SELECT CASE Statement

Fortran has one more selective execution statement, SELECT CASE, and could be very handy if it is used properly. The SELECT CASE statement is also usually referred to as the CASE statement. The following is its syntactic form:

SELECT CASE (selector)
   CASE (label-list-1)
      statements-1
   CASE (label-list-2)
      statements-2
   CASE (label-list-3)
      statements-3
     .............
   CASE (label-list-n)
      statements-n
   CASE DEFAULT
      statements-DEFAULT
END SELECT

where statements-1, statements-2, statements-3, ..., statements-n and statements-DEFAULT are sequences of executable statements, including the SELECT CASE statement itself, and selector is an expression whose result is of type INTEGER, CHARACTER or LOGICAL (i.e., no REAL type can be used for the selector). The label lists label-list-1, label-list-2, label-list-3, ..., and label-list-n are called case labels.
A label-list is a list of labels, separated by commas. Each label must be one of the following forms. In fact, three of these four are identical to an extent specifier for substrings:

value value-1 : value-2 value-1 : : value-2

where value, value-1 and value-2 are constants or alias defined by PARAMETER. The type of these constants must be identical to that of the selector.

The first form has only one value

The second form means all values in the range of value-1 and value-2. In this form, value-1 must be less than value-2.

The third form means all values that are greater than or equal to value-1

The fourth form means all values that are less than or equal to value-2
The rule of executing the SELECT CASE statement goes as follows:

The selector expression is evaluated

If the result is in label-list-i, then the sequence of statements in statements-i are executed, followed by the statement following END SELECT

If the result is not in any one of the label lists, there are two possibilities:

if CASE DEFAULT is there, then the sequence of statements in statements-DEFAULT are executed, followed by the statement following END SELECT

if there is no CASE DEFAULT is not there, the statement following END SELECT is executed.

There are some notes for writing good Fortran programs:

The constants listed in label lists must be unique

CASE DEFAULT is optional. But with a CASE DEFAULT, you are guaranteed that whatever the selector value, one of the labels will be used. I strongly suggest to add a CASE DEFAULT in every SELECT CASE statement.

The place for CASE DEFAULT can be anywhere within a SELECT CASE statement; however, put it at the end would be more natural.

Examples

The following uses Class to execute a selection. If Class is 1, Freshman is displayed; if Class is 2, Sophomore is displayed; if Class is 3, Junior is displayed; if Class is 4, Senior is displayed; and if Class is none of the above, Hmmmm, I don't know is displayed. After displaying one of the above, Done is displayed.

INTEGER :: Class SELECT CASE (Class) CASE (1) WRITE(*,*) 'Freshman' CASE (2) WRITE(*,*) 'Sophomore' CASE (3) WRITE(*,*) 'Junior' CASE (4) WRITE(*,*) 'Senior' CASE DEFAULT WRITE(*,*) "Hmmmm, I don't know" END SELECT WRITE(*,*) 'Done'

In the following, the value of CHARACTER variable Title is used as a selector. If its value is "DrMD" (resp., "PhD", "MS" and "BS"), then INTEGER variables DrMD (resp., PhD, MS and BS) is increased by 1. Otherwise, Others is increased.

CHARACTER(LEN=4) :: Title INTEGER :: DrMD = 0, PhD = 0, MS = 0, BS = 0, Others = 0 SELECT CASE (Title) CASE ("DrMD") DrMD = DrMD + 1 CASE ("PhD") PhD = PhD + 1 CASE ("MS") MS = MS + 1 CASE ("BS") BS = BS + 1 CASE DEFAULT Others = Others + 1 END SELECT

The follow example does not use CASE DEFAULT since the label lists have already covered all possibilities. If the value of Number is less than or equal to -1 (:-1), -1 is stored into Sign. If the value of Number is zero, 0 is stored into Sign. If the value of Number is greater than or equal to 1, 1 is stored into Sign.

INTEGER :: Number, Sign SELECT CASE (Number) CASE ( : -1) Sign = -1 CASE (0) Sign = 0 CASE (1 : ) Sign = 1 END SELECT WRITE(*,*) "Sign = ", Sign

If the value of c is one of the first ten letters (in lower case), 'One of the first ten letters' is displayed; if the value of c is one of l, m, n, o, p, u, v, w, x, y, 'One of l, m, n, o, p, u, v, w, x, y' is displayed; if the value of c is one of z, q, r, s, t, 'One of z, q, r, s, t' is displayed; and if the value of c is none of the above, 'Other characters, which may not be letters' is displayed. In the last case, the value of c could be one of these lower case letters not listed, an upper case letters, a digit, an operator, or any other character. So, if you only want to handle lower case letters, this program is incorrect.

CHARACTER(LEN=1) :: c SELECT CASE (c) CASE ('a' : 'j') WRITE(*,*) 'One of the first ten letters' CASE ('l' : 'p', 'u' : 'y') WRITE(*,*) 'One of l, m, n, o, p, u, v, w, x, y' CASE ('z', 'q' : 't') WRITE(*,*) 'One of z, q, r, s, t' CASE DEFAULT WRITE(*,*) 'Other characters, which may not be letters' END SELECT

The following program uses the value of Number to determine the value of Range.

INTEGER :: Number, Range SELECT CASE (Number) CASE ( : -10, 10 : ) Range = 1 CASE (-5:-3, 6:9) Range = 2 CASE (-2:2) Range = 3 CASE (3, 5) Range = 4 CASE (4) Range = 5 CASE DEFAULT Range = 6 END SELECT

Here is the result:

Number Range Why?

<= -10 1 CASE ( : -10, 10 : )

-9, -8, -7, -6 6 CASE DEFAULT

-5, -4, -3 2 CASE (-5:-3, 6:9)

-2, -1, 0, 1, 2 3 CASE (-2:2)

3 4 CASE (3, 5)

4 5 CASE (4)

5 4 CASE (3, 5)

6, 7, 8, 9 2 CASE (-5:-3, 6:9)

>= 10 1 CASE ( : -10, 10 : )

Number	Range	Why?
<= -10	1	CASE ( : -10, 10 : )
-9, -8, -7, -6	6	CASE DEFAULT
-5, -4, -3	2	CASE (-5:-3, 6:9)
-2, -1, 0, 1, 2	3	CASE (-2:2)
3	4	CASE (3, 5)
4	5	CASE (4)
5	4	CASE (3, 5)
6, 7, 8, 9	2	CASE (-5:-3, 6:9)
>= 10	1	CASE ( : -10, 10 : )