5.5. Data Initialization and the DATA Statement

An entity can be initialized in a type declaration statement. When an initialization expression appears in a declaration for an object that does not have the PARAMETER attribute, the object (which is a variable) is assigned the specified initial value. This object is a variable with explicit initialization. Alternatively, explicit initialization can be specified in a DATA statement unless the variable is a derived type or type alias for which default initialization has been specified.

• The variable is a derived type or type alias for which default initialization has been specified.

The same rules apply to the assignment of the initial value when an assignment statement is executed. For example, if the variable is of type real but the value is an integer value, the variable will be assigned the real equivalent of the integer value. If the kind of the variable is different from the kind of the value, the value will be converted to the kind of the variable. Array constructors and broadcast values can be used to initialize arrays, and structure constructors can be used to initialize variables of user-defined type. The format of a type declaration statement that provides an initial value for a variable is as follows:

type [, attribute_list ] :: object_name [ (array_spec) ] [ * char_length ] = initialization_expr

Subject to the rules governing combinations of these attributes, attribute_list can contain the following:

DIMENSION

POINTER

PRIVATE

PROTECTED (F2003)

PUBLIC

SAVE

STATIC (EXTENSION)

TARGET

VOLATILE (F2003)

For example:

INTEGER :: I = 0

The PARAMETER attribute can also appear in the attribute_list, but in this case, the object is declared to be a named constant.

The value associated with the name cannot be changed during the execution of the program. For example, PI or E can be associated with the familiar mathematical constants to provide more convenient access to these values. Named constants are also used to assign names to values (such as a sales tax rate) that could change at some later time. When a change is necessary, it can be made at one place in the program rather than every place where the value is used. The program can be recompiled to effect the change.

An array name that appears in a declaration statement that contains an initialization_expr must have its dimensionality declared in the same statement or a previous statement.

Initialization of a variable in a type declaration statement or any part of a variable in a DATA statement implies that the variable has the SAVE attribute unless the variable is in a named common block. The automatically acquired SAVE attribute may be reaffirmed by the appearance of SAVE as an attribute in its type declaration statement or by inclusion of the variable name in a separate SAVE statement.

The DATA statement is defined as follows:

The following notes pertain to the preceding format:

• digit must have one of the values 0 through 7 in octal_typeless_constant

• digit must have a value of 0 or 1 in binary_typeless_constant

• The B, O, X, and Z characters can be in uppercase or lowercase.

• The scalar_structure_component must contain at least one part_ref that contains a subscript_list.

Note that a constant value cannot be an array constructor. An array can be initialized by using the array name as the data_stmt_object and supplying values for all the elements of the array in array element order.

If an object is of type real or complex, the corresponding constant must be of type integer, real, or complex. The Cray Fortran Compiler permits a default real object to be initialized with a BOZ, typeless, or character (used as Hollerith) constant. No conversion of the BOZ value, typeless value, or character constant takes place.

The Cray Fortran Compiler permits an integer object to be initialized with a BOZ, typeless, or character (used as Hollerith) constant in a type declaration statement. The Cray Fortran Compiler also allows an integer object to be initialized with a typeless or character (used as Hollerith) constant in a DATA statement.

If an object is of derived type or of a type alias that describes a derived type, the corresponding structure constructor must be of the same type.

The value of the constant, structure constructor (in a DATA statement), or initialization expression (in a type declaration statement) must be such that its value could be assigned to the corresponding variable using an intrinsic assignment statement. The variable becomes initially defined with the value of the constant.

A variable, or the same part of a variable, must not be initialized more than once in an executable program.

An object declared to be of a type that has default initialization cannot be specified in a DATA statement. This object can be initialized in a type specification statement. The initialization in the type specification statement overrides the default initialization.

The following items cannot be initialized:

• A dummy argument

• An object made accessible by use or host association

• A function result

• An automatic object

• An allocatable array

• An external or intrinsic procedure

For an object being initialized, any subscript, section subscript, substring starting point, or substring ending point must be an integer initialization expression.

Each component of a structure constructor used for initialization must be an initialization expression.

If the variable being initialized has the POINTER attribute, then data_stmt_constant must be NULL(). The pointer has an initial association status of disassociated.

