 3.14;

(The getcomp and setcomp functions are equivalents for using the square brackets notation.)


The square-brackets operator also works on vectors. You can use it with matrices as well using a pair of brackets: float a = m3[0][1];

Slicing Arrays

The square-brackets can be used to extract sub-arrays using the Python slicing notation.

int nums[] = { 0, 1, 2, 3, 4, 5 };
int start[] = nums[0:2];  // { 0, 1 }
int end[] = nums[-2:];  // { 4, 5 }
int rev[] = nums[::-1];  // { 5, 4, 3, 2, 1, 0 }
int odd[] = nums[1::2]; // { 1, 3, 5 }

The slice function is the equivalent for using the slice-based square brackets notation.

Copying between arrays and vectors/matrices

The assignment operator supports assigning values between vector types and arrays of floats:

float x[];
// Cf and P are vectors

x = set(P);   // Assigns the components of P to the corresponding
              // members of the array x

Cf = set(x);  // Assigns the first 3 members of x as the
              // components of the vector Cf

If the array is not long enough to fill the vector/matrix, the last member is repeated as often as necessary.

float x[] = {1, 2} // Not long enough to fill a vector
Cf = set(x);  // Cf == {1, 2, 2}

You can also assign between matrix types and arrays of vector2/vector/vector4:

vector2     v2[];
vector      v[];
vector4     v4[];
matrix2     m2 = 1;
matrix3     m3 = 1;
matrix      m4 = 1;

v = set(m3);   // Each row of the 3x3 matrix is put into a vector
m3 = set(v);   // Copy the vectors into the row vectors of the matrix
v4 = set(m4);  // Extract the rows of the matrix into the vector4 array
m4 = set(v4);  // Create a matrix using the vector4's in the array as row vectors

In summary:

Left side = Right side Notes
vector2 float[] E.g. vector2 v = {1,2}
vector float[] E.g. vector v = {1,2,3}
vector4 float[] E.g. vector4 v = {1,2,3,4};
matrix2 float[] E.g. matrix2 m = {1,2,3,4};
matrix2 vector2[] E.g. matrix2 m = { {1,2}, {4,5} };
matrix3 float[] E.g. matrix3 m = {1,2,3,4,5,6,7,8,9};
matrix3 vector[] E.g. matrix3 m = { {1,2,3}, {4,5,6}, {7,8,9}};
matrix float[] E.g. matrix m = {1,2,3,4,5,6,7,8,9.., 16};
matrix vector4[] E.g. matrix m = { {1,2,3,4}, {5,6,7,8}, ... {13,14,15,16}};
float[] vector2 Create an array of 2 floats from the components
float[] vector Create an array of 3 floats from the components
float[] vector4 Create an array of 4 floats from the components
float[] matrix2 Create an array of 4 floats from the matrix2
vector2[] matrix2 Create an array of 2 vector2s from the matrix2
float[] matrix3 Create an array of 9 floats from the matrix3
vector[] matrix3 Create an array of 3 vectors from the matrix3
float[] matrix4 Create an array of 16 floats
vector4[] matrix4 Create an array of 4 vector4s.

Looping over an array

See foreach.

Working with arrays

The following functions let you query and manipulate arrays.


Sets the length of the array. If the array is enlarged, intermediate values will be 0 or "".


Returns the length of an array.


Removes the last item from the array (decreasing the size of the array by 1) and returns it.


Removes the first instance of a value in the array. Returns 1 if an item was removed, or 0 otherwise.


Removes an item at the given index and returns it.


Adds an item to the end of an array (increasing the size of the array by 1).


Gets the value of an array component, the same as array[num].


Sets the value of an array component, the same as array[num] = value.


Efficiently creates an array from its arguments.


Flattens an array of vectors or matrices into an array of floats.


Reverses the effect of serialize: assembles a flat array of floats into an array of vectors or matrices.


An array-based replacement for the neighbourcount/neighbour combo. Returns an array of the point numbers of the neighbors of a given point.

In addition, the following functions work with arrays:

VCC pragmas

The ramp pragma lets you specify a ramp user interface for a set of parameters.

#pragma ramp <ramp_parm> <basis_parm> <keys_parm> <values_parm>

See VCC pragmas for more information.


  • Currently VEX does not support multi-dimensional arrays.

  • Arrays cannot be passed between shaders (through simport, etc.).

  • Arrays cannot be written to image planes.



Next steps