FORTRAN LESSON 4

Lesson Topics
Statement FunctionsDo While Loops
Continuation LinesSign Function
View DemosDownload Demos
# 1 # 2 # 3 # 4 # 5 # 6 # 1 # 2 # 3 # 4 # 5 # 6
Main Fortran Page

Statement Functions

A statement function in Fortran is like a single line function definition in Basic. These are useful in defining functions that can be expressed with a single formula. A statement function should appear before any executable statement in the program, but after any type declaration statements. The format is simple - just type

f(x,y,z,…) = formula   .

You may replace f with any name you like for your function, and x, y, z, … with your own variable names. Instead of formula type the formula for your function.

Examples :
area(r) = pi * r * r
vol(r,h) = pi * r * r * h
f(x,y,z) = sqrt(x / y) * cos(z)

You should declare a type for the function in a declaration statement. Here is a program using a statement function, named "area", to compute areas of circles; the program computes in double precision the area of an annulus of inner radius a and outer radius b:

program annulus
double precision r, area, pi, a, b
parameter (pi = 3.1415926535897932D0)
area(r) = pi * r * r
print *, "Enter the inner and outer radii of the annulus: "
read *, a, b
write (*,10) "The area of the annulus is ", area(b) - area(a)
10format (a,f25.15)
end

In the type declaration statement just include the name of the function - do not include the parentheses or the function variables.

Observe that variables plugged into the function need not be the same variables used in defining the function.

It is possible to use a previous statement function in the definition of another. In the above program, for example, we have already defined the function area(r), so we could define further a second function "annarea", giving the area of the annulus as

annarea(a,b) = area(b) - area(a)    .

But this second function definition must appear later in the program than the first one.

Continuation Lines

Sometimes a Fortran statement will not all fit into columns 7-72. In such a case you may continue the statement onto the next line by placing a character in column 6 of that next line. Although any character is allowed, most programmers use "+", "&", or a digit (using 2 for the first continuation line, 3 for another if necessary, and so on).

Example :
det = a(1,1) * a(2,2) * a(3,3) + a(1,2) * a(2,3) * a(3,1)
& + a(2,1) * a(3,2) * a(1,3) - a(3,1) * a(2,2) * a(1,3)
& - a(2,1) * a(1,2) * a(3,3) - a(1,1) * a(3,2) * a(2,3)

Do While Loops

A do while loop in Fortran is similar to the same loop in Basic. However, in Fortran the test must be enclosed in parentheses, and the end of the loop is identified with either end do or a labeled continue statement. As in "if … then" constructions, in loop tests one uses letter abbreviations for relations such as "≤", ">", "=", etc. Here are two loops adding the squares of the integers from 1 to 10; they differ only in the way the loops are terminated:

N = 1 | N = 1
S = 0 | S = 0
do while (N .le. 10) | do 5 while (N .le. 10)
S = S + N ** 2 | S = S + N ** 2
N = N + 1 | N = N + 1
end do | 5 continue

Sign Function

The function sign in Fortran is called the sign transfer function. It is a function of two variables, and its definition involves two cases:

CASE 1:   If y ≥ 0 then
sign(x,y) = abs(x)   ,
CASE 2:   If y < 0 then
sign(x,y) = - abs(x)   .

The practical effect is that sign(x,y) has the same absolute value as x, but it has the same sign as y; thus the sign of y is transferred to x. (The case y = 0 is a little special - it gives sign(x,y) always a plus sign.)

Examples :

sign(2,3) = 2   ,   sign(2, -3) = - 2   ,   sign(-2,3) = 2   ,   sign(-2, -3) = - 2   .

The variables x and y in sign(x,y) may be integers or real numbers, and either single or double precision. (And x and y may even be of different types.)

If we substitute x = 1 in the sign transfer function, we get the sign of y; that is,

CASE 1:   If y ≥ 0 then
sign(1,y) = 1   ,
CASE 2:   If y < 0 then
sign(1,y) = - 1   .

Thus, sign(1,y) in Fortran is essentially the same as the function SGN(y) in Basic (except when y = 0, when the Fortran value is + 1 but the Basic value is 0).