 |
» |
|
|
|
An array constructor is used to assign
values to an array. The generated values are supplied from a list
of scalar values, arrays of any rank, and implied DO
specifications. An array constructor may appear in any context in
which a rank-one array expression is allowed. An array with a rank
greater than one may be constructed by using the RESHAPE
intrinsic function. The type of an array constructor is taken from
the values in the list, which must all have the same type and type
parameters (including character length). The extent is taken from
the number of values specified. The syntax of an array constructor is: (/ ac-value-list /) where ac-value-list is a comma-separated
list of one or more ac-values. Each ac-value
may be any of the following: Scalar expressions, for example: An array expression, for example: where the values in x(0)
through x(5) become the values
of the array constructor. If the array the value list has a rank
greater than one, the values are generated in column-major order,
as explained in “Array fundamentals”. An implied-DO specification,
taking the form: (ac-value-list, do-var =
expr1, expr2 [, expr3]) where do-var is the name of a scalar
integer variable, expr1 is the initial
value, expr2 is the final value, and
expr2 is the stride (the default is 1).
For example:
When used to initialize an array in a type declaration or
in an assignment statement, all elements in the array must be initialized.
For example, the following is illegal: INTEGER :: i(10) = (/ 1, 2, 3 /) ! ILLEGAL: too few
! initializers
|
If no values are supplied, the array constructor is zero-sized.
For example, the size of the following array constructor: depends on the value of the variable n;
if the value of the variable is less than 10, then the constructor
contains no values. If the list contains only constant values, the array constructor
may initialize a named constant or a type declaration statement.
An array constructor may not initialize variables in a DATA
statement, which may contain scalar constants only. As an extension, HP Fortran 90 allows the
use of [ and
] in place of
(/ and /). The following are examples of array constructors: ! array x is assigned three real values.
x = (/19.3, 24.1, 28.6/)
! One vector, consisting of 16 integer values, is assigned to j
j = (/4, 10, k(1:5), 2 + l, (m(n), n = -7,-2),16, 1/)
! assign 5 values
a = (/(base(k), k=1,5)/)
! The named constant t is a rank-one array initialized with
! the values 36.0 and 37.0
REAL,DIMENSION(2):: t
PARAMETER (t=(/ 36.0, 37.0/))
! the array constructor is reshaped as 1 3 5 7
! 2 4 6 8
! and is then assigned to z
z=RESHAPE((/1,2,3,4,5,6,7,8/), (/2,4/) )
! an array constructor is used for the second component of
! the structure constructor
alaska = site("NOME",(/-63,4/))
diagonal = (/ (b(i,i), i=1,n) /)
hilbert = RESHAPE( (/ ((1.0/(i+j), i=1,n), j=1,n) /), (/ n,n /) )
ident = RESHAPE ( (/ (1, (0, i=1,n), j=1,n-1), 1 /), (/ n,n /) ) |
As shown in last three examples, an array constructor with
implied- DO loops
and the RESHAPE
function permit construction of arrays that cannot otherwise be
expressed conveniently with alternative notations. |
Printable version
