Structures

Frequently, one needs a compound type where the different parts of the compound have different types. These different parts are called fields. This data object is called a structure, and is declared according to the following pattern:

	struct tag { declaration list } variable_list;

where:

	struct is the keyword for structures
	tag is an optional identifier which gives a unique tag name
		to this particular structure being declared
	declaration list is a list of declarations, i.e. types followed
		by field names (perhaps with a list separated by commas for
		the same type) and type-name_list pairs separated by a semicolon
	variable_list is an optional list of comma separated identifiers
		each of which will have type "struct" tag
	ending semicolon, ;, as with all declarations.

    struct student {
       char name[40];
       int  age, /* in months */
            height; /* in centimeters */
       float weight; /* in kg */
       struct student *next;
    } assistant, class[12], *first;

which means: the structure with tag name student has 5 fields. They are (in words):

	an array of 40 char called name
	a int called age, and one called height
	a float called weight
	a pointer to struct student called next

Lastly, three variables have been declared to be type struct student, one called assistant, and another called class which is actually an array of 12 structs with tag name'student, and first which is of type pointer to struct student. The last field, named next could be used to build a linked list of students starting with first.

The only operations allowed on whole structures is assignment, provided both the left side and right side are of the same structure type.

To refer to the fields of a structure, the field selection operator period, ., is used. Thus from the above example:

	assistant.height

is of type "int" it is the height of the student assistant, while:

	class[3].name

is of type pointer tochar (or: char *), and its value is the address of the array of characters which is the name of the fourth, i.e. index value 3, student in the array class.

If instead of an identifier whose type is struct we have a pointer to struct then, to select the height field, we can either write:

    (*first).height

or use the abbreviation:

    first->height

Although the abbreviation only saves one of four keystrokes, its main advantage visually is the operator is sequentially in one place and hence is clearer.