On this page:
array
mutable-array
make-array
build-array
array->mutable-array
mutable-array-copy
indexes-array
index-array
axis-index-array
diagonal-array

6.7 Construction

syntax

(array #[#[...] ...] maybe-type-ann)

 
maybe-type-ann = 
  | : type
Creates an Array from nested rows of expressions.

The vector syntax #[...] delimits rows. These may be nested to any depth, and must have a rectangular shape. Using square parentheses is not required, but is encouraged to help visually distinguish array contents from array indexes and other vectors. (See the examples for indexes-array for an illustration.)

Examples:

> (array 0)

- : (Array Zero)

(array 0)

> (array #[0 1 2 3])

- : (Array Byte)

(array #[0 1 2 3])

> (array #[#[1 2 3] #[4 5 6]])

- : (Array Positive-Byte)

(array #[#[1 2 3] #[4 5 6]])

> (array #[#[1 2 3] #[4 5]])

eval:121:0: array: expected rectangular data

  at: #(#(1 2 3) #(4 5))

  in: (array/syntax array list unsafe-list->array #(#(1 2 3)

#(4 5)))

As with the list constructor, the type chosen for the array is the narrowest type all the elements can have. Unlike list, because array is syntax, instantiating array with the desired element type is a syntax error:
> (list 1 2 3)

- : (Listof Positive-Byte) [more precisely: (List One Positive-Byte Positive-Byte)]

'(1 2 3)

> (array #[1 2 3])

- : (Array Positive-Byte)

(array #[1 2 3])

> ((inst list Real) 1 2 3)

- : (Listof Real)

'(1 2 3)

> ((inst array Real) #[1 2 3])

eval:125:0: array: not allowed as an expression

  in: array

There are two easy ways to annotate the element type:
> (array #[1 2 3] : Real)

- : (Array Real)

(array #[1 2 3])

> (ann (array #[1 2 3]) (Array Real))

- : (Array Real)

(array #[1 2 3])

Annotating should rarely be necessary because the Array type is covariant.

Normally, the datums within literal vectors are implicitly quoted. However, when used within the array form, the datums must be explicitly quoted.
> #(this is okay)

- : (Vector Symbol Symbol Symbol)

'#(this is okay)

> (array #[not okay])

eval:129:0: Type Checker: untyped top-level identifier okay

  in: #(not okay)

> (array #['this 'is 'okay])

- : (Array (U 'this 'is 'okay))

(array #['this 'is 'okay])

> (array #['#(an) '#(array) '#(of) '#(vectors)])

- : (Array (Vector Symbol))

(array #['#(an) '#(array) '#(of) '#(vectors)])

Arrays returned by array are strict. Another way to create immutable, strict arrays from literal data is to use list->array.

syntax

(mutable-array #[#[...] ...] maybe-type-ann)

 
maybe-type-ann = 
  | : type
Creates a Mutable-Array from nested rows of expressions.

The semantics are almost identical to array’s, except the result is mutable:
> (define arr (mutable-array #[0 1 2 3]))
> arr

- : (Mutable-Array Integer)

(mutable-array #[0 1 2 3])

> (array-set! arr #(0) 10)
> arr

- : (Mutable-Array Integer)

(mutable-array #[10 1 2 3])

Because mutable arrays are invariant, this form additionally accepts a type annotation for the array’s elements:
> (define arr (mutable-array #[0 1 2 3] : Real))
> arr

- : (Mutable-Array Real)

(mutable-array #[0 1 2 3])

> (array-set! arr #(0) 10.0)
> arr

- : (Mutable-Array Real)

(mutable-array #[10.0 1 2 3])

Another way to create mutable arrays from literal data is to use vector->array.

procedure

(make-array ds value)  (Array A)

  ds : In-Indexes
  value : A
Returns an array with shape ds, with every element’s value as value. Analogous to make-vector.

Examples:

> (make-array #() 5)

- : (Array Positive-Byte)

(array 5)

> (make-array #(1 2) 'sym)

- : (Array 'sym)

(array #[#['sym 'sym]])

> (make-array #(4 0 2) "Invisible")

- : (Array String)

(array #[#[] #[] #[] #[]])

The arrays returned by make-array do not allocate storage for their elements and are strict.

procedure

(build-array ds proc)  (Array A)

  ds : In-Indexes
  proc : (Indexes -> A)
Returns an array with shape ds and procedure proc. Analogous to build-vector.

procedure

(array->mutable-array arr)  (Mutable-Array A)

  arr : (Array A)
Returns a mutable array with the same elements as arr. The result is a copy of arr, even when arr is mutable.

procedure

(mutable-array-copy arr)  (Mutable-Array A)

  arr : (Mutable-Array A)
Like (array->mutable-array arr), but restricted to mutable arrays. It is also faster.

procedure

(indexes-array ds)  (Array Indexes)

  ds : In-Indexes
Returns an array with shape ds, with each element set to its position in the array.

Examples:

> (indexes-array #())

- : (Array Indexes)

(array '#())

> (indexes-array #(4))

- : (Array Indexes)

(array #['#(0) '#(1) '#(2) '#(3)])

> (indexes-array #(2 3))

- : (Array Indexes)

(array #[#['#(0 0) '#(0 1) '#(0 2)] #['#(1 0) '#(1 1) '#(1 2)]])

> (indexes-array #(4 0 2))

- : (Array Indexes)

(array #[#[] #[] #[] #[]])

The resulting array does not allocate storage for its return value’s elements, and is strict. (It is essentially the identity function for the domain ds.)

procedure

(index-array ds)  (Array Index)

  ds : In-Indexes
Returns an array with shape ds, with each element set to its row-major index in the array.

Examples:

> (index-array #(2 3))

- : (Array Index)

(array #[#[0 1 2] #[3 4 5]])

> (array-flatten (index-array #(2 3)))

- : (Array Index)

(array #[0 1 2 3 4 5])

As with indexes-array, the result does not allocate storage for its elements, and is strict.

procedure

(axis-index-array ds axis)  (Array Index)

  ds : In-Indexes
  axis : Integer
Returns an array with shape ds, with each element set to its position in axis axis. The axis number axis must be nonnegative and less than the number of axes (the length of ds).

Examples:

> (axis-index-array #(3 3) 0)

- : (Array Index)

(array #[#[0 0 0] #[1 1 1] #[2 2 2]])

> (axis-index-array #(3 3) 1)

- : (Array Index)

(array #[#[0 1 2] #[0 1 2] #[0 1 2]])

> (axis-index-array #() 0)

axis-index-array: contract violation

  expected: Index < 0

  given: 0

  argument position: 2nd

  other arguments...:

   '#()

As with indexes-array, the result does not allocate storage for its elements, and is strict.

procedure

(diagonal-array dims    
  axes-length    
  on-value    
  off-value)  (Array A)
  dims : Integer
  axes-length : Integer
  on-value : A
  off-value : A
Returns an array with dims axes, each with length axes-length. (For example, the returned array for dims = 2 is square.) The elements on the diagonal (i.e. at indexes of the form (vector j j ...) for j < axes-length) have the value on-value; the rest have off-value.

Example:

> (diagonal-array 2 7 1 0)

- : (Array (U Zero One))

(array

 #[#[1 0 0 0 0 0 0]

   #[0 1 0 0 0 0 0]

   #[0 0 1 0 0 0 0]

   #[0 0 0 1 0 0 0]

   #[0 0 0 0 1 0 0]

   #[0 0 0 0 0 1 0]

   #[0 0 0 0 0 0 1]])

As with indexes-array, the result does not allocate storage for its elements, and is strict.