The if Statement

To control whether statements are executed based on a condition in a Perl program, you usually use an if statement. The syntax of an if statement is as follows:

if (expression) block
					

The statement works like this: If the expression evaluates to true, the block of code is run. If the expression is false, the block of code is not run. Remember that the block includes the braces. Consider this example:

if ( $r == 5 ) {
    print 'The value of $r is equal to 5.';
}

The expression being tested is $r == 5. The == symbol is an equality operator. If the two operands on either side—$r and 5—are numerically equal to one another, the expression is considered to be true, and the print statement is executed. If $r is not equal to 5, the print statement is not executed.

The if statement can also run one set of statements if a condition is true and another set of statements if it's not. That structure is called an if-else statement. The syntax looks like this:

if (expression)      #If expression is true...
    block1           # ...this block of code is run.
else
    block2           # Otherwise this block is run.

The first block, block1, is run only if the expression is true; if the expression is not true, block2, following the else, is run. Now consider this example:

$r=<STDIN>;
chomp $r;
if ($r == 10) {
    print '$r is 10', "
";
} else {
    print '$r is something other than 10...', "
";
    $r=10;
    print '$r has been set to 10', "
";
}

Yet another way of structuring an if statement is to check multiple expressions and run code depending on which expressions are true:

if (expression1)     # If expression1 is true ...
    block1           # ...run this block of code.
elsif (expression2)  # Otherwise, if expression2 is true...
    block2           # ...Run this block of code.
else
    block3           # If neither expression was true, run this.

You can read the preceding block like this: If the expression labeled expression1 is true, the block block1 is run. Otherwise, control falls to the elsif and expression2 is tested; if it's true, block2 is run. If neither expression1 nor expression2 is true, block3 is run. The following is an example of real Perl code that demonstrates this syntax:

$r=10;
if ($r==10) {
    print '$r is 10!';
} elsif ($r == 20) {
    print '$r is 20!';
} else {
    print '$r is neither 10 nor 20';
}

The Other Relational Operators

So far, you've been comparing numeric quantities in your if statements with the equality operator, ==. Perl actually has quite a few operators for comparing numeric values, most of which are listed in Table 3.1.

To use these operators, you can simply put them in anywhere that your program needs to test relations between numeric values. An example of the use of these operators in an if statement is shown in Listing 3.1, which you can type in and run (do not type the line numbers—”1:” and so forth). The various parts of the program work as follows:

Listing 3.1. A Small Number Guessing Game

1:   #!/usr/bin/perl -w
2:
3:   $im_thinking_of=int(rand 10);
4:   print "Pick a number:";
5:   $guess=<STDIN>;
6:   chomp $guess;   # Don't forget to remove the newline!
7:
8:   if ($guess>$im_thinking_of) {
9:       print "You guessed too high!
";
10:  } elsif ($guess < $im_thinking_of) {
11:      print "You guessed too low!
";
12:  } else {
13:      print "You got it right!
";
14:  }

Line 1: This line is the standard first line of a Perl program; it indicates the interpreter you want to run and the -w switch to enable warnings. See Hour 1, “Getting Started with Perl”; your first line may need to look slightly different.

Line 3: The (rand 10) function picks a number between 0 and 10, and the int()function truncates it so that only integers 0 to 9 are assigned to $im_thinking_of.

Lines 4–6: These lines ask the user for the guess, assign it to $guess, and remove the trailing newline character.

Lines 8–9: If $guess is greater than the number in $im_thinking_of, these lines print an appropriate message.

Lines 10–11: Otherwise, if $guess is less than the number in $im_thinking_of, these lines print that message.

Lines 12–13: The only choice left is that the user guessed the number.

The operators in Table 3.1 are used only for testing numeric values. Using them to test nonalphabetic data results in behavior that you probably don't want. Consider this example:

$first="Simon";
$last="simple";
if ($first == $last) {    # == is not what you want!
    print "The words are the same!
";  
}

The two values $first and $last actually test equal to each other. The reason was explained in Hour 2: If nonnumeric strings are used when Perl is expecting numeric values, the strings evaluate to zero. So the preceding if expression looks something like this to Perl: if ( 0 == 0 ). This expression evaluates to true, and that's probably not what you wanted.

If you want to test nonnumeric values, you can use another set of Perl operators, which are listed in Table 3.2.

These operators decide “greater than” and “less than” by examining each character left to right and comparing them in ASCII order. This means that strings sort in ascending order: most punctuation first, and then numbers, uppercase, and finally lowercase. For example, 1506 compares less than Happy, which compares less than happy.

