Arrays and Intrinsic Functions | GFortran Reference

Arrays as first-class objects

In Fortran an array is a first-class object: it is declared, assigned, sliced, and passed to procedures as a unit, in much the same way as a scalar. Arrays may be fixed-size or allocatable:

real :: temperatures(365)          ! fixed size
real, allocatable :: grid(:,:)    ! size set later
allocate(grid(100,100))

Whole-array expressions

A whole-array expression operates on every element at once, which is both clearer and easier for compilers to optimize than an explicit loop. The two statements below are equivalent:

do i = 1, n
  y(i) = sin(x(i))
end do

y = sin(x)     ! same result, whole-array form

Common intrinsic functions

The language ships with a large set of intrinsic functions; the array-related ones do the most work in scientific code:

sum — the sum of an array's elements.
product — the product of an array's elements.
maxval and minval — the largest and smallest elements.
dot_product — the dot product of two rank-one arrays.
matmul — matrix multiplication of two rank-two arrays.
transpose, size, shape — shape and layout inquiries.

A compact example:

program array_sum
  implicit none
  real :: a(5)

  a = [1.0, 2.0, 3.0, 4.0, 5.0]   ! whole-array assignment
  print *, 'Total: ', sum(a)
end program array_sum

Array sections

A section selects part of an array with a subscript triplet — a(2:4) is elements 2 through 4, and a(1:10:2) takes every second element. Sections can be assigned, passed to procedures, and used in expressions. To copy a section into a separate variable, assign it to an allocatable array, which takes the correct size automatically:

real, allocatable :: part(:)
part = data(500:999)

The Fortran Intrinsics Quick Reference lists the most-used intrinsics in one place.