A variable that appears in a DATA statement and is thereby declared and typed implicitly can appear in a subsequent type declaration statement only if that declaration confirms the implicit declaration. An array name, array section, or array element appearing in a DATA statement must have had its array properties established previously.

If a DATA statement constant value is a named constant or a structure constructor, the named constant, derived type, or type alias must have been declared previously in the scoping unit or must have been made accessible by USE or HOST association.

An array element or structure component that appears in a DATA statement must not have a constant parent.

The DATA statement repeat factor value must be positive or zero. If it is a named constant, the value must have been specified in a prior statement in the scoping unit that contains the DATA statement or must have been made accessible by use or host association.

In a scalar_constant_subobject that is a data_stmt_repeat, any subscript must be an initialization expression.

In a scalar_constant_subobject that is a data_stmt_constant, any subscript, substring starting point, or substring ending point must be an initialization expression.

A subscript in an array element of an implied-DO list must contain as operands only constants or DO variables of the containing implied-DO s.

The scalar integer loop control expressions in an implied-DO must contain as operands only constants or DO variables of the containing implied-DO s. Each operation must be an intrinsic operation.

The data object list is expanded to form a sequence of scalar variables. An array or array section is equivalent to the sequence of its array elements in array element order. A data_implied_do is expanded to form a sequence of array elements and structure components, under the control of the implied-DO variable, as in the DO construct. A zero-sized array or an implied-DO with an iteration count of 0 contributes no variables to the expanded list, but a character variable declared to have zero length does contribute a variable to the list.

The data value list is expanded to form a sequence of scalar values. Each value must be a constant or constant expression (structure constructor). If a value is represented by a named constant, the named constant must be specified prior to the DATA statement. A DATA statement repeat factor indicates the number of times the following constant value is to be included in the sequence. If the repeat factor is 0, the following value is ignored.

If a data_stmt_constant is a boz_literal_constant, the corresponding object must be of type integer. A data_stmt_constant that is a boz_literal_constant is treated as if the constant were an int_literal_constant with a kind_param that specified the representation method with the largest decimal exponent range supported.

Scalar variables and values of the expanded sequence must be in one-to-one correspondence. Each value specifies the initial value for the corresponding variable. The lengths of the two expanded sequences must be the same.

The following examples show type declaration statement initializations:

CHARACTER(LEN = 10) :: NAME = "JOHN DOE" INTEGER, DIMENSION(0:9) :: METERS = (/ (0, I = 1, 10) /) TYPE(PERSON) :: ME = PERSON(21, "JOHN SMITH"), & YOU = PERSON(35, "FRED BROWN") INTEGER :: BIRD(3) = 1 REAL :: SKEW(100,100) = RESHAPE ( (/((1.0, K = 1, J-1), & (0.0, K = J, 100), J = 1, 100)/), (/ 100, 100 /) )

The following are examples of DATA statement initializations:

CHARACTER*10 NAME INTEGER METERS(0:9) DATA NAME /"JOHN DOE"/, METERS /10*0/

TYPE(PERSON) ME, YOU DATA ME /PERSON(21, "JOHN SMITH")/ DATA YOU%AGE, YOU%NAME /35, "FRED BROWN"/ INTEGER BIRD(3) DATA BIRD /3*1/ REAL SKEW(100, 100) DATA ((SKEW (K, J), K = 1, J-1), J = 1, 100) /4950 * 1.0/ DATA ((SKEW (K, J), K = J, 100), J = 1, 100) /5050 * 0.0/

In both forms, the character variable NAME is initialized with the value JOHN DOE with padding on the right because the length of the constant is less than the length of the variable. All ten elements of the integer array METERS are initialized to 0; an array constructor is used in the type declaration statement form; a repeat factor is used for the DATA statement form. ME and YOU are structures declared using the user-defined type PERSON defined in Section 4.8. In both forms ME is initialized using a structure constructor. In the DATA statement form YOU is initialized by supplying a separate value for each component.

In the type declaration statement form, the value 1 is broadcast to all 3 elements of BIRD. In the DATA statement form, a value must be supplied for each element of BIRD.

In both forms, the two-dimensional array SKEW is initialized so that the lower triangle is 0 and the strict upper triangle is 1. The RESHAPE(3i) intrinsic function is required in the first form because SKEW is of rank 2. Repeat factors are used in the second form.