What Truth Means to Perl

Up to this point, you've been reading about “if this expression is true…” or “…evaluates to true…,” but you haven't seen any formal definition of what Perl thinks “true” is. Perl has a few short rules about what is true and what is not true, and the rules actually make sense when you think about them for a bit. The rules are as follows:

• The number 0 is false.

• The empty string ("") and the string "0" are false.

• The undefined value undef is false.

• Everything else is true.

Make sense? The only other point to remember is that when you're testing an expression to see whether it's true or false, the expression is simplified—functions are called, operators are applied, math expressions are reduced, and so on—and then converted to a scalar value for evaluation to determine whether it is true or false.

Think about these rules, and then take a look at Table 3.3. Try to guess whether the expression is true or false before you look at the answer.

Until now, you've seen only expressions with relational operators as the conditions in if statements. Actually, you can use any expression that will evaluate to true or false the way you would want:

# The scalar variable $a is evaluated for true/false
if ($a) {  ...  }

# Checks the length of $b.  If nonzero, the test is true.
if (length($b)) { ....  }

Recall from Hour 2 that the assignment operator = returns a value—the value that was assigned. That value, of course, is also true or false:

$a = 1;
$b = 2;
print qq(The statement "$a = $b" is );
if ($a = $b) { # value is 2, therefore true
    print "true";
} else {
    print "false";
}

This code prints The statement "1 = 2" is true. Now you see why using = when you meant == is such a pitfall.

The value undef is a special value in Perl. Variables that have not yet been set have the value of undef, and some functions return undef on failure. It's not 0, and it's not a regular scalar value. It's kind of special. In a test for truth, undef always evaluates to false. If you try to use the undef value in a math expression, it's treated as though it were 0.

Using variables that haven't been set yet is usually a sign of a programming error. If you're running your Perl programs with warnings enabled, the value undef in an expression or as an argument to some functions causes Perl to generate the warning Use of uninitialized value.

Logical Operators

When you're writing programs, you sometimes need to code something like the following: Do this if $x is true and if $y is true, but not if $z is true. You can code this example into a series of if statements, but it's not pretty:

if ($x) {
    if ($y) {
        if ($z) {
            # do nothing
        } else {
            print "All conditions met.
";
        }
    }
}

Perl has a whole class of operators for connecting together true and false statements like this, called logical operators. The logical operators are shown in Table 3.4.

Using the operators in Table 3.4, you could rewrite the previous snippet much more concisely as follows:

if ($x and $y and not $z ) {
    print "All conditions met.
";
}

Expressions connected with logical operators are evaluated from left to right, until a value of true or false can be determined for the entire expression. Examine the following code:

1:   $a=0;
2:   $b=1;
3:   $c=2;
4:   $d="";
5:  if ($a and $b) {  print '$a and $b are true'; }
6:  if ($d or $b) { print 'either $d or $b is true'; }
7:  if ($d or not $b or $c) 
8:      { print '$d is true, or $b is false or $c is true'; }

Lines 1–4: These lines give the variables default values.

Line 5: $a is evaluated first. It is false, so the and expression cannot possibly be true. $b is never evaluated; it doesn't have to be because the truth of the expression is known after evaluating $a. The print is not executed.

Line 6: $d is evaluated first. It is false. Even if $d is false, the expression might still be true—because it contains a logical or—so $b is examined next. $b turns out to be true; therefore, the expression is true, and the print happens.

Line 7: $d is evaluated first. It is false. But although $d is false, the expression might still be true—as seen in line 4—because it contains a logical or. Next, the truth of $b—1, so true—is negated, so this expression becomes false. The truth of the or statement cannot be determined yet, so $c is evaluated. $c turns out to be true, so the whole expression is true, and the print happens.

This behavior—stopping the evaluation of a logical expression as soon as the truth can be determined—is called short-circuiting. This feature is used by Perl programmers to construct simple flow-control statements out of logical operators and to avoid the if statement entirely:

$message="A and B are both true."
($a and $b) or $message="A and B are not both true.";

In the preceding example, if either $a or $b is false, the right side of the or must be evaluated, and the message is changed. If both $a and $b are true, the or must be true, and it's not necessary to evaluate the right side. The truth value of the entire expression isn't used at all; this example uses the short-circuit side effects of the and and or operators to manipulate $message.

An interesting property of Perl's logical operators is that they don't simply return true or false. They actually return the last value evaluated. For example, the expression 5 && 7 doesn't just return true—it returns 7. This allows constructs like this:

# Set $new to old value if $old is true, 
# otherwise use the string "default".
$new=$old || "default"; 

which is a little more concise than the code

$new=$old;
if (! $old) {  # was $old empty (or false)?
    $new="default";
}

This trick can make your code less readable. Understanding how it works can be helpful, though, if you are going to be looking at much Perl code written by others.